Systematic and Applied Microbiology 40 (2017) 492–499

Contents lists available at ScienceDirect

Systematic and Applied Microbiology

journal homepage: www.elsevier.de/syapm

Assessing the impact of rumen microbial communities on methane

emissions and production traits in Holstein cows in a tropical climate
Camila S. Cunha a , Cristina M. Veloso a , Marcos I. Marcondes a , Hilario C. Mantovani b ,
Thierry R. Tomich c , Luiz Gustavo R. Pereira c , Matheus F.L. Ferreira a ,
Kimberly A. Dill-McFarland d,e,∗ , Garret Suen d,∗
a
Department of Animal Science, Universidade Federal de Viçosa, Peter Henry Rolfs Avenue, University Campus, Viçosa, Minas Gerais 36570-900,Brazil
b
Department of Microbiology, Universidade Federal de Viçosa, Peter Henry Rolfs Avenue, University Campus, Viçosa, Minas Gerais 36570-900, Brazil
c
Brazilian Agricultural Research Corporation, Embrapa Dairy Cattle, Eugênio do Nascimento Avenue, 610, Cascatinha, Juiz de Fora, Minas Gerais
36038-330, Brazil
d
Department of Bacteriology, University of Wisconsin-Madison, 1550 Linden Dr, Madison, Wisconsin 53706, USA
e
Department of Microbiology & Immunology, Life Sciences Institute, University of British Columbia, 2350 Health Sciences Mall, Vancouver, British
Columbia V6T 1Z3, Canada

a r t i c l e i n f o a b s t r a c t

Article history: The evaluation of how the gut microbiota affects both methane emissions and animal production is nec-
Received 28 May 2017 essary in order to achieve methane mitigation without production losses. Toward this goal, the aim of
Received in revised form 27 July 2017 this study was to correlate the rumen microbial communities (bacteria, archaea, and fungi) of high (HP),
Accepted 31 July 2017
medium (MP), and low milk producing (LP), as well as dry (DC), Holstein dairy cows in an actual tropical
production system with methane emissions and animal production traits. Overall, DC cows emitted more
Keywords:
methane, followed by MP, HP and LP cows, although HP and LP cow emissions were similar. Using next-
Methane emissions
generation sequencing, it was found that bacteria affiliated with Christensenellaceae, Mogibacteriaceae,
Milk production
Rumen microbiology
S24-7, Butyrivibrio, Schwartzia, and Treponema were negatively correlated with methane emissions and
Bacteria showed positive correlations with digestible dry matter intake (dDMI) and digestible organic matter
Archaea intake (dOMI). Similar findings were observed for archaea in the genus Methanosphaera. The bacte-
Fungi rial groups Coriobacteriaceae, RFP12, and Clostridium were negatively correlated with methane, but did
not correlate with dDMI and dOMI. For anaerobic fungal communities, no significant correlations with
methane or animal production traits were found. Based on these findings, it is suggested that manipula-
tion of the abundances of these microbial taxa may be useful for modulating methane emissions without
negatively affecting animal production.
© 2017 Elsevier GmbH. All rights reserved.

Introduction
The rumen is inhabited by a diversity of microorganisms, includ-
ing bacteria, archaea, and fungi. These microbes interact with each
other to coordinate the breakdown and fermentation of differ-
ent feed components ingested by the ruminant host. In particular,
∗ Corresponding authors at: Department of Bacteriology, University of Wisconsin-
Madison 5159 MSB, 1550 Linden Drive, Madison, WI, 53706, USA.
