Resource

Integrated Single-Cell Analysis Maps the Continuous

Regulatory Landscape of Human Hematopoietic
Differentiation
Graphical Abstract Authors
FACS known cell types Integrated single-cell analysis Jason D. Buenrostro, M. Ryan Corces,
HSC i) cell type analysis Caleb A. Lareau, ..., Ravindra Majeti,
Howard Y. Chang, William J. Greenleaf
Differentiation

HSC
CMP

GMP
Correspondence
jbuen@broadinstitute.org (J.D.B.),
FACS scATAC-seq clusters wjg@stanford.edu (W.J.G.)
Lymphoid Myeloid Erythroid

single-cell ATAC-seq ii) TF dynamics In Brief

Integrative analysis of single-cell
transcriptomics and chromatin
expression accessibility provides insights into
regulatory features and dynamics in

single-cell RNA-seq
+ Myeloid pseudo-time human hematopoiesis.
iii) enhancer-gene correlation
AAA
AAA

Highlights
d Single-cell chromatin accessibility reveals a heterogeneous
hematopoietic landscape

d TF motif-associated chromatin variability in HSCs follows

erythroid/lymphoid paths

d Characterization of two GMP subsets with chromatin and

transcriptome differences

d Integrative analysis enables regulatory insights into cis- and

trans-acting factors

Buenrostro et al., 2018, Cell 173, 1535–1548

May 31, 2018 ª 2018 Elsevier Inc.
https://doi.org/10.1016/j.cell.2018.03.074
Resource

Integrated Single-Cell Analysis Maps

the Continuous Regulatory Landscape
of Human Hematopoietic Differentiation
Jason D. Buenrostro,1,2,* M. Ryan Corces,3 Caleb A. Lareau,1,11 Beijing Wu,4 Alicia N. Schep,4 Martin J. Aryee,1,10,11
Ravindra Majeti,5,6 Howard Y. Chang,3,4,7 and William J. Greenleaf3,4,8,9,12,*
1Broad Institute of MIT and Harvard, Cambridge, MA 02142, USA
2Harvard Society of Fellows, Harvard University, Cambridge, MA 02138, USA
3Center for Personal Dynamic Regulomes, Stanford University, Stanford, CA 94305, USA
4Department of Genetics, Stanford University School of Medicine, Stanford, CA 94305, USA
5Institute for Stem Cell Biology and Regenerative Medicine, Stanford University School of Medicine, Stanford, CA 94305, USA
6Division of Hematology, Department of Medicine, Stanford University School of Medicine, Stanford 94305, CA, USA
7Program in Epithelial Biology, Stanford University School of Medicine, Stanford, CA 94305, USA
8Department of Applied Physics, Stanford University, Stanford, CA 94025, USA
9Chan Zuckerberg Biohub, San Francisco, CA 94158, USA
10Department of Pathology, Massachusetts General Hospital & Harvard Medical School, Boston, MA 02115, USA
11Department of Biostatistics, Harvard T.H. Chan School of Public Health, Boston, MA 02115, USA
12Lead Contact

*Correspondence: jbuen@broadinstitute.org (J.D.B.), wjg@stanford.edu (W.J.G.)

https://doi.org/10.1016/j.cell.2018.03.074

SUMMARY ation as a ball rolling down a bifurcating three-dimensional

Human hematopoiesis involves cellular differentiation
of multipotent cells into progressively more lineage-
restricted states. While the chromatin accessibility
landscape of this process has been explored in
defined populations, single-cell regulatory variation
has been hidden by ensemble averaging. We collected
single-cell chromatin accessibility profiles across 10
populations of immunophenotypically defined human
hematopoietic cell types and constructed a chromatin
accessibility landscape of human hematopoiesis to
characterize differentiation trajectories. We find varia-
tion consistent with lineage bias toward different
developmental branches in multipotent cell types.
We observe heterogeneity within common myeloid
progenitors (CMPs) and granulocyte-macrophage
progenitors (GMPs) and develop a strategy to partition
GMPs along their differentiation trajectory. Further-
more, we integrated single-cell RNA sequencing
(scRNA-seq) data to associate transcription factors
to chromatin accessibility changes and regulatory ele-
ments to target genes through correlations of expres-
sion and regulatory element accessibility. Overall, this
work provides a framework for integrative exploration
of complex regulatory dynamics in a primary human
tissue at single-cell resolution.
INTRODUCTION
In 1957, Conrad Waddington developed an influential analogy for
developmental cell biology by conceptualizing cellular differenti-
surface (Goldberg et al., 2007; Waddington, 1957). This develop-
mental landscape defines a descriptive path a cell might follow,
choosing different developmental fates as it reaches saddle
points that separate different, increasingly restricted, cellular
states. The shape of this landscape is largely defined by tran-
scription factors (‘‘guy-wires’’), which recruit chromatin effectors
to reconfigure chromatin (Calo and Wysocka, 2013; Long et al.,
2016) and promote new cellular phenotypes (Graf and Enver,
2009; Takahashi and Yamanaka, 2006). These concepts—the
first a descriptive notion of development (Figure S1A), and the
second a mechanistic description of the molecular actors that
drive state changes (Figure S1B)—have provided a conceptual
framework for understanding cell fate choices. Recent techno-
logical advances in single-cell epigenomic assays (Kelsey
et al., 2017) now provide the opportunity to ascribe epigenomic
features to this landscape by quantifying overall epigenomic
similarity of individual cells during a normal differentiation pro-
cess, as well as the activity of master regulators that influence
cell fate decisions.
Hematopoietic differentiation serves as an ideal model for
exploring the nature of multipotent cell fate decisions (Laurenti
and Göttgens, 2018; Orkin and Zon, 2008). The hematopoietic
system is maintained by the activity of a small number of self-
renewing, long-lived hematopoietic stem cells (HSCs) capable
of giving rise to the majority of blood cell lineages (Becker et al.,
1963; Laurenti and Göttgens, 2018; Orkin and Zon, 2008) whereby
multipotent cells transit multiple decision points while becoming
increasingly lineage-restricted (Figure 1A). The human hemato-
poietic system is an extensively characterized adult stem cell
hierarchy with diverse cell types capable of phenotypic isolation
with multi-parameter fluorescence activated cell sorting (FACS)
(Corces et al., 2016; Laurenti and Göttgens, 2018). This capacity
for phenotypic isolation has enabled measurement of the

Cell 173, 1535–1548, May 31, 2018 ª 2018 Elsevier Inc. 1535
 A B CD34 CD45RA
5 5
10 10
HSC 4
4 10 CLP (8.2%)
10

CD10
CD38
3
3 10
10
CD34+ HSPCs

2
10 0
Differentiation

02 2
-10
MPP -10
3 4 5 3 4 5
0 10 10 10 0 10 10 10

5 pDC (15.8%)
LMPP CMP 10 10
5
HSC (10.8%)
4 4 CMP (24.3%)
10 LMPP (2.1%) 10

CD123
GMP

CD90
CLP pDC GMP MEP 3 3 (23.0%)
10 10

2 2
10 10
02 MPP (3.3%) Unknown
02
B, T Peripheral Monocyte, mDC Ery, -10 -10 MEP (7.1%) (5.3%)
3 4 5
NK cells pDC Granulocytes Mega 0 10 10 10 0 10
3
10
4
10
5

CD45RA CD45RA
C Single-cell ATAC-seq
E
100 Filter (n=1,038) Pass Filter (n=2,034)

Percent of fragments in peaks

80

Human Freeze High-throughput

Bone Marrow FACS & bank Cell capture Transpose PCR 60
Sequencing
Integrated Fluidics Circuit (IFC)

D Scale 100 kb hg19

40 1
chr4: 105,950,000 106,000,000 106,050,000
1_ TET2 0
CD34+ ATAC Density
0_ 20
1_ 1 2 3 4 5
CD34+ scATAC Number of fragments in peaks, log10

HSC (N=347) F
Obs
MPP (N=142) GATA1
5 Perm
LMPP (N=160)
Cell-cell variability

CMP (N=502) 4 CEBPB

BCL11A
GMP (N=216)
3 IRF8
MEP (N=138) STAT1
TCF4
CLP (N=78) 2 EBF1
pDC (N=141) HOXA4
Monocyte (N=64) 1
2

Unkown (N=60)
0 0 400 800 1200 1600
Fragments Rank Sorted TF motifs

Figure 1. Single-Cell ATAC-Seq Profiles Chromatin Accessibility within Single Hematopoietic Progenitors
(A) A schematic of human hematopoietic differentiation.
(B) Sorting strategy for CD34+ cells.
(C) Single-cell ATAC-seq workflow used in this study.
(D) Single-cell epigenomic profiles along the TET2 locus.
(E) Percent fragments in peaks by number of fragments in peaks, red lines show cutoffs used for determining which cells pass filter; points are colored by density.
(F) TF motif variability analysis in all single-cell epigenomic profiles collected for this study.
See also Figure S1 and Table S1.

epigenomic and transcriptional dynamics associated with sorted
human progenitors across differentiation providing a foundation
for the dissection of regulatory variation in normal multi-lineage
cellular differentiation (Chen et al., 2014; Corces et al., 2016; Farlik
et al., 2016; Novershtern et al., 2011). Furthermore, recent work
measuring single-cell transcriptomes has revealed significant
transcriptional heterogeneity in isolated progenitors (Notta et al.,
2016) and across differentiation (Laurenti and Göttgens, 2018;
Velten et al., 2017). These observations set the stage for single-
cell epigenomic measurements that may define cis- and trans-
regulatory mechanisms underlying transcriptional and cell fate
commitment heterogeneity in hematopoiesis.
To define a single-cell chromatin accessibility landscape of
this developmental hierarchy, we applied a single-cell assay
for transposase-accessible chromatin by sequencing (scATAC-
seq) (Buenrostro et al., 2015b) to 10 sortable populations in
human bone marrow or blood comprising multipotent and line-
age restricted progenitors. We find that the regulatory landscape

1536 Cell 173, 1535–1548, May 31, 2018

of human hematopoiesis is continuous, with cell surface markers
reflecting ‘‘basins’’ within this landscape. This single-cell anal-
ysis also uncovered substantial heterogeneity within immuno-
phenotypically defined cellular populations, including variability
within multipotent progenitors strongly correlated along the di-
mensions of hematopoietic differentiation—an observation
consistent with lineage priming at the level of chromatin acces-
sibility. We observe especially strong variability within popula-
tions of immunophenotypically defined common myeloid
progenitor (CMP) and granulocyte-macrophage progenitor
(GMP) cell types. Using ATAC-seq and RNA sequencing (RNA-
seq), we confirm that GMPs are substantially heterogeneous
on both epigenomic and transcriptomic levels and demonstrate
a strategy to enrich for sub-populations within GMPs at different
developmental stages of a myeloid differentiation trajectory.
Last, we generate scRNA-seq data and integrate these data
with scATAC-seq to associate expression changes of transcrip-
tion factors (TFs) to changes in chromatin accessibility at
cis-regulatory elements. Using these integrated data, we also
link changes at cis-regulatory elements to changes in the
expression of nearby genes. These methods for assaying and
analyzing single-cell epigenomics data provide the opportunity
for de novo discovery of cell types and states, define regulatory
variability within immunophenotypically pure populations, and
capture the cis- and trans-regulatory dynamics across a cell-
resolved regulatory landscape of differentiation.
RESULTS
Single-Cell Chromatin Accessibility of Distinct
Hematopoietic Cell Types
We used FACS to isolate 8 distinct cellular populations
from CD34+ human bone marrow, which included cell types
spanning the myeloid, erythroid, and lymphoid lineages
(Figures 1A and 1B). In addition, we also profiled a
CD34+CD38
CD45RA+CD123
subset that has not been well
characterized (Manz et al., 2002). Cells analyzed after sorting
and cells cryopreserved after sorting provided comparable
data quality and yield (Figures S1C–S1E), and therefore we per-
formed all further scATAC-seq measurements on cryopreserved
cells (Figure 1C). Together, this sorting strategy captures 97%
of all CD34+ cells (Figure S1F) and using post-sort analysis, we
found that sorted cell types were on average 97% pure by cell
surface marker immunophenotype (Corces et al., 2016). Using
this approach, we profiled the chromatin accessibility land-
scapes (CALs) across a total of 30 independent single-cell ex-
periments representing 6 human donors, with each progenitor
population assayed from two or more donors (Figure S1G). We
did not profile CD34
bone marrow stem cells, as they are rare
and less well described (Matsuoka et al., 2015).
Aggregated single-cell chromatin accessibility profiles closely
resemble bulk CD34+ ATAC-seq profiles (Figures 1D, S1H, and
S1I). Including previously published scATAC-seq data from
LMPPs and monocytes (Corces et al., 2016), this dataset
comprised 3,072 single-cell CALs across 32 integrated fluidic
circuits (IFCs). Single-cell profiles were of consistent high-quality
with 2,034 cells passing stringent quality filtering, yielding a
median of 8,268 fragments per cell with 76% of those fragments
mapping to peaks, resulting in a median of 6,442 fragments in
peaks per cell (Figure 1E; see STAR Methods).
TF Activity Inference Using ChromVAR
We applied ChromVAR to calculate TF motif-associated CAL
changes and identify potential regulators of epigenomic vari-
ability (Buenrostro et al., 2015b; Schep et al., 2017). This
approach quantifies accessibility variation across single-cells
by aggregating accessible regions containing a specific TF motif,
then compares the observed accessibility of all peaks containing
a TF motif to a background set of peaks normalizing for known
technical confounders. ChromVAR identifies high-variance TF
motifs across CALs representing known master regulators of he-
matopoiesis such as GATA1, BATF, and CEBPB (Figure 1F).
Notably, TFs of the same family often share a similar motif and
thus are difficult to disambiguate, therefore TF motifs highlighted
throughout the text are representative TF motifs that may
encompass the individual activities of multiple expressed TFs.
Hierarchical clustering of single-cell profiles using TF Z scores
generally classifies single-cells by their immunophenotypically
defined cell type identity (Figure 2A). Interestingly, despite the
overall high quality of the HSC profiles (Figures S1J–S1L),
HSCs exhibit low TF Z scores for lineage specifying TF motifs.
Furthermore, the HOX TF motif was most enriched in HSCs, pre-
viously shown to regulate stem cell activity (Lawrence et al.,
1997; Magnusson et al., 2007), however, this TF motif also ex-
hibited relatively low-level activity compared to lineage defining
TFs in more-differentiated cells. Low level expression of lineage
specifying TFs in other multipotent cell systems has been
described (Grün et al., 2016) and is hypothesized to generally
promote multipotency (Graf and Enver, 2009), which may also
explain the low level TF Z scores in this analysis of HSCs.
Using the vector of TF Z scores as features, we visualize he-
matopoietic differentiation within these data using t-SNE, which
clearly displays the expected branching into four distinct differ-
entiated final states representing erythroid, myeloid, lymphoid,
and pDC differentiation (Figures 2B and S2A–S2D). In past
work, we also used chromatin immunoprecipitation sequencing
(ChIP-seq) data as annotations to explore chromatin accessi-
bility differences in single-cell profiles (Buenrostro et al.,
2015b). Here, we found that ChIP-seq data from K562 cells, a
cell line described as an erythroid progenitor model, discrimi-
nated between cells at different stages of erythroid differentia-
tion, however, failed to capture the variance associated with
myeloid and lymphoid trajectories (Figure S2E). Therefore, due
to the relatively paucity of TF ChIP-seq in primary bone
marrow-derived CD34+ cells or in early myeloid and lymphoid
cell models, we chose to use TF motifs in downstream analyses.
Mapping Single Profiles on Hematopoietic Principal
Components
The TF Z scores can be used to cluster single-cell profiles,
however, this unsupervised analysis may not distinguish chro-
matin accessibility changes associated with differentiation
from changes associated with other biological phenomenon
such as the cell cycle or niche-dependent cell-cell signaling
(Crane et al., 2017). Furthermore, the use of t-SNE and TF
Z scores for clustering makes relative cell-cell distances difficult
Cell 173, 1535–1548, May 31, 2018 1537
 A B
40
KLF2
CTCF
GATA3 30
GATA1
HOXA4
Lymphoid
MAFF 20
ERG
ETV4
RARA
ATF4 10
CEBPB
PAX2

Dim. 2
EBF1 0
SPIB
STAT1
IRF8
SPI1 -10
BCL11A
RUNX2
MEIS3 -20
-30
Erythroid
ID3
TCF12 -40 Myeloid
-50
HSC LMPP GMP CLP Unknown -60 -40 -20 0 20 40 60
MPP CMP MEP pDC Z-score -5 5
Mono Dim. 1
Z-score Z-score
C E -11.7 11.2 G -6.1 6.4
GATA1
GA
ATA1
A A1
1 motif
o IID3 motif
t

0 0 0
PC3 (10.42%)

PC3 (10.42%)

PC3 (10.42%)
-2 -2 -2

-4 -4 -4
%)

%)

%)
15 15 15
20
(45.5

20

(45.5
20

(45.5
-6 25 -6 25 -6 25
30 30 30
PC1

35

PC1
35

PC1
-12 -10 35
-8 -6 -4 -2 0 2 -12 -10 -8 -6 -4 -2 0 2 -12 -10 -8 -6 -4 -2 0 2
PC2 (37.8%) PC2 (37.8%) PC2 (37.8%)
Density Z-score Z-score
D 0 1 F -8.4 7.9 H -2.9 3.2
CEBPB
CE
EBPB
E PB
Bmmotif HOXA9
HO
OX
OXXA9
9 motif
o
Lymphoid
Lym
L mpho
m ho
oid
id
d

0 0 0
PC3 (10.42%)

PC3 (10.42%)

PC3 (10.42%)

-2 -2 -2

-4 Erythroid
Eryythro
ro
oiid
d M l
Myeloid -4 -4
%)

%)

%)
15 15 15
20
(45.5

20
(45.5

20

(45.5
-6 25 -6 25 -6 25
30 30 30
PC1

35
PC1

35

PC1
35
-12 -10 -8 -6 -4 -2 0 2 -12 -10 -8 -6 -4 -2 0 2 -12 -10 -8 -6 -4 -2 0 2
PC2 (37.8%) PC2 (37.8%) PC2 (37.8%)

Figure 2. Lineage Projection of Human Hematopoietic Progenitors

(A) Top: hierarchical clustering of single-cell epigenomic profiles (columns) and TF motif accessibility Z scores (rows). Bottom: single-cell profiles colored by their
sorted immunophenotype identity.
(B) t-SNE of TF Z scores shown in (A), cells are colored by their sorted immunophenotype identify.
(C and D) Single-cell epigenomic landscape defined by PCA projection (see STAR Methods) colored by (C) cell type identity using immunophenotype and (D)
density (see STAR Methods) overlaid with nominal trajectories expected from the literature, as shown in Figure 1A.
(E–H) PC projection colored by (E) GATA, (F) CEBPB, (G) ID3, and (H) HOXA9 TF motif accessibility Z scores.
See also Figure S2, Table S2, and Data S1 and S2.

to interpret. We reasoned a reference guided approach that uti-
lizes accessibility co-variance of regulatory elements in bulk
hematopoietic samples, combining previously published (Cor-
ces et al., 2016) and additional reference profiles (see STAR
Methods), would provide a natural and intuitive subspace for
dimensionality reduction of single-cell data. To achieve this,
we implemented a computational strategy, similar to recent
methods for single-cell RNA-seq analysis (Li et al., 2017), that
first identifies principal components (PCs) of variation in bulk
ATAC-seq samples (Figure S2F) (Corces et al., 2016), then
scores each single-cell by the contribution of each PC. Cells
are subsequently clustered using the Pearson correlation coeffi-
cients between these normalized PCs scores and all other cells
(Figure S2G). For low-dimensional visual representation, we per-
formed PCA on this correlation matrix (Figures 2C and 2D) and
display the first 3 principal components, which represent
93.7% of the total subspace variance (Figures S2H–S2M).
We validated this computational approach by down sampling
bulk profiles to 104 fragments and find that the PCA-projection
approach closely follows sample clustering using the bulk

1538 Cell 173, 1535–1548, May 31, 2018

dataset (Figure S2N). We next down-sampled ensemble single-
cell profiles to match the sequence depths observed in single-
cells to quantify the expected mean error to be 1.95%, 1.70%,
and 3.1% per cell of the total signal for PCs 1–3, respectively
(Figures S2O and S2P). To test our sensitivity for identifying
intermediate cell states, we created synthetic mixtures from
ensemble profiles, down-sampled to 104 fragments and found
the synthetic mixtures to closely follow the expected paths (Fig-
ures S2Q and S2R). Last, we found this approach to be robust to
expected experimental confounders in single-cell data (Figures
S2S–S2V). Overall, this visual representation of the data provides
a reference-guided landscape of differentiation with similarities
to Waddington’s developmental landscape (Figure S1A), further-
more layering TF Z scores onto this representation provides
insight into the ‘‘guy wires’’ that may underlie epigenomic
changes during differentiation (Figure S1B).
The Continuous Landscape of Human Hematopoiesis
Using this computational approach, we find the CAL of human
hematopoiesis radiates away from a common basin of early he-
matopoietic progenitors (Figure 2C). HSC/MPP (left) and LMPP
(right) localize at the center of the projection, followed by CMP
and GMP cells that comprise a large and diverse basin. Differen-
tiation into CLP (lymphoid), MEP (erythroid), and monocytes
(myeloid) appear as distinct differentiation trajectories that
swoop away from the central HSC basin (Figure 2D). Further-
more, motifs associated with master lineage regulators ID3,
CEBPB, and GATA1 (Orkin and Zon, 2008) show continuous
gradients of activity across lymphoid, myeloid, and erythroid
development, while the HSC and LMPP compartment show
higher accessibility associated with the HOX motif (Figures 2E–
2H). We also observe examples of FACS misclassification of
cell types, particularly between CMP:MPP, GMP:LMPP (sepa-
rated by CD38), and GMP:CLP (separated by CD10), likely due
to the continuous nature of the cell surface markers (Figures
S3A–S3D).
CMP, GMP, and MEP profiles appear markedly heteroge-
neous in this projected space. We quantified the statistical
significance of this observed heterogeneity by comparing
the observed variability to down-sampled aggregate profiles
and found CMPs to be the most heterogeneous cell type (p <
10
111), with all cell types displaying statistically significant
heterogeneity (Figure S3E). To further assess the statistical sig-
nificance of this heterogeneity, we permuted peaks matched in
mean accessibility and GC content (see STAR Methods) and
find that variability across all cell types, with the exception of
monocytes (p = 0.35), remained statistically significant (Figures
S3F and S3G). Finally, single-cell TF Z scores (Schep et al.,
2017), which are calculated without using bulk ATAC-seq data
as a reference, also exhibit significant variability for all cell types
(Figure S3H). Thus, rather than identifying a series of discrete
cellular states (Figure 1A), these results suggest that the CAL
in early hematopoietic differentiation (HSC, MPP, LMPP, and
CMP) comprise a fairly broad basin of allowable states, while
paths of later differentiation become more canalized into distinct
and continuous differentiation trajectories (see Data S1; single-
cell CALs can be further explored using our web resource:
http://schemer.buenrostrolab.com/).
De Novo Identification of Uncharacterized Chromatin
States
Given the observed limitations of our sort markers, we sought to
define hematopoietic cell types de novo by applying k-medoids
clustering on the first five principal components from this PC
projection approach. We defined 14 unique clusters (Figures
3A and 3B; see STAR Methods) that largely overlap with
previously defined cell surface marker-based definitions of hu-
man hematopoietic subsets (Figure 3C) and changes in the
accessibility of TF motifs associated with hematopoiesis (Fig-
ure 3D). This analysis identified key hematopoietic regulators
de novo, including motifs associated with well-described master
regulators GATA1 (erythroid), CEBPD (myeloid), and EBF1
(lymphoid) lineage-specifying factors (Orkin and Zon, 2008).
Notably, we also find a specific HSC cluster of TFs that include
HOX, ERG, and MAFF motifs.
Using this clustering approach, we find that CMPs separate
into 4 clusters denoted here as clusters 2–5, which includes a
cluster with mixed contribution from CMP and MEP cells (cluster
5). We observe that the 4 CMP clusters within our data show
significant variability across motifs associated with GATA1,
BCL11A, and SPI1 (PU.1), TFs implicated in myeloid/erythroid
specification (Figure S3I). We identify 1,801 differentially acces-
sible regions across these CMP clusters, including two previ-
ously validated erythroid enhancers (Fulco et al., 2016)
regulating GATA1 expression (Figures S3I and S3J). Given these
differences, we assigned CMP clusters as CMP-K3 (early
erythroid), CMP-K5 (late erythroid), CMP-K4 (unknown), and
CMP-K2 (myeloid primed). These strong chromatin accessibility
differences further validate recent work describing functional
and transcriptional heterogeneity within mouse (Paul et al.,
2015; Perié et al., 2015) and human (Notta et al., 2016) CMPs
and strongly suggest that CMPs can be partitioned into myeloid
and erythroid committed progenitors. In addition, we also find
that MEP (K5–K7), GMP (K9,K10), and pDCs (K12,K13) predom-
inantly separate as two or more distinct clusters each, likely
representing early- and late-stage progenitor differentiation.
Chromatin Accessibility Variability within Data Driven
Clusters
We next sought to measure chromatin accessibility differences
within stringently defined HSC and LMPP progenitors, popula-
tions previously described as primed toward different lineage
fates (Busch et al., 2015; Karamitros et al., 2018; Laurenti and
Göttgens, 2018; Naik et al., 2013; Pei et al., 2017). To quantify
this variability, we created stringent cluster definitions that
required cells to be both CAL cluster-pure and immunopheno-
typically (FACS identity) pure populations, which we call
epigenomically and immunophenotypically pure (EIPP) clusters.
We then computed TF Z scores (Schep et al., 2017) for cells
within each EIPP group and found substantial heterogeneity
within these subsets (Figure S3K). To explore the relationship
of TF-associated variability within HSCs to directions of differen-
tiation, we categorized individual EIPP HSCs by their TF Z scores
(high or low), computed the distance between high/low centroids
in the PC space, and calculated statistical significance by
comparing high/low distances to permuted HSC EIPP profiles
(Figures 3E and S3L–S3N). Using this analysis approach, we
Cell 173, 1535–1548, May 31, 2018 1539
 Freq.
A B C 0 1
CLP
CLP 14
13
HSC
pD
pDC
pDC 12
11 MPP
10 K14
K1
14
1 4
M P
MEP HSC
C 9 CMP
8 K13
0 7 0 K
K7 MEP
6 K1
K
K8
K
PC3 (10.42%)

PC3 (10.42%)
5
4 K6
K K12 LMPP
-2
3
-2 K2
K2
2 K5 GMP
-4 1
-4 K3 K9
K 9 Mono

%)

%)
CMP 15 K4
K 15
20 K10
K 0
(45.5
20

(45.5
-6 GMP 25 -6 25 pDC
30 P C1 K11
K1
11
1 30
35

P C1
35 CLP
-12 -10 -8 -6 -4 -2 0 2 -12 -10 -8 -6 -4 -2 0 2
PC2 (37.8%) PC2 (37.8%) 1 2 3 4 5 6 7 8 9 10 11 12 13 14
TF Z-score Clusters Z-score
D 0 Max E F -3.8 3.4
PAX5 2.2 RELA motif
EBF1 GATA motif 2
RUNX2 CTCF ID motif
CTCF motif
NKFB motif 0

Variability, TF z-score
1.8 CTCFL ETS motif
RELA
TCF12
LYL1 NFKB1
FLI1 -2

PC3
GATA2
TCF4 1.4 ETV4 MESP2 GATA1
ID3 MESP1 GATA3 GATA5
MEIS3 ID4 -4
TBX15
IRF1
PRDM1 1
STAT1 -6
MEF2A variability=1.80
BCL11A
KLF4 direction p-value=0.28 Low High
KLF14 0.6
0 5 10 15 20 25 -12 -8 -4 0 4
ESRRA Direction score, -log10 p-value PC2
RARA
RELA Z-score
STAT3 G Z-score H -2.7 2.7
MAFF -3.1 2.6
2 GATA2 motif 2 MESP1 motif
BACH2
ATF4
CEBPD
JUND 0 0
SPI1
ATF3
-2 -2
ERG
HOXA9
CTCF
HOXA4 -4 -4
ALX4
GATA1
FOXO4 -6 variability=1.43 -6 variability=1.36
direction p-value=10-17 Low High direction p-value=10-8 Low High
FOXP3
-12 -8 -4 0 4 -12 -8 -4 0 4
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Clusters

Figure 3. Molecular Characterization of Data-Defined Clusters

(A) Single-cell epigenomic landscape defined by PCA projection, colored by data-driven cluster number.
(B) Medoids of data-driven centroids depicted on the PCA sub-space.
(C) Confusion matrix of data-driven clusters representing the percent frequency of immunophenotypically defined cell types.
(D) TF motif accessibility Z scores averaged across data defined clusters and hierarchically clustered. Scores are normalized by the max value of each TF motif.
(E) TF motif variability and direction –log10 p value for each TF motif for the HSC EIPP cluster, TFs sharing a similar motif are highlighted.
(F–H) TF motif accessibility Z scores of HSC profiles for (F) RELA, (G) GATA2, and (H) MESP1 motifs, arrows denote the direction of the signal bias and are colored
by the target cell type.
See also Figure S3.

find CTCF, nuclear factor kB (NF-kB) (represented by the RELA this TF bias is consistent with previous studies that lineage trace
motif), and ETS motifs to be significantly variable in HSCs, HSCs that suggest that HSCs exhibit oligo-lineage bias toward
however, uncorrelated with any specific direction of differentia- erythroid-myeloid or lymphoid cell fates (Pei et al., 2017).
tion (Figure 3F). Interestingly, NF-kB signaling (inflammatory We next turned to LMPPs, a subset previously shown to
signaling) has been implicated in mouse HSC stem-cell mainte- demonstrate lineage priming toward dendritic (pDC), myeloid
nance (Zhao et al., 2012) and HSC emergence mediated by (GMP:monocytes), and lymphoid (CLP:B cell) fates in mice
neutrophil secretion of tumor necrosis factor alpha (TNF-a) (Busch et al., 2015; Naik et al., 2013) and in human (Karamitros
(Espı́n-Palazón et al., 2014; Sawamiphak et al., 2014). In et al., 2018). Interestingly, we found motifs associated with the
contrast, we find the GATA (p = 10
17) and MESP/ID (p = 10
8) TFs TCF4, STAT1, and CEBPE to be significantly correlated
motifs (represented by GATA2 and MESP1 motifs) TF Z scores with directionality toward CLP, pDC, and GMP differentiation,
to be significantly correlated to erythroid and lymphoid trajec- respectively (Figures S3O–S3R), notably the TCF4 motif is similar
tories respectively (Figures 3G, 3H, and S3N). The direction of to the TCF3, ID3, and MESP1 motifs. TCF4 and CEBPE motif