E-mail addresses: camilasoares4@yahoo.com.br (C.S. Cunha),
cristina.veloso@ufv.br (C.M. Veloso), marcos.marcondes@ufv.br (M.I. Marcondes),
hcm6@ufv.br (H.C. Mantovani), thierry.tomich@embrapa.br (T.R. Tomich),
luiz.gustavo@embrapa.br (L.G.R. Pereira), matheus.fellipe234@gmail.com
(M.F.L. Ferreira), dillmcfarlan@wisc.edu (K.A. Dill-McFarland), gsuen@wisc.edu
(G. Suen).
fungi and bacteria degrade host-indigestible fiber, and the result-
ing breakdown products are fermented mainly by bacteria into
important nutrients, such as volatile fatty acids (VFAs) [4,19]. Thus,
the rumen microbiota is responsible for producing the nutrients
used by the host for growth, reproduction, and ultimately, the gen-
eration of agriculturally-relevant products like milk and meat. In
addition to the production of nutrients such as VFAs, byproducts of
fiber degradation include hydrogen (H2 ) and carbon dioxide (CO2 ),
which are utilized by methanogenic archaea to produce methane
(CH4 ) [43]. The release of methane not only deprives the host of
important carbon resources [17] but also significantly contributes
to greenhouse gas emissions and global warming [15].
The diet is influenced by the type of feed [31] and quality of
forage [29], which are primary factors in modulating the avail-
ability of hydrogen used by methanogens in the rumen. However,

http://dx.doi.org/10.1016/j.syapm.2017.07.008
0723-2020/© 2017 Elsevier GmbH. All rights reserved.
 C.S. Cunha et al. / Systematic and Applied Microbiology 40 (2017) 492–499
dietary strategies for mitigating methane emissions are costly, and
different feeds and forages are often used according to milk pro-
duction levels (high, low, dry) to meet the dairy cow’s nutritional
requirements at lower costs [44]. Considering this, there is a need
for studies that mirror what occurs on commercial dairy farms in
tropical conditions, since feeding animals at different diets accord-
ing to the lactation stage and milk production levels is a common
practice. For example, cheaper, more fibrous diets fed to dry and
low-milk producing cows lead to more residual hydrogen because
of the greater amounts of acetate produced, while more expensive,
high-milk producing diets based on high starch content induce a
greater production of propionate, thereby lowering the produc-
tion of hydrogen and methane in the rumen [4]. Differences in
methane production due to diet are further impacted by climate
and geography, as forages in tropical conditions often present lower
digestibility and their degradation in the rumen results in a higher
proportion of acetate and methane [10,27]. Thus, adjusted methane
mitigation practices may be necessary for animals with different
production levels fed different diets in different environments.
The composition of the rumen microbial community is also
thought to contribute significantly to methane production. There
are a number of studies that have focused on the relationship
between methane emissions and the rumen bacterial or archaeal
microbiota, since these communities contribute directly to hydro-
gen or methane production in the rumen [8,38,48]. However, there
is little information regarding how the overall community of bacte-
ria, archaea, and fungi in the same animal contributes to methane
emissions, and this knowledge could be used to modulate methane
emissions through manipulation of the ruminal microbiota. This is
of particular importance as direct modulation of archaeal popula-
tions to reduce CH4 emissions is difficult and can result in decreased
feed digestion due to ruminal hydrogen accumulation [28].
Thus, feed degradation and methanogenesis in the rumen are
dependent on a combination of diet, the composition of the rumen
microbiota [7], and the host [37]. However, it remains unclear
which attributes (diet, microbiota, genetics, etc.) result in the com-
bination of high production with low methane emissions in the
same animal. Therefore, it was hypothesized for this study that
differences in bacterial, archaeal and/or fungal community com-
position would contribute to differences in the methane emissions
of Holstein cows under a feeding management strategy based on
milk production level. To address this, the relationship between
the overall ruminal microbial community, enteric methane pro-
duction, and milk production traits in dairy cows was investigated
in a tropical production system.
Materials and methods
Animals and management
All experiments were conducted at the Unit for Teaching,
Research, and Extension on Dairy Cattle (UEPE-GL), Universidade
Federal de Viçosa (UFV), Viçosa, Minas Gerais, Brazil. Microbiolog-
ical analyses were performed at the Department of Bacteriology,
University of Wisconsin-Madison, Madison, WI, USA. The project
was approved by the Ethics Committee in the Use of Production
Animals from the UFV (protocol number 99/2014).