1540 Cell 173, 1535–1548, May 31, 2018

A B G

C D H

E F

Figure 4. Identifying Continuous Differentiation Trajectories

(A–D) PC2 by PC3 projection of single-cells highlighting cells progressing through the inferred (A) erythroid, (B) lymphoid, (C) pDC, and (D) myeloid developmental
trajectory (black line), cells used for inference are colored by sorted identity, all other cells are shown in gray.
(E) Sorting schema for different GMP progenitors defined by CD123 expression, marked by CD123 low (GMP-A, light-gray), CD123 medium (GMP-B, gray), and
CD123 high (GMP-C, dark-gray).
(F) Bulk RNA-seq log2-fold-change and -(log p value) for expressed genes comparing GMP-C and GMP-A.
(G) Single-cells used for the myeloid trajectory colored by (left) their cluster identity (cluster colors as in Figure 3) or (right) their density along the trajectory.
(H) Density of myeloid progression scores for immunophenotypically defined cell types, including the GMP subsets.
See also Figure S4.

accessibility were anti-correlated with each other, and each
defined a unique direction toward CLP (lymphoid) or GMP
(myeloid) differentiation, respectively, suggesting antagonism
between myeloid/lymphoid differentiation programs. Sepa-
rately, the STAT1 motif accessibility appeared to be directed
toward CLP (lymphoid) and pDC (dendritic) cell fates. Overall,
this reference-guided computational approach provides a
statistical framework for assigning TF motif-associated vari-
ability to lineage-associated CAL variation providing a resource
for identifying molecular factors that may be involved in lineage
priming across multipotent cell populations.
Heterogeneous Cell Types Can Be Further Divided along
Developmental Trajectories
We next sought to order cells along continuous differentiation
trajectories across branches of hematopoietic development.
To achieve this, we first determined the shortest path between
cluster centroids and assigned cells to the closest point along
that path; a similar approach has been described for analyzing
scRNA-seq data (Shin et al., 2015). This approach aligned
cells to well-defined lineage pathways (Orkin and Zon, 2008)
producing an ordering of single cells along continuous
erythroid (K1,K3,K5,K6,K7), lymphoid (K1,K2,K8,K14), pDC
(K1,K2,K8,K12,K13), and myeloid (K1,K2,K9,K10,K11) differenti-
ation trajectories (Figures 4A–4D and S4A–S4D). These
trajectories allow for interpretation of CAL heterogeneity within
progenitors and provide methods to further parse cellular sub
states across differentiation.
We examined variability within two clusters (K9 and K10) of
GMPs, which show significant differences in accessibility among
myeloid-defining factors SPI1 (PU.1) and CEBP-associated
motifs across the myeloid developmental trajectory (Figure S4E).

Cell 173, 1535–1548, May 31, 2018 1541

To further partition this population, we sought to identify cell sur-
face markers that may differentially enrich for K9 and K10 GMPs.
We hypothesized that CD123 expression may correlate with
early and late GMP differentiation for two reasons: (1) the UNK
population, which is CD123
/lo, is enriched in the GMP K9
cluster, and (2) CD123, also known as IL3RA, is a high-affinity
receptor for the myeloid promoting cytokine IL3. We therefore
performed scATAC-seq, bulk ATAC-seq (Buenrostro et al.,
2013, 2015a), and bulk RNA-seq on cells from three distinct
bins of CD123 expression (Figure 4E). Bulk ATAC-seq and
RNA-seq data revealed substantial chromatin accessibility and
transcriptomic differences across the GMP-A and GMP-C
populations (Figures 4F and S4F–S4H). The list of differentially
expressed genes included important developmental regulators,
including downregulation of HSPC TFs GATA2 and TAL1 and
upregulation of myeloid genes SPIB, IRF8, TLR7, and MPEG1
in the GMP-C cell population (Figure 4F). In addition, projection
of the scATAC-seq data from the three cell fractions revealed
that this strategy provides strong separation of early (GMP-A)
and late (GMP-C) stages of myeloid differentiation (Figures 4G,
4H, and S4I–S4K). Altogether, we validate the heterogeneity
within GMPs and more generally demonstrate a data driven
approach for defining cell populations from single-cell epige-
nomic data.
Motif Accessibility Dynamics along Myeloid Cell
Differentiation
The myeloid trajectory described above transits two heteroge-
neous cell populations (CMP and GMP), as such, regulatory
analysis of myeloid differentiation has been previously obscured
in bulk studies due to the limitations of the immunophenotypic
markers of these populations. We therefore sought to charac-
terize TF dynamics across myeloid development by mapping
TF Z scores to cells along the continuous myeloid differentiation
trajectory (Figures S4L–S4N). Using this approach, we find
6 clusters (see STAR Methods) of TF Z score profiles during
myeloid development (Figures 5A and S5A–S5C). Accessibility
at TF motifs associated with regulators HOXB8 and GATA1 (clus-
ter 1) is high in HSCs and decreases through differentiation to
CMPs. Interestingly, loss of GATA motif accessibility (repre-
sented by the GATA1 motif) begins within the HSC compartment,
while HOX motif accessibility (represented by the HOXB8 motif)
is lost at the transition of HSC to CMP differentiation, suggesting
that loss of GATA motif accessibility may be an early event in
lineage commitment within HSCs (Figure 5B). We also observe
two distinct modes of activation for myeloid-associated TF
motifs; cluster 4 TFs (CEBPD- and SPIB-associated motifs)
display early and gradual gain in activity beginning within
CMPs, while cluster 5 TFs (STAT1-, IRF8-, and BCL11A-associ-
ated motifs) increase sharply in activity across the GMP-A to
GMP-C transition, implicating the CEBP family of TFs (repre-
sented by the CEBPD motif) as an initiating factor for myeloid-
erythroid specification (Figure 5C). In addition to the activity
patterns associated with canonical myeloid-defining factors,
we also identify a pulse of activity within CMPs from cluster
2 TFs (TCF3/12 associated TF motifs upregulated in CLP/
pDC), which may reflect transient activation of a lymphoid pro-
gram within pre-committed myeloid-biased CMPs (Figure S5C).
1542 Cell 173, 1535–1548, May 31, 2018
Matching Transcriptomes with Chromatin Accessibility
We next aimed to develop a means to pair single-cell epigenomic
and transcriptomic measurements, with the goal of linking chro-
matin accessibility changes associated with DNA sequence
motifs to expressed TFs, as well as linking accessibility changes
at putative enhancers to expression changes at target genes. We
first performed single-cell RNA-seq (10X genomics platform)
across HSC, CMP, and GMPs, collecting a total of 7,818 cells
passing filter (2,268, 4,454, and 1,096, respectively; Figure 5D).
In addition, we included publically available scRNA-seq data
from CD34+ and CD14+ monocyte cells (Zheng et al., 2017),
altogether analyzing transcriptional dynamics of 14,432 cells
across myeloid differentiation. Using these data, we developed
a reference-guided approach to pair scATAC-seq and scRNA-
seq profiles (see Data S2). To do this, we first fit a linear model
to match the measured bulk ATAC-seq PCs, which measure
global variation in chromatin accessibility, to changes in gene
expression as measured by bulk RNA-seq across sorted popu-
lations (Figures S5D–S5F). We then used this map between
ATAC-seq PCs and gene expression to assign ‘‘inferred tran-
scriptomes’’ to each cell in the scATAC-seq dataset (Figure S5G).
Finally, to pair each scRNA-seq profile to a scATAC-seq cell, we
assigned scRNA-seq profiles to the most correlated scATAC-
seq ‘‘inferred transcriptome’’ (Figure S5H). Using this approach,
we found that the sorted identity of scRNA-seq profiles were
enriched for the corresponding matched sorted identity for
scATAC-seq profiles (Figure S5I). Furthermore, by pairing
single-cell RNA-seq to scATAC-seq cells, we found scRNA-
seq profiles of FACS-sorted CMPs associated with the four
scATAC-seq-defined CMP clusters discussed above (Fig-
ure S5J). Further validating the recent reports of heterogeneity
in mouse (Paul et al., 2015; Perié et al., 2015) and human (Notta
et al., 2016) CMPs, we found expression heterogeneity of known
hematopoietic regulators in CMPs, which included the TFs
HOXA5, GATA1, and CEBPB (Figures S5K and S5L).
Our approach provides a computational method to fit gene
expression changes across bulk ATAC-seq and RNA-seq ‘‘an-
chor points’’ generated from well-defined sorted populations,
providing a reference for analysis of single-cell gene expression
and chromatin changes spanning these anchor points to resolve
continuous regulatory changes in cell differentiation. This refer-
ence-guided strategy resulted in a total of 9,312 scRNA-seq cells
positioned across myeloid pseudo-time with high concordance
in the enrichment of immunophenotypically defined cells across
the trajectories (Figure 5E). Using this unified lineage order, we
mapped expression dynamics across myeloid cell differentiation
and found expected patterns across known regulators of myelo-
poesis (Figures 5F and 5G). To further validate this pairing
approach, we compared the ATAC:RNA paired lineage order
with ordering scRNA-seq cells using diffusion pseudotime
(DPT) (Haghverdi et al., 2016). In this comparison, we find that
the two approaches for cell ordering are overall highly correlated
(R = 0.86; Figure S5M). However, we find that unsupervised
ordering of HSCs using DPT was more correlated to the number
of genes detected than the ATAC:RNA pairing approach
described above (R = 0.68 versus R = 0.14), suggesting
computational ordering of scRNA-seq data with DPT may be
more sensitive to dropout (Figures S5N and S5O). This may be
A B HSC CMP GMP Mono C HSC CMP GMP Mono

HSC CMP GMP Mono TF Z-score

0 Max 1 1
36 motifs

0.8 0.8
HOXA9

Deviation Score
Deviation Score
k=1

GATA1
HOXB8 0.6 0.6

0.4
52 motifs

CTCF 0.4
k=2

TCF3 0.2
TCF12 HOXB8, K1 0.2 CEBPD, K4
GATA1, K1 BCL11A, K5
0
25 motifs

95% CI 0 95% CI
k=3

BPTF -2 0 2 4 6 8 10 12 -2 0 2 4 6 8 10 12
Myeloid pseudo-time Myeloid pseudo-time
FOXJ3
Rel. Density
D E
35 motifs

CEBPG 10x scRNA-seq 0 1

k=4

SPIB 60 GMP
HSC HSC
DBP Mono

scATAC
CMP
CEBPA 50
CMP
STAT1 40
20 motifs

GMP
IRF8
k=5

BCL11A 30 mono
Dim. 2

IRF4
20
HSC
37 motifs

scRNA
ESRRA 10 CMP
k=6

RORB
KLF2 0 GMP
-2 0 2 4 6 8 10 12 14 -10 mono
Myeloid pseudo-time -1 0 1 2 3 4 5 6 7 8 9 10 11 12
-40 -30 -20 -10 0 10 20 30 40 50
Dim. 1 Myeloid pseudo-time
F 5 CEBPD expression H
log2 expression

RNA expression of TF Motif accessibility of TF Pearson

HOXB7
0
HOXB6
2 NFIB
1 GATA3
0 MECOM
-2 0 2 4 6 8 10 12
GATA2
G 5
GATA2 expression
SPI1
log2 expression

CEBPD

0 SPIB

1
Pearson
1 IRF8

IRF2

0.5
0 -2 0 2 4 6 8 10 12 -2 0 2 4 6 8 10 12
-2 0 2 4 6 8 10 12
Myeloid pseudo-time Myeloid pseudo-time
Myeloid pseudo-time

Figure 5. Transcription Factor Dynamics across Myeloid Differentiation

(A) K-medoids clustering of TF motif accessibility (left) and PWM logos (right) for dynamic TF motif profiles across myeloid development.
(B and C) Smoothed profiles of TF motif accessibility Z scores in myeloid progression for (B) HSC active TFs GATA1 (blue) and HOXB8 (green), and (C) monocyte
active regulators CEBPD (yellow) and BCL11A (red). Error bars (gray) denote 95% confidence intervals.
(D) t-SNE of scRNA-seq data showing HSC, CMP, GMP, and monocyte cells.
(E) Density of myeloid pseudo-time scores for (top) scATAC-seq and (bottom) computationally matched scRNA-seq profiles (see STAR Methods).
(F and G) Log2 mean expression profiles for TFs (F) CEBPD and (G) GATA2 across myeloid pseudo-time, (top) individual cells are colored by their sorted identity,
CD34+ cells are shown in black and (bottom) smoothed profiles are shown in red.
(H) Left: expression and (right) TF motif accessibility dynamics across myeloid pseudo-time for correlated (R > 0.5) gene-motif pairs.
See also Figure S5.

expected as DPT does not explicitly model cell-cell differences Linking TF Expression with Associated Accessibility
in dropout (zero counts for genes) and further suggests that Variation in Binding Motif
computational tools for joint analysis of scATAC-seq and In effort to disambiguate TFs that bind the same or similar motif
scRNA-seq may be more robust to technical confounders. and thus assign expression of TFs to downstream changes
Most importantly, the gene expression trajectories (Figure S5P) in chromatin accessibility at TF motifs, we correlated the
are largely similar between the two approaches, supporting our expression of TFs with the TF motif Z scores across myeloid
ATAC:RNA pairing approach. pseudo-time. We then filtered for motif accessibility-TF

Cell 173, 1535–1548, May 31, 2018 1543

expression correlations of R > 0.5, this approach yielded 11 TFs
that defined different stages of myeloid development (Figure 5H),
including loss in the expression of HOX factors (HOXB7 and
HOXB8) (Argiropoulos and Humphries, 2007) and activation of
well-known master regulators of myeloid cell development
including SPI1 (PU.1) and IRF8 (Satpathy et al., 2012). Resolution
of the developmental order of these activated TFs across
myeloid differentiation has been previously obscured in bulk
studies, in part, due to the cellular heterogeneity within CMPs
and GMPs. Interestingly, we also observed a strong correlation
between GATA3 expression and GATA motif accessibility, dele-
tion of GATA3 has been shown to promote self-renewal in HSCs
(Frelin et al., 2013), together leading to the hypothesis that
GATA3 may be associated with HSC lineage priming. Here,
single-cell chromatin accessibility, paired with single-cell tran-
scriptomics, resolves the temporal dynamics of master regulator
expression and associated chromatin changes in myeloid cell
development, providing a resource for further functional studies
and for the analysis of regulatory changes associated with
differentiation.
Regulatory Element and Gene Activation across
Myelopoiesis
We next sought to characterize locus-specific cis-regulatory
dynamics during myeloid differentiation. We first filtered for reg-
ulatory elements with high fragment counts and with significant
variability across the ordered cells identifying 14,005 cis-regula-
tory elements for analysis (see STAR Methods). These regulatory
elements exhibited highly heterogeneous patterns of accessi-
bility changes (Figures 6A–6C and S6A–S6C)—suggesting that
a limited number of TF motif accessibility patterns (k = 6) could
induce a surprising level of variation of chromatin accessibility
at individual regulatory elements. For example, within the regula-
tory elements surrounding the myeloid regulator CEBPD
(numbered for simplicity, see Figure 6B), the distal element
CEBPD-1 was ‘‘fast-to-activate’’ and showed stepwise gains
of activity while the distal element CEBPD-2 was ‘‘slow-to-acti-
vate’’ and showed a more discrete pulse of activity (Figure 6B).
To visualize the complete repertoire of dynamic regulatory
profiles, we ordered elements based on their accessibility
changes over this trajectory (Figures 6C and S6). This analysis
reveals multiple broad classes of regulatory element behaviors,
ranging from fast- to slow-to-repress HSC regulatory elements
and fast- to slow-to-activate monocyte regulatory elements (Fig-
ure 6C). We also observe a collection of ‘‘transition’’ cis-regula-
tory elements that exhibit peak accessibility at intermediate
stages of myeloid development, as well as ‘‘reactivation’’ ele-
ments that are initially lost and subsequently reactivated in later
stages of myeloid differentiation (Figures S6D and S6E). Thus,
from a small number of discrete clusters of TF motif accessibility,
highly diverse cis-regulatory profiles likely arise from the combi-
natorial control of trans-factor binding to their target regulatory
elements (Figure S6F).
We reasoned that correlation between dynamically activated
patterns of distal regulatory elements with nearby expressed
genes may be used to connect enhancers to target genes
(Figure 6C). Indeed, we found dynamic regulatory elements sur-
rounding CEBPD were highly correlated with CEBPD expression
1544 Cell 173, 1535–1548, May 31, 2018
(Figure 6D). More generally, we calculated the correlation of reg-
ulatory elements to dynamic genes within 10 Mb of annotated
transcription start sites (‘‘peak-gene pairs’’) and found that
proximal regulatory elements (<100 kb) were significantly more
correlated to the expression of nearby genes than distant ele-
ments (>100 kb) (Figure 6E). Further validating this approach,
we also found that the correlation of regulatory elements to
target genes improved as a function of loop confidence within
promoter capture HiC (PCHiC) data (Figure 6F), here defined
by PCHiC loops from both CD34+ (Mifsud et al., 2015) and mono-
cyte (Javierre et al., 2016) cells. Importantly, we find loop interac-
tions at this resolution do not necessarily define correlated
peak-gene pairs. In our analysis, only 45% of dynamic en-
hancers within high confidence loops are called as correlated
to the expression of PCHiC defined target genes (Figure S6G).
This observation may be due to the fact that PCHiC loops link
relatively large genomic regions, often encompassing multiple
regulatory elements, which may independently regulate down-
stream genes.
We next sought to test whether previously defined cis-linked
expression quantitative trait loci (cis-eQTLs) overlapped
enhancer-gene interactions identified using these integrated sin-
gle-cell data. We reasoned that correlated peak-gene pairs
could be used to functionally connect relevant genetic variation
at regulatory elements to consequences in gene expression
important in normal monocyte function. We therefore collected
previously published cis-eQTLs, derived from interferon-g and
lipopolysaccharide stimulation of monocytes (Fairfax et al.,
2014), and filtered for SNPs within developmentally dynamic reg-
ulatory elements linked to dynamic genes (n = 370 peak-gene
pairs). To directly compare enrichment of either correlated
peak-gene pairs or PCHiC loops at cis-eQTL defined peak-
gene interactions, we determined significant enrichment of
each dataset by normalizing to a background set of peak-genes
matched for distance (Figure S6H). We found that cis-eQTLs
were strongly enriched for scATAC/scRNA-seq correlated
peak-gene pairs (p = 4.9 3 10
5) and observed only a modest
enrichment PCHiC loop interactions (p = 0.19) (Figures 6G and
S6I). Thus, statistical linkage between single-cell chromatin
accessibility and gene expression can serve as a means to
functionally link enhancers to target gene promoters.
DISCUSSION
We used single-cell chromatin accessibility and transcriptomic
analysis to identify regulatory heterogeneity and continuous
differentiation trajectories in early human hematopoiesis by
developing a broadly applicable computational framework for
analysis of these single-cell data. This framework includes a
means for visualizing single-cell chromatin accessibility, and
computationally pairing these data with single-cell RNA-seq,
by using bulk data as a reference. With this approach, we find
that immunophenotypically defined cell populations often flow
from one state to another and further we dissociate TF motif
activity variability within these populations as correlated or un-
correlated to axis of differentiation. In this effort, we find the
activity of TF motifs, such as the GATA motif in HSCs, may repre-
sent indicators of lineage priming pulling cells toward different
A B C scATAC-seq scRNA-seq Peak activity
CEBPD:promoter CEBPD-2:distal
CEBPD-1:distal CEBPD-3:intronic 0 Max 0 Max HSC mono
3 0.6
2 CEBPD expression HSC CMP GMP Mono
1 0.5

Expression per cell (L0g2)

0
Fragments per cell

Fragments per cell

0.5 0.4 Early to
CEBPD-1:distal repress

14,005 variable regulatory elements

0.4 95% CI
0.3
0.3
1
0.2
0.2 Late to
repress
0.1 0.1

0
0 0
-2 0 2 4 6 8 10 12 14 -2 0 2 4 6 8 10 12 14 Transition
Myeloid pseudo-time Myeloid pseudo-time peaks

D Chromosome 8 Early to
Scale 1 Mb hg19 activate
chr8: 48,500,000 49,000,000 49,500,000 50,000,000

K1-HSC
K2-CMP
Late to
K9-GMP
activate
K10-GMP
Correlate dynamics
K11-Mono REs to nearby genes (+/-10Mb)
CD34+_HSPCs Early to
CEBPD MCM4 SNAI2 C8orf22 1 repress

Pearson
EFCAB1
PRKDC BC042029
KIAA0146 UBE2V2

0.5 Late to
<0.5 repress

1,983 variable genes

E 0.5
F 0.3
G
4 Transition
0.4 expression
Interaction enrichment at
cis-eQTLs (-log10 p-val)
Mean Pearson

Mean Pearson

0.2 3
0.3
Early to
2
0.2 activate
0.1

0.1 1
Late to
activate
0 0 0
s s
b b b b
<1k -10k 100k b-1M -10M
b op op op
s airs scATAC/RNA capture -2 0 2 4 6 8 10 12 14
1 f. lo f. lo nf. lo ene p
10- 100k 1
c on . con c o - g correlation HiC Myeloid pseudo-time
igh d k
h me low l pea
Al

Figure 6. Regulatory Element Dynamics Links Distal Elements to Genes

(A) Fragments per cell for a CEBPD distal element ordered by myeloid pseudo-time, (top) cells are colored by their sorted identity and (bottom) values are
smoothed (blue). Error bars (gray) denotes 95% confidence intervals.
(B) cis-Regulatory and expression dynamics across four regulatory elements near the myeloid regulator CEBPD.
(C) Accessibility (top) and expression (bottom) dynamics across myeloid pseudo-time, rows are sorted by their peak intensity in the myeloid trajectory.
(D) Regulatory profiles surrounding the CEBPD gene, dynamic enhancers are highlighted in gray with significant (blue) and non-significant (gray) correlated peak-
gene pairs shown as loops.
(E and F) Mean Pearson correlation coefficients binned by (E) genomic distance to the gene and (F) loop confidence. Error bars represent 1 SD on the estimate of
the mean.
(G) p value of enriched peak-gene correlation or promoter capture HiC at cis-eQTLs overlapping dynamic enhancers.
See also Figure S6.

developmentally committed states. While this reference-guided already or soon-to-be collected (Regev et al., 2017). As such,
approach enabled us to pair scATAC-seq and scRNA-seq data the data generated here and associated computational methods
along a common lineage trajectory, this approach may be gener- may be broadly adapted to further develop computational tools
alized to pair cells along more-diverse cell fate transitions. to pair different single-cell data types.
Notably, methods for computationally pairing multi ‘‘-omic’’ pro- Furthermore, single-cell CALs can be aggregated to define
files have advantages over experimentally coupled approaches, unique cis-regulatory elements active at different stages of
for example, a computational approach may provide (1) more differentiation. The intersection of genetic variants with these reg-
flexible experimental workflows, (2) allow pairing data across ulatory elements may provide new insights into cell types or stages
experimental methods that may not be easily combined, and of differentiation relevant to disease (Corces et al., 2016; Guo
(3) the reanalysis of the large repertoire of scRNA-seq data et al., 2017). Current experimental methods that aim to associate