In a functioning dairy system at UEPE-GL, five dry cows (DC,
746 ± 38.61 kg; 5.24 ± 1.84 years) and 14 lactating cows were
selected according to their milk production. Lactating cows were
split into high-producing (HP, 25 ± 2.74 L milk d−1 ; 540 ± 70.26 kg;
2.67 ± 0.67 years; n = 5), medium-producing (MP, 15 ± 0.85 L milk
d−1 ; 611 ± 57.29 kg; 4.41 ± 1.20 years; n = 5) or low-producing
(LP, 10 ± 0.76 L milk d−1 ; 624 ± 122.85 kg; 3.91 ± 1.46 years; n = 4)
groups. Each of the four cohorts was measured and sampled
493
sequentially over a unique 10-day period, due to limited methane
emission equipment.
All cows were housed in individual tie stalls and kept under
the feeding management protocol they were already adapted to in
the production system. These animals were not maintained on the
same diet because we purposefully wanted to mirror what occurred
on an actual dairy farm in tropical conditions. All diets consisted
of corn silage and concentrate comprising corn, soybean meal and
minerals (Table S1). The forage:concentrate ratios were 50:50 for
HP cows, 70:30 for the MP cows, 80:20 for LP cows, and 90:10 for
DC cows. High-producing cows also received Coastcross grass (Cyn-
odon sp.) hay in the proportion of 8.5% of the total dry matter (DM)
(41.5% of corn silage). Animals were fed twice a day, at 7:00 am and
4:00 pm, and water was available ad libitum to all animals. Cows
in lactation were mechanically milked twice a day, at 6:00 am and
3:00 pm.
Methane emission measurements
The trial for each production group was performed separately
for a 10-day period. Animals were adapted to tie stall housing and
yokes from days 1 to 5. The adaptation period was appropriate
since slight interferences of feed intake were observed only in the
first two days of this period. Gas collections from animal eructa-
tion for evaluation of methane (CH4 ) emissions were performed
on days 6–10 in 24 h periods. In situations where problems were
encountered in the sampling process, the collection for that day
was repeated after day 10.
Animal methane emission measurements were performed by
the tracer gas sulfur hexafluoride (SF6 ) technique [18] with the
adaptations described in Oss et al. [34]. Briefly, the flow of CH4
released by the animal was calculated in relation to the flow of
SF6 , measured from the SF6 release rate from a permeation capsule
lodged in the rumen and from the concentrations of CH4 and SF6 in
gas samples [16]. Analysis of CH4 and SF6 levels were performed at
the Laboratory of Gas Chromatography at the Empresa Brasileira de
Pesquisa Agropecuária (EMBRAPA) Dairy Cattle, Juiz de Fora, Minas
Gerais, Brazil. The concentrations of CH4 and SF6 in the samples
were evaluated by gas chromatograph, as described in Schloss et al.
[40]. Raw CH4 emissions were corrected for digestible fractions
of dry matter intake (dDMI) and digestible organic matter intake
(dOMI) by dividing raw values by dDMI and dOMI (Table 1 and
Table S3). As these characteristics were highly variable between
animals, the corrections were required to make the comparison
between animals more reliable, considering that microorganisms
would degrade only a fraction of ingested food.
Food, refusals and feces sampling
From days 6 to 10, forage and refusals from each animal were
sampled. The quantities of food provided and the weight of the
refusals were recorded. Concentrate ingredients were sampled dur-
ing mixing. At day 6, feces were collected during a 24-h period,
weighed, and a subsample was then taken from this collection.
The food, refusal and feces samples of each animal were grouped
per trial and analyzed for dry matter (DM), according to Detmann
et al. [11] method INCT-CA n◦ G-003/1, as well as the crude protein
(CP), ether extract (EE), neutral detergent fiber (NDF), non-fiber car-
bohydrates (NFC) and ash (MM) contents [2]. These results were
used for the calculation of the dry matter intake (DMI), dry mat-
ter digestibility (DMD), organic matter intake (OMI) and organic
matter digestibility (OMD) (Table S3) [3,24].
494 C.S. Cunha et al. / Systematic and Applied Microbiology 40 (2017) 492–499
Table 1
Means and standard deviations for methane emissions and production traits of Holstein cows.
Group CH4 g day−1 CH4 g kg−1 DMI CH4 g kg−1 dDMI CH4 g kg−1 dOMI
HP 286.22 ± 67.43B 19.13 ± 4.16B 21.70 ± 4.68C 23.60 ± 5.04C
MP 369.92 ± 41.24A 23.08 ± 2.21AB 29.25 ± 3.07B 31.44 ± 3.43B
LP 174.68 ± 31.80C 13.29 ± 2.11C 19.52 ± 3.47C 20.64 ± 3.38C
DC 277.35 ± 22.22B 25.09 ± 3.81A 38.10 ± 5.79A 40.09 ± 6.26A
CH4 = methane; DMI = dry matter intake; DMD = dry matter digestibility; dDMI = digestible dry matter intake; dOMI = digestible organic matter intake; MY = milk yield;
HP = high-producing (n = 5); MP = medium-producing (n = 5); LP = low-producing (n = 4); DC = dry cows (n = 5). Means followed by different letters in a column are different
by Tukey’s test (P < 0.10).
Ruminal sampling and fermentation pattern
Ruminal contents were sampled 4 h after morning feeding using
the stomach tube technique [13,23,26]. The initial 200 mL of col-
lected fluid were discarded to avoid saliva contamination, and
rumen samples therefore represented only the liquid phase of
ruminal content. A sample of approximately 70 mL was filtered
through four layers of cheesecloth and split into two aliquots. The
first one was used for DNA extraction (50 mL), and the other (20 mL)
for volatile fatty acid (VFA) concentration analysis. Both samples
were frozen at −80 ◦ C until they were analyzed.
VFA concentrations were analyzed using high-performance liq-
uid chromatography (HPLC). Briefly, after a centrifugation step,
1.5 mL of a cell-free supernatant was treated as described by
Siegfried et al. [41] and analyzed in a Dionex Ultimate 3000 Dual
chromatograph (Dionex Corporation, Sunnyvale, CA, USA) system
with a Shodex RI-101 refractive index detector at 45 ◦ C, using a
Phenomenex Rezex ROA, 300 × 7.8 mm ion-exclusion column, also
at 45 ◦ C. The mobile phase was H2 SO4 (5 mM) at a flow rate of
0.7 mL min−1 .
DNA extraction, amplicon library preparation and sequencing
DNA extractions were performed on 25 mL rumen samples, as
described by Stevenson and Weimer [42], and then treated with
RNAse Ribonuclease A (10 mg mL−1 ). Total DNA was quantified
®
(Table S2) using a Qubit 2.0 Fluoremeter (Invitrogen, San Diego,
CA, USA). A two-step PCR was employed to amplify the V3 and V4
regions of the 16S ribosomal RNA gene (16S rRNA) for bacteria, the
V6 to V8 regions of 16S rRNA for archaea, and the ITS1 gene for fungi
[20], which corresponded to amplicon lengths of 571 bp, 608 bp and
366 bp, respectively.
The first PCR aimed to amplify the region of interest using spe-
cific primers attached to Illumina sequencing primers, and the
second one to attach dual indices and Illumina sequencing adapters.