Cell 173, 1535–1548, May 31, 2018 1545

non-coding genetic variation to changes in gene expression
generally measure either physical interactions using chromatin
conformation capture approaches (Javierre et al., 2016) or direct
genetic perturbation (Fulco et al., 2016). Here, we show that corre-
lation of naturally occurring regulatory heterogeneity across sin-
gle-cells can be used to pair regulatory elements to target genes.
This single-cell inference approach for linking regulatory elements
to genes may be particularly useful for inferring enhancer-gene
interactions in rare cells or across cells states where FACS
markers are not well defined. We expect future studies will
combine an integrated single-cell inference approach with phys-
ical interaction or genetic perturbation maps for improved linking
of enhancers to target genes, providing a single-cell resolved
interaction landscape of non-coding genetic variation.
Overall this work has defined one representation of the
epigenomic states underlying hematopoiesis, reminiscent of
Waddington’s landscape of differentiation. However, given the
static snapshot of the CAL profiles we have quantified it remains
uncertain to what degree density of this landscape might allow
inference of cell state transition kinetics and potential. Joint
measures with the emerging repertoire of CRISPR-based tools
for lineage tracing (Woodworth et al., 2017) will be essential for
quantifying the epigenomic contribution of lineage priming on
cell fate decisions over time. It also remains to be seen to what
extent lineage priming is reflected in transcriptional diversity
within HSCs and whether the lineage-associated CAL variability
we observe within HSCs is tightly coupled with transcriptional
changes (Yu et al., 2016). We expect future work to couple
single-cell epigenomic, transcriptomic, proteomic, and lineage
measures may reveal important insights into the molecular
details and temporal order of initiating regulatory factors govern-
ing multipotent cell fate transitions. Altogether, we expect further
improvements in experimentally or computationally integrating
multiple single-cell data types will unravel a dynamic regulatory
landscape providing a single-cell resolved systems perspective
for developmental or disease cell fate decisions.
STAR+METHODS
Detailed methods are provided in the online version of this paper
and include the following:
d KEY RESOURCES TABLE
d CONTACT FOR REAGENT AND RESOURCE SHARING
d EXPERIMENTAL MODEL AND SUBJECT DETAILS
B Cell collection and isolation
d METHOD DETAILS
B Single-cell ATAC-seq and single-cell RNA-seq
B Single-cell RNA-seq
B Bulk ATAC-seq and RNA-seq
d QUANTIFICATION AND STATISTICAL ANALYSIS
B Data pre-processing and TF scores
B PCA projection
B Clustering, K-medoids and computing density
B Lineage bias analysis
B Significantly differential CMP peaks
B Ordering cells for pseudo-time and smoothing
B Filtering and regulatory element analysis
1546 Cell 173, 1535–1548, May 31, 2018
B Bulk and single cell RNA-Seq analysis
B Matching scRNA-seq to scATAC-seq
B Integration of promoter capture Hi-C data
B Monocyte cis-eQTL data
d DATA AND SOFTWARE AVAILABILITY
d ADDITIONAL RESOURCES
SUPPLEMENTAL INFORMATION
Supplemental Information includes six figures, two tables, and two data files
and can be found with this article online at https://doi.org/10.1016/j.cell.
2018.03.074.
ACKNOWLEDGMENTS
We thank members of Greenleaf, Chang, Majeti, and Buenrostro labs for valu-
able discussions. We acknowledge the C. Bustamante lab for help with
sequencing. This work was supported by NIH (P50HG007735 and
UM1HG009442 to H.Y.C. and W.J.G. and U19AI057266 to W.J.G.), Stine-
hart-Reed Foundation (to R.M. and H.Y.C.), the Rita Allen Foundation (to
W.J.G.), the Baxter Foundation Faculty Scholar Grant, and the Human Fron-
tiers Science Program grant RGY006S (to W.J.G). W.J.G is a Chan Zuckerberg
Biohub investigator and acknowledges grants 2017-174468 and 2018-182817
from the Chan Zuckerberg Initiative. J.D.B. acknowledges support from the
Harvard Society of Fellows and Broad Institute Fellowship. J.D.B. also ac-
knowledges the Allen Distinguished Investigator Program, through The Paul
G. Allen Fronteirs Group, for funding. R.M. is a Leukemia and Lymphoma So-
ciety Scholar. M.R.C is a Fellow of The Leukemia & Lymphoma Society.
AUTHOR CONTRIBUTIONS
J.D.B., M.R.C., H.Y.C., and W.J.G. conceived the project. M.R.C. and R.M.
performed cell sorting. J.D.B. performed ATAC-seq and scATAC-seq data
analysis and oversaw scATAC-seq library generation and protocol optimiza-
tion performed by B.W. B.W. generated the scATAC, bulk ATAC-seq, bulk
RNA-seq, and scRNA-seq data. C.A.L. performed the RNA-seq and PCHi-C
data analysis with help from M.J.A.. A.N.S. developed the TF motif analysis
tools. C.A.L. developed the associated web resource with help from M.J.A.
J.D.B. and W.J.G. wrote the manuscript with input from all authors.
DECLARATION OF INTERESTS
Stanford University has filed a provisional patent on ATAC-seq; J.D.B., H.Y.C.,
and W.J.G. are named as inventors. H.Y.C. and W.J.G. are scientific co-foun-
ders of Epinomics.
Received: August 14, 2017
Revised: January 3, 2018
Accepted: March 27, 2018
Published: April 26, 2018
REFERENCES
Argiropoulos, B., and Humphries, R.K. (2007). Hox genes in hematopoiesis
and leukemogenesis. Oncogene 26, 6766–6776.
Becker, A.J., McCulloch, E.A., and Till, J.E. (1963). Cytological demonstration
of the clonal nature of spleen colonies derived from transplanted mouse
marrow cells. Nature 197, 452–454.
Brennecke, P., Anders, S., Kim, J.K., Ko1odziejczyk, A.A., Zhang, X., Proser-
pio, V., Baying, B., Benes, V., Teichmann, S.A., Marioni, J.C., et al. (2013).
Accounting for technical noise in single-cell RNA-seq experiments. Nat.
Methods 10, 1093–1095.
Buenrostro, J.D., Giresi, P.G., Zaba, L.C., Chang, H.Y., and Greenleaf, W.J.
(2013). Transposition of native chromatin for fast and sensitive epigenomic
profiling of open chromatin, DNA-binding proteins and nucleosome position.
Nat. Methods 10, 1213–1218.
Buenrostro, J.D., Wu, B., Chang, H.Y., and Greenleaf, W.J. (2015a).
ATAC-seq: a method for assaying chromatin accessibility genome-wide.
Curr. Protoc. Mol. Biol. 109, 21.29.1–21.29.9.
Buenrostro, J.D., Wu, B., Litzenburger, U.M., Ruff, D., Gonzales, M.L., Snyder,
M.P., Chang, H.Y., and Greenleaf, W.J. (2015b). Single-cell chromatin acces-
sibility reveals principles of regulatory variation. Nature 523, 486–490.
Busch, K., Klapproth, K., Barile, M., Flossdorf, M., Holland-Letz, T., Schlenner,
S.M., Reth, M., Höfer, T., and Rodewald, H.-R. (2015). Fundamental properties
of unperturbed haematopoiesis from stem cells in vivo. Nature 518, 542–546.
Calo, E., and Wysocka, J. (2013). Modification of enhancer chromatin: what,
how, and why? Mol. Cell 49, 825–837.
Chen, L., Kostadima, M., Martens, J.H.A., Canu, G., Garcia, S.P., Turro, E.,
Downes, K., Macaulay, I.C., Bielczyk-Maczynska, E., Coe, S., et al. (2014).
Transcriptional diversity during lineage commitment of human blood progeni-
tors. Science 345, 1251033.
Corces, M.R., Buenrostro, J.D., Wu, B., Greenside, P.G., Chan, S.M., Koenig,
J.L., Snyder, M.P., Pritchard, J.K., Kundaje, A., Greenleaf, W.J., et al. (2016).
Lineage-specific and single-cell chromatin accessibility charts human hema-
topoiesis and leukemia evolution. Nat. Genet. 48, 1193–1203.
Crane, G.M., Jeffery, E., and Morrison, S.J. (2017). Adult haematopoietic stem
cell niches. Nat. Rev. Immunol. 17, 573–590.
Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut,
P., Chaisson, M., and Gingeras, T.R. (2013). STAR: ultrafast universal RNA-seq
aligner. Bioinformatics 29, 15–21.
Espı́n-Palazón, R., Stachura, D.L., Campbell, C.A., Garcı́a-Moreno, D., Del
Cid, N., Kim, A.D., Candel, S., Meseguer, J., Mulero, V., and Traver, D.
(2014). Proinflammatory signaling regulates hematopoietic stem cell emer-
gence. Cell 159, 1070–1085.
Fairfax, B.P., Humburg, P., Makino, S., Naranbhai, V., Wong, D., Lau, E., Jos-
tins, L., Plant, K., Andrews, R., McGee, C., and Knight, J.C. (2014). Innate
immune activity conditions the effect of regulatory variants upon monocyte
gene expression. Science 343, 1246949.
Farlik, M., Halbritter, F., Müller, F., Choudry, F.A., Ebert, P., Klughammer, J.,
Farrow, S., Santoro, A., Ciaurro, V., Mathur, A., et al. (2016). DNA methylation
dynamics of human hematopoietic stem cell differentiation. Cell Stem Cell 19,
808–822.
Frelin, C., Herrington, R., Janmohamed, S., Barbara, M., Tran, G., Paige, C.J.,
Benveniste, P., Zuñiga-Pflücker, J.-C., Souabni, A., Busslinger, M., and
Iscove, N.N. (2013). GATA-3 regulates the self-renewal of long-term hemato-
poietic stem cells. Nat. Immunol. 14, 1037–1044.
Fulco, C.P., Munschauer, M., Anyoha, R., Munson, G., Grossman, S.R., Perez,
E.M., Kane, M., Cleary, B., Lander, E.S., and Engreitz, J.M. (2016). Systematic
mapping of functional enhancer-promoter connections with CRISPR interfer-
ence. Science 354, 769–773.
Goldberg, A.D., Allis, C.D., and Bernstein, E. (2007). Epigenetics: a landscape
takes shape. Cell 128, 635–638.
Graf, T., and Enver, T. (2009). Forcing cells to change lineages. Nature 462,
587–594.
Grün, D., Muraro, M.J., Boisset, J.-C., Wiebrands, K., Lyubimova, A., Dhar-
madhikari, G., van den Born, M., van Es, J., Jansen, E., Clevers, H., et al.
(2016). De novo prediction of stem cell identity using single-cell transcriptome
data. Cell Stem Cell 19, 266–277.
Guo, M.H., Nandakumar, S.K., Ulirsch, J.C., Zekavat, S.M., Buenrostro, J.D.,
Natarajan, P., Salem, R.M., Chiarle, R., Mitt, M., Kals, M., et al. (2017).
Comprehensive population-based genome sequencing provides insight into
hematopoietic regulatory mechanisms. Proc. Natl. Acad. Sci. USA 114,
E327–E336.
Haghverdi, L., Büttner, M., Wolf, F.A., Buettner, F., and Theis, F.J. (2016). Diffu-
sion pseudotime robustly reconstructs lineage branching. Nat. Methods 13,
845–848.
Javierre, B.M., Burren, O.S., Wilder, S.P., Kreuzhuber, R., Hill, S.M., Sewitz, S.,
Cairns, J., Wingett, S.W., Várnai, C., Thiecke, M.J., et al.; BLUEPRINT
Consortium (2016). Lineage-specific genome architecture links enhancers
and non-coding disease variants to target gene promoters. Cell 167,
1369–1384.
Karamitros, D., Stoilova, B., Aboukhalil, Z., Hamey, F., Reinisch, A., Samitsch,
M., Quek, L., Otto, G., Repapi, E., Doondeea, J., et al. (2018). Single-cell anal-
ysis reveals the continuum of human lympho-myeloid progenitor cells. Nat.
Immunol. 19, 85–97.
Kelsey, G., Stegle, O., and Reik, W. (2017). Single-cell epigenomics: recording
the past and predicting the future. Science 358, 69–75.
Laurenti, E., and Göttgens, B. (2018). From haematopoietic stem cells to com-
plex differentiation landscapes. Nature 553, 418–426.
Lawrence, H.J., Helgason, C.D., Sauvageau, G., Fong, S., Izon, D.J., Humph-
ries, R.K., and Largman, C. (1997). Mice bearing a targeted interruption of the
homeobox gene HOXA9 have defects in myeloid, erythroid, and lymphoid he-
matopoiesis. Blood 89, 1922–1930.
Li, H., Courtois, E.T., Sengupta, D., Tan, Y., Chen, K.H., Goh, J.J.L., Kong, S.L.,
Chua, C., Hon, L.K., Tan, W.S., et al. (2017). Reference component analysis of
single-cell transcriptomes elucidates cellular heterogeneity in human colo-
rectal tumors. Nat. Genet. 49, 708–718.
Long, H.K., Prescott, S.L., and Wysocka, J. (2016). Ever-changing landscapes:
transcriptional enhancers in development and evolution. Cell 167, 1170–1187.
Magnusson, M., Brun, A.C.M., Lawrence, H.J., and Karlsson, S. (2007).
Hoxa9/hoxb3/hoxb4 compound null mice display severe hematopoietic de-
fects. Exp. Hematol. 35, 1421–1428.
Manz, M.G., Miyamoto, T., Akashi, K., and Weissman, I.L. (2002). Prospective
isolation of human clonogenic common myeloid progenitors. Proc. Natl. Acad.
Sci. USA 99, 11872–11877.
Matsuoka, Y., Sumide, K., Kawamura, H., Nakatsuka, R., Fujioka, T., Sasaki,
Y., and Sonoda, Y. (2015). Human cord blood-derived primitive CD34-negative
hematopoietic stem cells (HSCs) are myeloid-biased long-term repopulating
HSCs. Blood Cancer J. 5, e290.
Mifsud, B., Tavares-Cadete, F., Young, A.N., Sugar, R., Schoenfelder, S., Fer-
reira, L., Wingett, S.W., Andrews, S., Grey, W., Ewels, P.A., et al. (2015). Map-
ping long-range promoter contacts in human cells with high-resolution capture
Hi-C. Nat. Genet. 47, 598–606.
Naik, S.H., Perié, L., Swart, E., Gerlach, C., van Rooij, N., de Boer, R.J., and
Schumacher, T.N. (2013). Diverse and heritable lineage imprinting of early hae-
matopoietic progenitors. Nature 496, 229–232.
Notta, F., Gan, O.I., Wilson, G., Kaufmann, K.B., Mcleod, J., Laurenti, E., Dun-
ant, C.F., John, D., Stein, L.D., Dror, Y., et al. (2016). Distinct routes of lineage
development reshape the human blood hierarchy across ontogeny. Science
351, aab2116.
Novershtern, N., Subramanian, A., Lawton, L.N., Mak, R.H., Haining, W.N.,
McConkey, M.E., Habib, N., Yosef, N., Chang, C.Y., Shay, T., et al. (2011).
Densely interconnected transcriptional circuits control cell states in human he-
matopoiesis. Cell 144, 296–309.
Orkin, S.H., and Zon, L.I. (2008). Hematopoiesis: an evolving paradigm for
stem cell biology. Cell 132, 631–644.
Paul, F., Arkin, Y., Giladi, A., Jaitin, D.A., Kenigsberg, E., Keren-Shaul, H.,
Winter, D., Lara-Astiaso, D., Gury, M., Weiner, A., et al. (2015). Transcriptional
heterogeneity and lineage commitment in myeloid progenitors. Cell 163,
1663–1677.
Pei, W., Feyerabend, T.B., Rössler, J., Wang, X., Postrach, D., Busch, K.,
Rode, I., Klapproth, K., Dietlein, N., Quedenau, C., et al. (2017). Polylox bar-
coding reveals haematopoietic stem cell fates realized in vivo. Nature 548,
456–460.
Perié, L., Duffy, K.R., Kok, L., de Boer, R.J., and Schumacher, T.N. (2015). The
branching point in erythro-myeloid differentiation. Cell 163, 1655–1662.
Regev, A., Teichmann, S.A., Lander, E.S., Amit, I., Benoist, C., Birney, E., Bod-
enmiller, B., Campbell, P.J., Carninci, P., Clatworthy, M., et al. (2017). Science
Forum: The Human Cell Atlas. eLife 6, e27041.
Cell 173, 1535–1548, May 31, 2018 1547
Satpathy, A.T., Wu, X., Albring, J.C., and Murphy, K.M. (2012). Re(de)fining the
dendritic cell lineage. Nat. Immunol. 13, 1145–1154.
Sawamiphak, S., Kontarakis, Z., and Stainier, D.Y.R. (2014). Interferon gamma
signaling positively regulates hematopoietic stem cell emergence. Dev. Cell
31, 640–653.
Schep, A.N., Wu, B., Buenrostro, J.D., and Greenleaf, W.J. (2017). chromVAR:
Inferring transcription factor variation from single-cell epigenomic data. Nat.
Methods 14, 975–978.
Shin, J., Berg, D.A., Zhu, Y., Shin, J.Y., Song, J., Bonaguidi, M.A., Enikolopov,
G., Nauen, D.W., Christian, K.M., Ming, G.-L., and Song, H. (2015). Single-cell
RNA-seq with waterfall reveals molecular cascades underlying adult
neurogenesis. Cell Stem Cell 17, 360–372.
Takahashi, K., and Yamanaka, S. (2006). Induction of pluripotent stem cells
from mouse embryonic and adult fibroblast cultures by defined factors. Cell
126, 663–676.
Velten, L., Haas, S.F., Raffel, S., Blaszkiewicz, S., Islam, S., Hennig, B.P.,
Hirche, C., Lutz, C., Buss, E.C., Nowak, D., et al. (2017). Human haemato-
1548 Cell 173, 1535–1548, May 31, 2018
poietic stem cell lineage commitment is a continuous process. Nat. Cell Biol.
19, 271–281.
Waddington, C. (1957). The Strategy of the Genes: A Discussion of Some
Aspects of Theoretical Biology (Allen & Unwin).
Woodworth, M.B., Girskis, K.M., and Walsh, C.A. (2017). Building a lineage
from single cells: genetic techniques for cell lineage tracking. Nat. Rev. Genet.
18, 230–244.
Yu, V.W.C., Yusuf, R.Z., Oki, T., Wu, J., Saez, B., Wang, X., Cook, C., Bar-
yawno, N., Ziller, M.J., Lee, E., et al. (2016). Epigenetic memory underlies
cell-autonomous heterogeneous behavior of hematopoietic stem cells. Cell
167, 1310–1322.
Zhao, C., Xiu, Y., Ashton, J., Xing, L., Morita, Y., Jordan, C.T., and Boyce, B.F.
(2012). Noncanonical NF-kB signaling regulates hematopoietic stem cell self-
renewal and microenvironment interactions. Stem Cells 30, 709–718.
Zheng, G.X.Y., Terry, J.M., Belgrader, P., Ryvkin, P., Bent, Z.W., Wilson, R.,
Ziraldo, S.B., Wheeler, T.D., McDermott, G.P., Zhu, J., et al. (2017). Massively
parallel digital transcriptional profiling of single cells. Nat. Commun. 8, 14049.
STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies
Full list in SupplementaryTable1 This paper Table S1
Biological Samples
Healthy adult bone marrow allcells https://www.allcells.com/products/whole-
bone-marrow-aspirate
Critical Commercial Assays
Nextera DNA Library Preparation Kit Illumina FC-121-1030
C1 Single-Cell Auto Prep IFC for Open App Fluidigm 100-8133
Chromium Single Cell 30 Library & Gel Bead Kit v2 10X genomics 120267
Deposited Data
Raw sequencing data This paper GEO: GSE96772
Software and Algorithms
Bowtie2 http://bowtie-bio.sourceforge.net/bowtie2/
index.shtml
Samtools http://samtools.sourceforge.net/
chromVAR Schep et al., 2017 https://bioconductor.org/packages/release/
bioc/html/chromVAR.html
CellRanger v1.2.0 10X genomics https://support.10xgenomics.com/single-cell-
gene-expression/software/downloads/latest
STAR 2.5.1b Dobin et al., 2013 https://github.com/alexdobin/STAR
Other
Bulk ATAC-seq and bulk RNA-seq Corces et al., 2016 GEO: GSE74246
Promoter capture HiC, CD34+ Mifsud et al., 2015 STAR Methods
Promoter capture HiC, monocytes Javierre et al., 2016 STAR Methods
Cis-eQTL, monocytes Fairfax et al., 2014 STAR Methods

CONTACT FOR REAGENT AND RESOURCE SHARING

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, William J.
Greenleaf (wjg@stanford.edu).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell collection and isolation

Human bone marrow was sorted as previously described (Corces et al., 2016). In addition to the cell types previously described, we
also isolated plasmacytoid dendritic cells (pDCs), an unknown population (UNK) and megakaryocytes. To isolate pDCs from human
bone marrow using FACS we gated for live, lineage negative, CD34+ CD38+ CD10- CD45RA+ CD123+ cells. To isolate UNK cells
from human bone marrow by FACS, we gated for live, lineage negative, CD34+ CD38+ CD10- CD45RA+ CD123-. Megakaryocytes
were isolated using in vitro differentiation of bone marrow derived CD34+ cells to megakaryocytes. To do this, CD34+ cells were
cultured in StemSpan SFEM with Megakaryocyte Expansion Supplement (Stem Cell Technologies) for 14 days, yielding
approximately 100-fold expansion in cell number. After cell isolation of all populations using FACS, 15,000 single-cells were
resuspended in 100 mL of BAMBANKER serum-free cell freezing medium (Wako Chemicals, 302-14681) and cryopreserved in liquid
nitrogen. Samples were collected commercially from allcells. Further information about the donors profiled and FACS protocols can
be found in Table S1.