The reactions were amplified under the following conditions: 95 ◦ C
for 3 min, 95 ◦ C for 30 s, 55 ◦ C for 30 s, 72 ◦ C for 30 s and 72 ◦ C for
5 min. A total of 25 cycles were performed for bacterial and archaeal
amplicons, 35 cycles for fungal, and eight cycles for the second
PCR for all amplicons. The first PCR was performed using 5–200 ng
DNA, 0.4 ␮M of each primer, 1X KAPA HiFi HotStart ReadyMix (Kapa
®
Biosystems , Wilmington, MA, United States) and ultrapure water
to a total of 25 ␮L. This PCR product was purified using an Invitro-
gen PureLink Pro96 PCR Purification Kit and 5 ␮L of product was
used in the second PCR. The products of the second PCR were
run on a 1% AquaPor LM low-melt agarose gel and the gel extrac-
tion was performed using the ZR-96 Zymoclean Gel DNA Recovery
Kit (Zymo Research, Irvine, CA). DNA was then quantified using a
®
Qubit 2.0 Fluoremeter, diluted, and equimolar pooled to 2 nM. The
library was denatured at 96 ◦ C using 0.2 N NaOH, then diluted with
hybridization buffer to 20 pM, and combined with the PhiX con-
trol library (20 pM). The library was loaded into an Illumina MiSeq
cartridge and sequenced on an Illumina MiSeq using a 2 × 300 bp
DMI (kg) DMD dDMI (kg) dOMI (kg) MY (L day−1 )
15.15 ± 2.79B 0.76 ± 0.02A 13.39 ± 2.66A 12.31 ± 2.40A 25.86 ± 6.73A
16.09 ± 1.99A 0.71 ± 0.02B 12.72 ± 1.70A 11.84 ± 1.63A 15.54 ± 1.91B
13.22 ± 2.05BC 0.68 ± 0.05C 9.07 ± 1.88B 8.56 ± 1.76B 10.55 ± 1.51C
11.16 ± 0.03C 0.66 ± 0.03C 7.35 ± 0.62B 6.99 ± 0.62B NA
paired-end sequencing kit (v3) according to the manufacturer’s
guidelines.
DNA sequence processing analysis
The program mothur, version 1.36.1 [40], was used to analyze
the sequenced libraries following a protocol adapted from [22].
Sequences that could not be aligned were chimeric, and all con-
taminants were removed. The sequences from bacteria and archaea
were aligned against the SILVA 16S rRNA gene database [35] and
classified using the Greengenes database [9]. For anaerobic fungal
analysis, sequences were classified using the UNITE database [21]
and contaminants were also removed. A sequence similarity of 97%
was used as a cut-off to classify operational taxonomic units (OTUs)
at the species level [39]. Taxonomic assignment of OTUs was also
performed with mothur. Microbial communities were normalized
to equal sequence counts before further analyses. All sequences
used for this study were publicly deposited in the NCBI Short Read
Archive under BioProject PRJNA 380769.
Statistical analysis
Relative abundances were calculated as the number of
sequences of the taxa divided by the number of total sequences
present in the sample multiplied by 100. Alpha diversity indices
for richness (Chao1) and diversity (Shannon), as well as Good’s
coverage, were calculated in mothur.
The program R [36] was used to test differences in CH4 emis-
sions from each group using analysis of variance (ANOVA), followed
by Tukey’s honest significance test (TukeyHSD) for multiple com-
parisons, when necessary. Differences in richness and diversity of
bacterial, archaeal and anaerobic fungal communities were also
tested with ANOVA and TukeyHSD. For these tests, P < 0.10 was
accepted as significant due to the high variability of the methane
measurement technique [46].
Bacterial, archaeal and anaerobic fungal communities were
visualized using non-metric multidimensional scaling (nMDS)
plots of the Bray-Curtis dissimilarity index in the vegan package
[33]. Analysis of similarity (ANOSIM) [5] was performed to deter-
mine differences in the overall microbial community structure
(Bray–Curtis) and composition (Jaccard) between groups (HP, MP,
LP and DC). Similarity percentage analysis (SIMPER) [6] was used
to evaluate the contribution of each OTU to the differences seen in
the ANOSIM. The relative abundances of OTUs identified by SIMPER
(at least 3% for bacteria and 8% for fungi) were compared using the
Kruskal–Wallis non-parametric test (Agricolae package) [30]. Other
differences in relative abundance of interest were also tested using
Kruskal–Wallis. For these tests, P < 0.05 was considered significant
and P < 0.10 was a trend.
To investigate if microbial community data and VFA concen-
trations co-varied, a Mantel test [27] was performed. Mantel was
also used to evaluate if bacterial, archaeal and anaerobic fungal
communities co-varied. All these analyses were performed with
the aid of the vegan package. For these analyses, P < 0.05 was con-
sidered significant. Spearman’s correlation was used to correlate
 C.S. Cunha et al. / Systematic and Applied Microbiology 40 (2017) 492–499
Table 2
Means and standard deviations for volatile fatty acids in the rumen fluid of Holstein cows.