Cell 173, 1535–1548.e1–e5, May 31, 2018 e1

METHOD DETAILS

Single-cell ATAC-seq and single-cell RNA-seq

Single-cells not cryopreserved after FACS (fresh) were assayed as previously described (Buenrostro et al., 2015b). To assay cells
cryopreserved after FACS (frozen), cells were allowed to recover for 10 min at 37 C in IMDM with 10% FBS. After recovery, cells
were washed twice in cold 1x PBS and once in with the C1 DNA Seq Cell Wash Buffer (Fluidigm). Cells were then resuspended in
6 mL of C1 DNA Seq Cell Wash Buffer, and were combined with 4 mL of C1 Cell Suspension Reagent, 7 mL of this cell mix was loaded
onto the Fluidigm IFC. Single-cells were then assayed using scATAC-seq as previously described (Buenrostro et al., 2015b).

Single-cell RNA-seq
Single-cell RNA-seq data was collected using the recommended protocol for the 30 scRNA-seq 10X genomics platform
using v2 chemistry. DNA shearing used the Covaris S220, libraries were sequenced on a NextSeq with 26x8x98 read lengths.

Bulk ATAC-seq and RNA-seq

ATAC-seq and RNA-seq libraries were generated as previously described (Buenrostro et al., 2015b; Corces et al., 2016) with slight
modifications for frozen cells. One vial of 15,000 frozen cells in 100 mL of BAMBANKER freezing medium was quickly thawed at 37 C.
70 mL for bulk ATAC-seq and 30 mL for bulk RNA-seq were then added to 500 mL of warm IMDM with 10% FBS. For bulk ATAC-seq,
the cells were split into 2 tubes of 5,000 cells used as technical replicates. Cells were washed twice in 1x PBS, all supernatant was
carefully removed without disturbing the cell pellet, and cells were resuspended in 40 mL of transposition mix (20 mL of 2x TD buffer,
2 mL of TDE1, 0.2 mL of 2% digitonin, 13.33 mL of 1x PBS, and 4.47 mL of nuclease-free water) (Illumina, FC-121-1030; Promega,
G9441), here the transposition reactions were scaled down to compensate for cell loss during washes. Transposition reactions
were incubated at 37 C for 30 min in an Eppendorf ThermoMixer with agitation at 300 rpm. The transposed DNA fragments were
purified and amplified as described (Corces et al., 2016).
For bulk RNA-seq, cells were split into 2 technical replicates. RNA was isolated using the QIAGEN RNeasy Plus Micro kit, and RNA
was eluted in 10 mL of RNase-free water. 5 mL of total RNA was used as input for NuGen Ovation V2 cDNA synthesis kit. The yield of
purified SPIA-amplified cDNA was measured using Qubit dsDNA HS Assay kit. 50 ng of SPIA cDNA was fragmented using Nextera
DNA library preparation kit (Illumina, FC-121-1030). Fragmented SPIA cDNA was then purified using QIAGEN MinElute Reaction
Cleanup Kit, and purified DNA was eluted in 10 mL of elution buffer (10mM Tris-HCl, pH 8). Purified SPIA cDNA fragments was
amplified and purified as previously described for ATAC-seq (Corces et al., 2016). ATAC-seq and RNA-seq libraries were quantified
using qPCR, amplified libraries were sequenced using paired-end, dual-index sequencing on a NextSeq 500 instrument with 76 bp
read lengths.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data pre-processing and TF scores

Single-cell and bulk ATAC-seq alignment, quality filtering and peak calling was performed as previously described (Corces et al.,
2016), with one exception. For single-cell profiles, any fragment that occurred in two cells from the same experiment (a single
96-cell IFC) was removed from further analysis. Using the previously described approach (Corces et al., 2016), we defined a peak
list using all bulk hematopoietic data analyzed here, resulting in 491,437 500bp non-overlapping peaks which we use for the
remainder of this study. To count the number of fragments per peak in a sample or cell, for computing TF z-scores per cell and
for determining background peak sets matched in GC and peak intensity, we used the default settings in the chromVAR package
(Schep et al., 2017). Single-cells were filtered for quality requiring at least 60% of fragments in peaks and requiring greater than
1,000 fragments passing quality filters, quality filters are previously described (Corces et al., 2016) which includes removal of mito-
chondrial reads and low alignment quality (Q30). Bulk counts were normalized as previously described (Corces et al., 2016) using
quantile normalization and the CQN package.

PCA projection
To calculate PCs on the bulk datasets, later used for projecting single-cell profiles, we first removed peaks in annotated promoters or
aligning to chrX, peaks associated with chrY and unmapped contigs were filtered out in the preprocessing steps described above.
455,057 peaks remained after filtering and were used for the PCA projection analysis. To normalize the bulk count matrix by library
size, we identified 19,287 low variance promoters, using a MATLAB implementation of a previously described approach (Brennecke
et al., 2013), across all bulk samples and normalized each sample by the mean fragment counts within the low variance promoters.
We subsequently took the mean counts of all normalized bulk sample replicates (HSC, MPP, LMPP, CMP, GMP, MEP, Mono, CD4,
CD8, NK, NKT, B, CLP, Ery, UNK, pDC and Megakaryocyte) and performed PCA-SVD, resulting in 15 principal components.
To score single-cells by the activity for each component, we first centered the counts for each cell by dividing each peak by the
mean fragment counts in peaks for a given cell. We then used the weighted coefficients for each peak and PC (determined using
PCA-SVD of the bulk data above) to take the product of the weighted PC coefficients by the centered count values for each cell,
taking the sum of this value resulted in a matrix of cells by PCs. Last, we calculate a cell-cell similarity matrix using Pearson correlation

e2 Cell 173, 1535–1548.e1–e5, May 31, 2018

and perform PCA on the similarity matrix of correlation values. To assess this computational approach, we repeated this procedure
for the bulk data, down sampled bulk data, mean single-cell profiles and synthetic mixtures. Notably, we found a patient-specific
batch effect in the HSC single-cell profiles in the PCA projected sub-space, this batch signal was strongly correlated with Jun/Fos
TF z-scores. We therefore normalized each HSC batch to the mean value of all HSC profiles, and in addition, we blacklisted all motifs
correlated with Jun/Fos accessibility. Correlated Jun/Fos motifs were determined by calculating motif similarity (Pearson) (Schep
et al., 2017) and removing all motifs with an R > 0.8. We use these corrected PC values and filtered TF motifs for all subsequent
visualizations of these data.
To determine significant cell-cell variability in the PC projected sub-space, we either down sampled or permuted peaks by their GC
content and mean accessibility. To permute peaks that match GC% and mean accessibility, we take the sum accessibility of all cells
of a given immunophenotypically defined cell type (e.g., HSCs) and use the ChromVAR function ‘‘get_background_peaks’’ with the
default settings. Notably, ChromVAR samples background peaks with replacement and may select the observed peak as a
background peak, and therefore provides a conservative estimation of excess variability.

Clustering, K-medoids and computing density

All hierarchical clustering performed used Pearson correlation as the distance function. All k-medoids clustering shown in this work is
performed using 10 replicates with the distance function Pearson correlation. For k-mediods clustering throughout this work, the gap
statistic is used to determine the appropriate number of clusters, here the optimal cluster number is determined wherein the minimum
K satisfies the follow criteria: the gap value for a given K is greater than the gap value of K+1 minus the standard error of the clustering
solution for K+1. The first 5 PCs were chosen for clustering cells by their projected PC values. All metrics of cell data density shown in
this work are weighted by the expected in vivo frequency of each cell type, as measured from flow cytometry data. 2D data density is
calculated using KDE weighted by the in vivo frequency.

Lineage bias analysis

To compute TF variability within epigenome and immunophenotype pure (EIPP) cells, we first determined the most represented cell
type for each of the 14 k-medoids determined clusters. We defined EIPP clusters as cells that were immunophenotypically-marked
by this most-represented type, and also were within one of the k-medoids clusters. We then collected cluster-pure and immunophe-
notype-pure profiles for HSCs (k1 cluster) and LMPPs (k8 cluster), and proceeded to compute variability and TF z-scores using
ChromVAR (Schep et al., 2017). To determine the magnitude and direction of the TF lineage bias, we first partitioned TF z-scores
as greater than zero (high) or less than zero (low). We then computed the mean of the first 5 PCs from the PCA projection for the cells
assigned to the high or low TF z-score, distance of the high and low centroids was calculated using Euclidean distance. We
determined significance using ‘‘direction z-scores,’’ whereby we repeated the analysis described above for PCs calculated using
50 background peak permutations matching GC and mean accessibility, peaks were determined using ChromVAR (Schep et al.,
2017). Direction z-scores were computed comparing the observed to the 50 permuted distances.

Significantly differential CMP peaks

To determine differential CMP peaks, aggregate CMP profiles for each k-medoids cluster were collected. Pairwise binomial tests
were performed for each aggregate profile (4 CMP clusters) for each peak. Peaks with a p value of < 10
5 in one or more comparisons
was used for further analysis (n = 1,801 peaks). For clustering and visualization, counts were normalized using column z-scores and
clustered using k-medoids, the gap statistic (as described above) was used to determine a K of 6.

Ordering cells for pseudo-time and smoothing

To determine continuous differentiation trajectories, we developed a supervised approach, similar to a recently published method
(Shin et al., 2015). To do this, we first fit a line through each cluster centroid for the relevant clusters using linear interpolation across
the first 5 PCs from the PC projection described above. These relevant clusters were determined using prior literature describing
functional cell differentiation trajectories. To assign each cell to a given trajectory, we determined the closest point of each cell within
the implicated clusters to the interpolated fit line that connected each cluster centroid across the 5 projected PCs (the closest point
was determined using Euclidean distance). The pseudo-time values represented in the main figures represent Euclidean distance
along the interpolated line across the 5 PCs from the projected PC space, all values are relative to the mean of HSCs (defined as
the start point for all trajectories) and thus represent the value 0 in all pseudo-time trajectories. To order cells by myeloid develop-
ment, only cells within clusters K1, K2, K9, K10, K11 were considered. For the erythroid, lymphoid and pDC trajectories, only the
following clusters were considered: erythroid (K1,3,5,6,7), lymphoid (K1,2,8,14) and pDC (K1,2,8,12,13). To determine the TF motif
accessibility dynamics across this inferred trajectory, we smoothed the TF z-scores along myeloid progression with a lowess function
with a span of 200, implemented within the MATLAB function smooth. Continuous profiles were normalized by their min/max value for
plotting and downstream clustering. The accessibility of individual peaks were smoothed and normalized as was done with TFs,
however, with the notable difference of smoothing using a span of 500. To determine error in the smoothed TF z-score profiles
and cis-regulatory elements we computed 95% confidence intervals by resampling cells (n = 100 permutations) with replacement.
We next repeated the smoothing for each permutation and used the MATLAB function paramci to determine the 95% confidence
interval per TF or regulatory element.

Cell 173, 1535–1548.e1–e5, May 31, 2018 e3

Filtering and regulatory element analysis
To determine TFs that were significantly variable within the smoothed myeloid trajectories, we compared the standard deviation of
the observed smoothed scores to a set of similarly smoothed, permuted, TF scores generated by randomly permuting the myeloid
cell order. Selecting TFs with a standard deviation greater than 0.5 provided 205 motifs with an FDR < 1%. We used two criteria to
determine highly variable regulatory elements, to do this we first filtered regulatory elements with greater than a mean
accessibility of 0.01 fragments per cell and subsequently filtered for regulatory elements with a coefficient of variation of greater
than 0.5 (n = 3,403), resulting in a high-quality list. This list was used for all analysis of regulatory elements except for quantifying
enrichment at cis-eQTLs. For cis-eQTL analysis we reduced the coefficient of variation filter to 0 to yield a larger list of peaks for
analysis (n = 14,005).
To determine motif enrichment across dynamic accessible peaks, we first collected log-PWM scores per peak using ChromVAR
(Schep et al., 2017) for motifs that were selected in the TF analysis described above. Peak-to-peak distances were computed using
Euclidean distance of the PC scores determined by PCA of the min/max normalized accessibility dynamics across the myeloid
trajectory. For each peak, the mean log-PWM score was computed for the nearest 300 peaks.

Bulk and single cell RNA-Seq analysis

Raw RNA-Seq reads were aligned to hg19 and quantified per cell barcode using CellRanger v1.2.0 (https://support.10xgenomics.
com/single-cell-gene-expression/software/downloads/latest). While the full quantification pipeline is automated and reproducible
through CellRanger, we briefly describe the workflow here. First, raw sequencing data was demultiplexed using Illumina bcl2fastq
v2.17.1.14. Raw sequencing reads per cell barcode were aligned using STAR version 2.5.1b (Dobin et al., 2013) before duplicates
were removed based on unique molecular indicies (UMIs). Per cell, per transcript counts were aggregated, and cells with fewer
than 1,000 UMIs/gene counts were excluded. To augment our sorted scRNA-Seq data, we downloaded processed monocytes
and CD34+ scRNA-Seq data (as previously described (Zheng et al., 2017)) from the 10X website.
To match the exact feature set quantified in the scRNA-Seq pipeline, all bulk RNA-Seq (including samples originally described in
our previous study (Corces et al., 2016) and new samples described in this study) were aligned and quantified using STAR version
2.5.1b (Dobin et al., 2013) and the same gene transfer format file provided in the CellRanger v1.2.0 distribution. Read counts for
biological replicates were summed over each transcript to provide a single transcript count per sorted cell type.

Matching scRNA-seq to scATAC-seq

scRNA-seq profiles were matched to scATAC-seq profiles by first linking bulk ATAC-seq PCs to bulk RNA-seq expression profiles.
To do this, we first normalized the projected PC scores per cell by the sum-of-squares, which effectively normalizes to differences in
the number of reads per cell, and determined the mean projected PC score (in scATAC-seq space) for each immunophenotypically
defined cell type with matched bulk RNA-seq. We then trained a linear model, using stepwise regression (MATLAB implementation),
to fit the log2 expression of all bulk gene expression profiles across sorted populations as the linear combination of ATAC-seq PC’s.
Using the fit coefficients from the stepwise regression, we inferred the expression of all genes for each scATAC-seq profile to produce
a reference-assisted ‘‘inferred transcriptome.’’ To match scRNA-seq data to scATAC-seq cells, we first selected for genes that were
both variable (defined by a standard deviation across cells greater than 3) and well described by the regression model of bulk PCs
(R > 0.9) resulting in a total of 853 genes. To improve matching by denoising profiles, we performed PCA on the expression level
inferred for these variable genes, then scored profiles for both scATAC-seq ‘‘inferred transcriptomes’’ and scRNA-seq transcrip-
tomes by the PC loadings from this PCA. Next, we calculated the correlation (Pearson), using the top 20 PCs, between each
scATAC-seq ‘‘inferred transcriptome’’ and each cell from the scRNA-seq dataset. Single-cell RNA-seq profiles were then matched
to ATAC-seq data by selecting the single-cell ATAC-seq cell with the maximum correlation coefficient. ScRNA-seq cells with low
correlation values (R < 0.9) were discarded from further analysis.

Integration of promoter capture Hi-C data

Processed promoter-capture Hi-C loops for CD34+ (Mifsud et al., 2015) and monocytes (Javierre et al., 2016) were downloaded from
the supplemental resources associated with each of the original publications. While the authors reported only significant loops for the
CD34+ data, we filtered loops from the full dataset, thresholding the interaction score to > 5 as recommended by the authors. In
instances where the promoter bait spanned two or more defined gene promoters (as reported in the original data file), loops were
considered for each promoter gene separately. In total, 275,848 significant loops for CD34+ and 443,980 significant loops for
monocytes were considered in our downstream analyses, 36,962 loop were overlapping (same target promoter; same distal regu-
latory annotation).

Monocyte cis-eQTL data

A list of statistically significant cis-eQTL associations were downloaded from the supplemental materials from Fairfax et al., (2014).
Only cis-eQTLs within the longer list of dynamic regulatory elements were considered (n = 14,005), no other filtering was performed.
The filter for variable genes was reduced to generate a longer gene list (std. > 2, n = 1,983). Using this expanded list 370 cis-eQTLs
with associated dynamic peak-gene pairs were available for analysis.

e4 Cell 173, 1535–1548.e1–e5, May 31, 2018

DATA AND SOFTWARE AVAILABILITY

The accession number for the sequencing data reported in this paper is GEO: GSE96772. Processed scATAC-seq data, which
include PC values and TF scores per cell can be found in Data S1. The software developed for analyzing these data, which includes
projecting scATAC-seq profiles onto bulk hematopoietic PCs and pairing single-cell ATAC-seq with single-cell RNA-seq, as well as
processed data from this manuscript can be found in Data S2.

ADDITIONAL RESOURCES

The data can be visualized in the UCSC genome browser using the track hubs representing bulk data (https://s3.amazonaws.com/
JasonBuenrostro/scATAC_heme_label/hub.txt) and clusters of single-cell ATAC-seq data (https://s3.amazonaws.com/
JasonBuenrostro/scATAC_heme/hub.txt). This manuscript is accompanied by a web resource for visualizing the single-cell
ATAC-seq data, which can be found at: http://schemer.buenrostrolab.com/.

Cell 173, 1535–1548.e1–e5, May 31, 2018 e5

Supplemental Figures

Figure S1. Quality Characteristics of Single-Cell Epigenomes, Related To Figure 1

(A and B) Waddington landscape representing (A) the sinuous epigenetic landscape wherein a cell (ball) can roll down different cell fates, and (B) ‘‘guy-wires’’ that
shape the epigenetic landscape.
(C and D) Comparison of scATAC-seq from ‘fresh’ (blue) and cells frozen after FACS sorting (red), only cells passing quality filtering are shown. Profiles from HSCs
showing the (C) fragment yield per cell and (D) fraction of fragments in peaks.
(E) Comparison of the log2 accessibility between donor-matched fresh and frozen aggregate accessibility profiles, R = 0.88.
(F) Fraction of immunophenotypically defined cell types from CD34+ cells for each bone marrow donor, ungated cells are marked in gray.
(G) The measured (blue) and average in vivo (yellow) frequency of cells in the dataset.
(H) Fragment size (bp) distribution and
(I) enrichment at transcription start sites (TSSs) for bulk (blue) and aggregate single-cell (red) profiles.
(J) Number of cells passing filter for each cell type assayed.
(K and L) Mean (K) fragment counts and (L) fraction of reads in peaks of cells passing filter for each cell type assayed. Error bars represent SEM.
 A 2081-GATA1-D-N7
B 558-CEBPD-D-N2
C Fragment counts, log10
D Fraction of reads in peaks (FRiP)
40 40 40 40
30 30 30 30
20 20 20 20
10 10 10 10

Dim. 2
Dim. 2
Dim. 2

Dim. 2
0 0 0 0
-10 -10 -10 -10
-20 -20 -20 -20
-30 -30 -30 -30
Z-score Z-score log10 counts FRiP
-40 -40 -40 2.7 4.8 -40 .55
-11.8 10.9 -7.2 7.1 .99
-50 -50 -50 -50
-60 -40 -20 0 20 40 60 -60 -40 -20 0 20 40 60 -60 -40 -20 0 20 40 60 -60 -40 -20 0 20 40 60
Dim. 1 Dim. 1 Dim. 1 Dim. 1
E Cells scored by ENCODE ChIP-seq F G Pearson
1
H 40 PCA of bulk samples
HSC LMPP GMP CLP Unknown 300
50 MPP CMP MEP pDC centroid
Mono 30
CD8

% Variancee Explained
40 bulk sample
CD4 20
30 200 NK -0.1
10
PC2 18.96%

20 HSC
100 Bcell MPP 0
10
Dim. 2

LMPP 0 2 4 6 8 10 12 14 16 18
0 Component #
CMP
-10 0 pDC CLP GMP 50 PCA of projected samples
GMP Mono 40
-20 LMPP MEP
HSC 30
-30 -100 CMP Mega CLP
MEP 20
-40 Ery pDC 10
-50 -200 Unknown 0
-80 -60 -40 -20 0 20 40 60 80 -200 0 200 400 600 0 2 4 6 8 10 12 14 16 18
PC1 37.34% scATAC-seq cells scored by bulk PCs Component #
Dim. 1
I PC1 more stem J PC2 lym-ery K PC3 myel-lym L PC4 CLP-pDC
4 2 8 7
2 1
7 6
0 0
-1 6
PC2 (37.8%)

PC3 (10.4%)

PC4 (2.6%)

PC5 (2.1%)
-2 5
-2 5
-4
-3 4
-6 4
-4
-8 -5 3
3
-10 -6
2 2
-12 -7
-14 -8 1 1
10 15 20 25 30 35 40 -14 -12 -10 -8 -6 -4 -2 0 2 4 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 1 2 3 4 5 6 7 8
PC1 (45.5%) PC2 (37.8%) PC3 (10.4%) PC4 (2.6%)
M N Peak accessibility based correlation PCA-based correlation PCA-based correlation across bulk samples
across bulk samples across bulk samples Downsampled to 10,000 fragments
PC5 CLP 1
Pearson

2.5
2
0
PC6 noise
PC6 (0.9%)

1.5
1
0.5 HSC
MPP
0 LMPP
CMP
-0.5 GMP
-1 MEP
1 2 3 4 5 6 7 CLP
PC5 (2.1%) pDC

O Downsampled
D n m d to m
mean d
depth
t (10k)
th 0
P Downsampled
D
Dow
wn
nsampled
n m d to m
matched
h depths
d hs
de h
Q Synthetic mixture R Synthetic mixture
Downsampled
Do ssampled
p to mean
ea depth
de
e Downsampled
Do s pl d to to mean
e de depth
epth
e
Ratio CMP/GMP Ratio LMPP/CLP

0.1 0.5 0.9 0.1 0.5 0.9

0 0 0 0

-2 -2 -2 -2

-4 -4 -4 -4
15 15 15 15
20 20 20 20
-6 25 -6 25 -6 25 -6 25
30 30 30 30
35 35 35 35
-12 -10 -8 -6 -4 -2 0 2 -12 -10 -8 -6 -4 -2 0 2 -12 -10 -8 -6 -4 -2 -12 -10 -8 -6 -4 -2
0 2 0 2
S Fragment counts, log10
T Fraction of reads in peaks (FRiP) U Fresh HSC vs frozen profiles V DonorID
2 2 2 2
1 1 1 1
0 0 0 0
-1 -1 -1 -1
-2 -2 -2 -2
PC3
PC3

PC3
PC3

-3 -3 -3 -3 BM4983
-4 -4 -4 -4 BM0106
BM0828
-5 -5 -5 -5 PB1022
-6 -6 -6 -6 BM1077
log10 counts FRiP BM1137
-7 -7 -7 -7
2.7 4.8 .55 .99 Frozen Fresh BM1214
-8 -8 -8 -8
-14 -12 -10 -8 -6 -4 -2 0 2 4 -14 -12 -10 -8 -6 -4 -2 0 2 4 -14 -12 -10 -8 -6 -4 -2 0 2 4 -14 -12 -10 -8 -6 -4 -2 0 2 4
PC2 PC2 PC2 PC2

Figure S2. Analytical Frameworks for Clustering scATAC-Seq Profiles, Related to Figure 2
(A–D) t-SNE embedding of cells colored by the TF z-score activity of (A) GATA1 and (B) CEBPD motifs, or the quality metrics (C) log10 fragment counts and (D)
fraction of reads in peaks.
(legend continued on next page)
(E) t-SNE embedding of single-cell profiles using ENCODE ChIP-seq of K562s as a feature vector, instead of the TF motifs shown above, cells are colored by their
cell identity.
(F) PC1 and PC2 from PCA of fragments in peaks for bulk samples (gray) and their centroids (red).
(G) (left) Hierarchical clustering of the correlation of single-cell epigenomic profiles scored by bulk PCs, (right) profiles colored by their sorted immunophenotype
identity.
(H) Percent variance explained for each PC from (top) PCA derived from bulk data using fragments in peaks and (bottom) PCA of the PCA projected subspace (see
STAR Methods).
(I–M) PC projection of single-cell ATAC-seq data showing cells scored by PC components (I) PC1 and PC2, (J) PC2 and PC3, (K) PC3 and PC4, (L) PC4 and PC5,
and (M) PC5 and PC6.
(N) Bulk sample-sample correlation matrix determined by correlation of (left) fragments in peaks, (middle) PCA projection values and (right) PCA projection values
after down sampling data to 10,000 fragments per sample. Far left represents the sorted immunophenotype of each bulk sample.
(O and P) PCA projection of mean single-cell profiles of immunophenotypically defined cell types down-sampled to (O) 10,000 or (P) matched fragment counts to
the observed single-cell dataset.
(Q and R) Synthetic mixtures of two immunophenotypically defined single-cell profiles down sampled to 10,000 fragments with varying mixtures of (Q) CMP/GMP
and (R) LMPP/CLP cell types, unmixed cell types from (O) are shown for reference.
(S–V) PCA projection of single-cells colored by (S) log10 fragment counts, (T) fraction of reads in peaks, (U) fresh HSC versus frozen profiles, and (V) donor.
 A MPP and GMP B CMP and LMPP C Sample Name Subset Name Count D Sample Name Subset Name Count
BM1077_PrimarySort CMP 11668 BM1077_PrimarySort GMP 16428
BM1077_PrimarySort CD34+CD38-CD10- 11885 BM1077_PrimarySort CD34+CD38-CD10- 11885

5 5
10 10

HSC LMPP HSC LMPP

0 0 10
4 4
10

CD90

CD90
PC3 (10.42%)

PC3 (10.42%)
-2 -2 3 3
10 10

-4 -4

%)

%)
15 15 10
2
10
2
20 20

(45.5

(45.5
-6 25 -6 25 0
MPP 0 MPP
30 30 -10
2
-10
2
35
PC1
35

PC1
0 1 2 3 4 5
-12 -10 -8 -6 -4 -2 0 2 -12 -10 -8 -6 -4 -2 0 2 10 10 10 10 10 10 10
0
10
1
10
2
10
3
10
4
10
5

PC2 (37.8%) PC2 (37.8%) CD45RA CD45RA

E 14 H I TF Z-score Column Z-score
5 Motifs 0 Max Peaks (N=1,801) -2 2
Fold variance over expected

12
Cell-cell TF variability

10 5
4

CMP Clusters
8 3
3
6
4
4 2

2 2
1
0 1 4 D I1 A 3 2 4 Erythroid bias Myeloid bias
TAATF EBP SP L11 ATF ACH TCF
GA
SC
PP

PP

P
P
EP
LP

0 400 800 1200 1600 C BC B

on
M
M

pD
C
M
LM

M
H

C
G

Sorted TF motifs
F Peak permutation by GC & peak intensity G 6 J Scale 10 kb hg19
chrX: 48,645,000 48,650,000 48,655,000 48,660,000
Fold variance over expected

5 GATA1 HDAC6
1_
4 5
0 CMP Clusters 0_
1_
3
PC3 (10.42%)

3
-2
0_
2 1_
-4 4
%)

15 1 0_
20
(45.5

1_
-6 25
30 0
2
35
PC1

-12 -10 0_
-8 -6 -4 -2 0 2
SC

PP

PP

P
EP
LP

o
on
M

pD

PC2 (37.8%)
C
M
LM

M
H

C
G

K 1 Max L Peak permutation by GC and read count M

2
RELA 2.2
CLP pDC GATA motif
HSC CTCF ID motif
0 LMPP CTCF motif
Calculate NKFB motif
Variability, TF z-score

RUNX2 hi/low distance 1.8 CTCFL ETS motif

CTCF for each TF RELA
-2 NFKB1
PC3

ERG FLI1 GATA2

MEP Calculate 1.4 ETV4 MESP2 GATA1
hi/low distance for ID3 MESP1 GATA3 GATA5
-4 ID4
GMP permuted scores
SPI1 CMP 1
-6 Compute
p-value
GATA1 using z test
CEBPD -8 0.6
ATF4 -12 -8 -4 0 4 0 5 10 15 20 25
PC2 Direction score, -log10 p-value
STAT1
IRF1
PRDM1
N O -3.4
Z-score
3.3
BCL11A TF motifs 2 TCF4 motif
TCF4
GATA5
0
GATA3

GATA2
PC3

LYL1 -2
TCF12 GATA1

MESP1 -4
PAX5
EBF1
MESP2
variability=1.68
- C
- P
- P
- P
- P
P
-L P

K1 -GMP
K1 -G P
K1 -Mo P
K1 -pD o
3 C
4- C
LP
2 n

-6
K3 CM
K4 CM
K5 CM
K6 CM
K7 ME
K8 -ME
K9 MP

1 M
K2 HS

K1 -pD
C

HSC EIPP pure epigenomes TF Z-score -3 3 direction p-value=10-81 Low High

-
K1

-12 -8 -4 0 4
PC2
P Q -3.5
Z-score
2.4 R -2.9
Z-score
3.2
TF motifs 2 STAT1 motif 2 CEBPE motif

REL
0 0

CEBPE
PC3
PC3

STAT1 -2 -2

ID3 -4 -4
TCF4
MESP1
variability=1.54
-6 variability=1.48 -6
direction p-value=10-11 Low High direction p-value=10-10 Low High
LMPP EIPP pure epigenomes TF Z-score -3 3 -12 -8 -4 0 4 -12 -8 -4 0 4
PC2 PC2

(legend on next page)

Figure S3. Sources of Variability within Defined Cell Types, Related to Figure 3
(A and B) PCA projection of highlighted cell types for (A) MPP and GMP, and (B) CMP and LMPP.
(C and D) Flow cytometry back gating of (C) CMPs and (D) GMPs to show that a subset of cells exhibit CD90 and CD45RA cell surface marker expression without
significant CD38 signal. These potentially mis-gated CMPs localize to the MPP gate while mis-gated GMPs localize to the LMPP gate.
(E) Fold variance of the PCA projection over the variability expected due to count noise, determined by down-sampling counts from the mean of each im-
munphenotypically defined cell type to matched sequencing depths of the observed single-cell profiles. Error bars represent 1 standard deviation estimated
using bootstrap sampling (1,000 iterations) of cells.
(F) Peaks are permuted by their GC content and the mean fragment count for each aggregate immunophenotypically defined cell type, and permuted single-cell
profiles are projected onto the PC subspace, un-permuted cells shown in gray for reference.
(G) Fold variance over expected for each cell type, quantified as described in (E) using the permuted scores shown in (F).
(H) TF motif z-score variability sorted by the rank score for each cell type.
(I) (left) Differential motifs and (right) regulatory elements across CMP clusters (K2-5), motifs are normalized by max-min values and regulatory elements are
normalized as z-scores and clustered using k-medoids.
(J) Accessibility at GATA1 locus across the CMP clusters highlighting (gray) two validated (Fulco et al., 2016) enhancers of GATA1.
(K) Cell-cell TF motif variability within each EIPP cluster (see STAR Methods).
(L) Peaks were permuted by their GC content and mean peak fragment count for each aggregate single-cell profile, single cell profiles were then projected onto
the PC subspace.
(M) (left) Schematic for determining direction p value using permuted PCA scores (n = 50) described in (F) and (L), (right) TF motif variability and
direction –log10 p value for each TF motif for the HSC EIPP cluster.
(N–R) Hierarchical clustering of single-cell (N) HSC and (P) LMPP EIPP profiles (columns) for TF motifs appearing as highly variable and directional (rows). (O-R)
PC2 and PC3 projection of single LMPP profiles colored by high (yellow) or low (blue) TF motif accessibility z-scores for (O) TCF4, (Q) STAT1 and (R) CEBPE
motifs, arrows denote the direction of the signal bias.
 A Myel. trajectory B Ery. trajectory C Lym. trajectory D pDC. trajectory
180 300 180 200
160 160 180
250 160
140 140

Num of cells
Num of cells
Num of cells

140

Num of cells
120 200 120
100 120
100
150 100
80 80
80
60 100 60 60
40 40 40
50
20 20 20
0 0 0 0
-2 0 2 4 6 8 10 12 14 0 5 10 15 20 25 -2 0 2 4 6 8 10 12 14 -2 0 2 4 6 8 10 12 14 16
Dev. progression Dev. progression Dev. progression Dev. progression

E F ATAC-seq, biological replicates

G ATAC-seq, GMP-A vs GMP-C
5 848-BCL11A-D
Mean TF z-score difference, K10-K9

11 11
4 R=0.91 R=0.86
1813-SPI1-D-N5

GMP-A, log2 fragments

10 10

GMP-C, log2 fragments

3 3478-STAT1-D-N1
2729-IRF1-D-N2 9 9
2 458-ATF3-D-N1
611-PRDM1-D-N4 8 8
558-CEBPD-D-N2
1 7 7
0 6 Density 6 Density
-1 303-TCF4-D-N3
5 5
2081-GATA1-D-N7
-2 747-CTCF-D-N67 4 4

-3 4 5 6 7 8 9 10 11 4 5 6 7 8 9 10 11
0 400 800 1200 1600 UNK, log2 fragments GMP-A, log2 fragments
Rank sorted TF motifs

H Scale
20 kb hg19
chr3:
128,190,000 128,200,000 128,210,000 128,220,000 128,230,000
500 _
K1-HSC
0_
500 _
K2-CMP
0_
500 _
K9-GMP
0_
500 _
K10-GMP
0_
500 _
K11-Mono
0_
500 _
Bulk HSC
0_
500 _
Bulk GMP-A
0_
500 _
Bulk GMP-B
0_
500 _
Bulk Mono
0_
RefSeq Genes
DNAJB8 GATA2

I J K
1200
GMP-A
GM
MP-A
M A (10.2%)
(10
( 2 2%) GMP
GMP-B
GMP
G M B (15.
(15.0%)
.0%)
..0 1000
Rank Sorted Cells

GMP-C
G - (8.9%)
(8 9%) UNK
800
0
GMP-A
600
PC3 (10.42%)

-2
0.8

GMP-B 400
-4
%)

Freq.

15 200
20
(45.5

-6 25 GMP-C
30
0

0
PC1

35
-12 -10 -8 -6 -4 -2 0 2 K2 K8 K9 K10 K11 K12 -2 0 2 4 6 8 10 12 14
PC2 (37.8%) CMP LMPP GMP GMP Mono pDC
Myeloid developmental progression
Clusters

L 848-BCL11A-D
M 558-CEBPD-D-N2
N 2081-GATA1-D-N7
12 12
5

8 8
0
TF z-score
TF z-score
TF z-score

4 4

-5
0 0

-10
-4 -4

-2 0 2 4 6 8 10 12 14 -2 0 2 4 6 8 10 12 14 -2 0 2 4 6 8 10 12 14
Dev. trajectory Dev. trajectory Dev. trajectory

Figure S4. Heterogeneity and Development Ordering of GMPs, Related to Figure 4

(A–D) Histogram denoting number of cells within each point in the (A) myeloid, (B) erythroid, (C) lymphoid and (D) pDC by pseudo-temporal developmental
ordering.
(legend continued on next page)
(E) TF motifs sorted by their rank difference in the mean TF z-score between GMP K10 and K9 clusters, important regulators are highlighted.
(F and G) Scatterplots colored by density denoting log2 fragments in individual peaks across samples for (F) UNK versus GMP-A and (G) GMP-A versus GMP-C
profiles.
(H) Genome browser track highlighting two differentially accessible regions surrounding the GATA2 gene.
(I) PCA projection of (light gray) GMP-A, (gray) GMP-B and (dark gray) GMP-C single-cell profiles.
(J) Frequency of single-cell profiles from differing GMP sorted immunophenotypes within previously defined data-driven epigenomic clusters.
(K) Single-cell profiles colored by their immunophenotypic cell type identity rank sorted by myeloid developmental progression.
(L–N) Single-cell myeloid progression and TF z-scores for (L) BCL11A, (M) CEBPD and (N) GATA1 motifs, smoothed motif accessibility trajectories are shown
in red.
A B C HSC CMP GMP Mono HSC CMP GMP Mono
1 Observed TFs 2
FDR <1% 6 clusters
Permuted NULL
1
0.8 1.5
Relative Density

0.8

Cluster Centroids
Gap Values
0.6 1
0.6
cluster cluster
0.4 1 3
0.5 2 4
0.4
5
0.2 6
0 0.2

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 2 4 6 8 10 12 14 16 18 20 -2 0 2 4 6 8 10 12 14 -2 0 2 4 6 8 10 12 14
Standard Dev. of TF scores Number of Clusters Developmental traj. (euc. distance) Developmental traj. (euc. distance)

D E F 11 Mono
Infer RNA expression Assign scRNA to scATAC cell 1

Observed CEBPD expression (log2)

2 10
R=0.996
UNK
3
9 GMP
Bulk RNA Step-wise 4
Bulk ATAC Filter for Filter cells with low 5 8
Regression 6
variable genes similarity values 7
7 LMPP

PC
8 6 pDC
Score 9 CMP
5
10
scATAC cells 11 4 MEP
CLP
Calculate Take max value to 12
13
3
scATAC x scRNA assign scRNA cell 14 MPP
2
Inferred expression 15
similarity (Pearson) to scATAC-seq profile 1 HSC
profiles -10 -5 0 5 10 15 1 2 3 4 5 6 7 8 9 10 11
CEBPD fit coefficients Fit CEBPD expression (log2)

G H I HSC
MPP
LMPP
CMP
GMP
MEP
CLP
pDC
Unknown J FACs isolated CMPs
Pearson Mono K2 K3 K4 K5
Inferred CEBPD expression 1 1.0 1.0
2

Fraction of CMPs in each cluster

1
-1 0.8 0.8
0
Fraction of cells
scRNA-seq

-1
0.6 0.6
-2
PC3

-3
0.4 0.4
-4
-5
0.2 0.2
-6
log2 expression
-7
-2 12
0.0 0.0
-8 P P o 4 4 8 7 7 q
C
-14 -12 -10 -8 -6 -4 -2 0 2 4 scATAC-seq HS CM GM Mon CD3 121 M082 M107 M113 NA-se
BM B B B cR
PC2 10X scRNA-seq sorted identity s
scATAC-seq
Norm. -1 1
K L expression M
0.6 HSC CMP GMP
K5
K3 R = 0.86
0.5
K4
K2 0.4
DPT order

304 variable genes 0.3

K5 0.2
K3
0.1
K4
K2 0

I1 X1 F4 1 2B F1 A1 E 2 C S1 O 4 A5 FF ID1 ID2 T3 74 T2 PB -2 0 2 4 6 8 10
FL PB P FPM GA KL AT PO PRG CL AL MP XCR OX MA FL CD BS EB
A ATAC:RNA pairing order
Z IT G LG C H C
ATAC:RNA pairing order

Smoothed
N O P HSC CMP GMP
scRNA-seq of HSCs scRNA-seq of HSCs HOXA9 CEBPD
Rel. expression

1
log2 expression

3000 ATAC:RNA pairing order 3000 DPT order 4 1

4
2500 R = 0.14469 2500 R = 0.6799 2 2
Genes detected
Genes detected

2000 2000 0 0 0 0
0 5 10 0 5 10
ATAC:RNA pairing order ATAC:RNA pairing order
1500 1500
HOXA9 CEBPD
log2 expression

Rel. expression

4 1 1
1000 1000
DPT order

Counts Counts 4
2500 2500
500 500 2 2

500 500 0 0 0 0
0 0
-2 0 2 4 6 8 10 0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.1 0.2 0.3 0.4 0.5 0.6
DPT order DPT order
ATAC:RNA pairing order DPT order

Figure S5. Chromatin Accessibility and Expression Dynamics across Myeloid Cell Differentiation, Related to Figure 5
(A) The standard deviation of smoothed TF accessibility scores, as shown above, for observed scores (blue) and cells randomly permuted (red), dotted line
represents an FDR less than 1%.
(B) Gap values for 1 to 20 k-medoids clusters, with optimal cluster number 6 highlighted. Error bars represent 1 SD on the within cluster dispersion.(C) Centroids
for each k-medoid TF cluster for (left) early and (right) late acting TFs, cells ordered by myeloid development (top) are shown for reference.

(legend continued on next page)

(D) Computational workflow for pairing scATAC-seq with scRNA-seq profiles using bulk data as reference.
(E) Stepwise regression fit coefficients linking the activity of scATAC-seq PCs to the expression of CEBPD.
(F) Fit and observed expression values (Log2) for bulk transcriptomes of sorted populations (R = 0.996).
(G) PCA projection of scATAC-seq cell profiles colored by inferred expression of CEBPD.
(H) Hierarchical clustering of scATAC-seq by scRNA-seq profiles, values represent Pearson correlation coefficient (see STAR Methods).
(I and J) Fraction of cell assignments for (I) scRNA-seq profiles to immunophenotypically defined cell types in scATAC-seq data and (J) CMP scATAC-seq of
different donors or scRNA-seq profiles to defined clusters as shown in Figure 3B.
(K) scRNA-seq counts of CMP cells matching to cluster 2 or cluster 3 scATAC-seq cells.
(L) Hierarchical clustering of genes for scRNA-seq CMPs matched to the four scATAC-seq defined CMP clusters showing (top) significantly variable genes or
(bottom) known marker genes.
(M) Correlation of transcriptome profiles ordered by their developmental trajectory, comparing scATAC and scRNA pairing (as described in the main text) with
diffusion pseudo-time (DPT) of cells (HSCs in green, CMP in yellow and GMP in orange).
(N and O) Pseudotime order and number of genes detected for HSCs using (N) ATAC:RNA pairing and (O) DPT, cells are colored by the number of genes detected.
(P) Log2 mean expression profiles for TFs (left) HOXA9 and (right) CEBPD across myeloid pseudo-time determined using (top) ATAC:RNA pairing or (bottom) DPT,
cells are colored by their sorted identity and smoothed profiles are shown in red.
 max activity
A C k-mediods (k=20) peaks
D HSC mono Transition elements
0.8 peaks 80
TAL1;-95 per K Accessibility 0 Max
GATA2;4567
77 Reactivation
0.6 318 40
Fragments per cell

329 Reactivation

HSC specific

Dim. 2
0.4 404 0
298
338
0.2 -40
203
132
0 -80
66
-2 0 2 4 6 8 10 12 -100 -50 0 50 100
Developmental traj. (euc. distance) 55 Dim. 1
B E

Transition
0.3 78
IL7;-364457
62 80 Mono. activity
65 0 1
Fragments per cell

0.2 58 40
78

Dim. 2
86
0

Monocyte specific
0.1 156
248
-40
285
0 67
-2 0 2 4 6 8 10 12 14
Developmental traj. (euc. distance) -2 0 2 4 6 8 10 12 14 -80
Developmental traj. (euc. distance) -100 -50 0 50 100
Dim. 1

F 2081-GATA1-D-N7 2288-HOXB8-D 428-CEBPG-D-N2 747-CTCF-D-N67 18376-RORB-I-N3

Collect raw
PWM scores
per peak

Compute PWM
peak-peak log-score 3 8
distance
2081-GATA1-D-N7 848-BCL11A-D 1880-SPIB-D-N2 216-TCF12-D-N3 3073-ESRRA-D-N4
For each peak
average over
K-nearest (300)

Local
PWM score min max
G I rs1978487: C->T
0.5 100 kb hg19
31,150,000 31,200,000 31,250,000 31,300,000 31,350,000 31,400,000 31,450,000 31,500,000
Fraction of correlated peaks

K1-HSC
0.45

K2-CMP
0.4

K9-GMP
0.35

K10-GMP
0.3

K11-Mono
0.25
CD34+_HSPCs
0.2 s s
pairs ce loops ce loop ce loop BCKDK PRSS36 FUS PYDC1 ITGAM ITGAX ZNF843 C16orf58
ene n en en
eak-g w evide ed. evid igh evid 1
PRSS8 PYCARD ITGAD TGFB1I1
Pearson

All p lo m h KAT8 TRIM72 COX6A2 SLC5A2

0.5
H Expected peak-gene pairs
<0.5
7
Background peaks-gene pairs
6

5
Rel. density

0
-1.5Mb 0 1.5Mb
Distance of cis-eQTL peak to predicted target gene

Figure S6. Association of TFs and Genetic Variants to Regulatory Element Dynamics across Myeloid Differentiation, Related to Figure 6
(A and B) Example cis-regulatory dynamics across myeloid development for (A) HSC active peaks and (B) a reactivated peak near IL7.
(C) Hierarchical clustering of k-medoids centroids across all peaks, values (left) denote number of peaks per cluster and labels (right) denote categorization into
different global patterns.

(legend continued on next page)

(D and E) t-SNE plots of dynamic enhancer profiles highlighting (D) reactivation or transition elements, points are colored by the max accessibility of the element in
myeloid pseudo-time, and (E) colored by accessibility in monocytes.
(F) (left) Schematic for determining motif enrichment across similar peak profiles, (right) raw log PWM score or averaged (see STAR Methods) local PWM score for
a representative subset of dynamic TF motifs across myeloid development for the t-SNE embedding shown in Figure 5.
(G) Fraction of correlated peaks (R > 0.5) as a function of loop confidence. Error bars represent 1 standard deviation on the estimate of the mean.
(H) Distribution of distances of cis-eQTLs (purple) or background peak-gene pairs (gray) to the predicted target gene.
(I) Regulatory profiles surrounding the ITGAX gene (also known as CD11c), dynamic enhancers are highlighted in gray with significant (blue) and non-significant
(gray) correlated peak-gene pairs shown as loops, cis-eQTL rs1978487 is highlighted.