Group Total VFA (mmol L−1 ) Acetate (mmol L−1 )
HP 124.47 ± 11.20A 72.47 ± 4.61A
MP 108.27 ± 11.80AB 68.39 ± 7.45A
LP 75.51 ± 21.84C 57.77 ± 5.89B
DC 101.33 ± 3.60B 65.87 ± 1.90AB
VFA = volatile fatty acid (acetate, propionate, butyrate, lactate, isobutyrate, isovalerate, valerate); A:P = acetate:propionate ratio. HP = high-producing; MP = medium-
producing; LP = low-producing; DC = dry cows. Values are mean ± standard deviation. Means followed by different letters in a column are different by Tukey’s test (P < 0.10).
continuous variables (CH4 emissions, dDMI, dOMI) with bacterial
families and genera, as well as archaeal and fungal genera rela-
tive abundances, using the corrplot package [50]. Only correlations
significant at P < 0.05 were added to the resultant correlogram.
Results
Methane emissions and physiological traits related to milk
production
Raw CH4 emissions (Table 1) differed by production group
(ANOVA, P < 0.01) and were highest in medium-producing cows
(MP, TukeyHSD P < 0.05) (Table S4). High-producing cows (HP) and
dry cows (DC) had similar emission levels, but they were less than
MP cows and low-producing cows (LP) (TukeyHSD P < 0.05). When
CH4 emissions were corrected for digestible dry matter intake
(dDMI) and digestible organic matter intake (dOMI) (Table S4), DC
had the highest CH4 emissions (TukeyHSD P < 0.05), followed by
MP, and HP/LP cows. MP was greater than LP (TukeyHSD P < 0.05)
but a trend was only observed when compared to HP (TukeyHSD
P < 0.10). When comparing the dDMI and dOMI, HP and MP cows
were similar (TukeyHSD P > 0.10) but higher than LP and DC cows
(TukeyHSD P < 0.05, Table 1 and Table S4).
VFA concentrations in rumen fluid (Table 2 and Table S4) were
lower for LP (TukeyHSD P < 0.05) cows, and HP cows had the highest
propionate concentration (TukeyHSD P < 0.05), which led to a lower
acetate:propionate ratio (TukeyHSD P < 0.05).
Rumen liquid fraction microbiota characterization
After all filtering analyses, the number of sequences obtained
was 214,676 for bacteria (mean = 11,298 ± 5109), 64,774 for
archaea (mean = 3409 ± 1555) and 12,544 for anaerobic fungi
(mean = 660 ± 505). For all amplicons, Good’s coverage of all sam-
ples was >0.98.
The most abundant bacterial phyla within all samples included
the Firmicutes (49.18 ± 6.47%), Bacteroidetes (31.55 ± 9.07%), and
TM7 (8.36 ± 8.21%) (Fig. 1A). Unidentified sequences accounted
for 0.37 ± 0.32% of the total sequence counts. Among the Firmi-
cutes, the genera Ruminococcus, Butyrivibrio and Coprococcus were
the most abundant in all animals (8.25 ± 5.34%, 6.97 ± 3.75% and
3.46 ± 2.28%). Within the Bacteroidetes, the genus Prevotella had
the highest relative abundance at 62.48 ± 19.88%, followed by
CF231 in the family Paraprevotellaceae with a relative abundance
of 2.72 ± 1.02%.
Methanobrevibacter was the most abundant genus in the
archaeal community (93.43 ± 3.18%), followed by Methanosphaera
(5.32 ± 2.84%) and Vadin CA11 (1.23 ± 1.56%) (Fig. 1B). For the
anaerobic fungal community, 73.19 ± 19.99% of the sequences were
unclassified at the phylum level, 21.95 ± 17.40% were assigned
to the genus Caecomyces, 4.16 ± 7.46% to genus Orpinomyces and
0.70 ± 1.30% to genus Neocallimastix (Fig. 1C).
A total of 32 bacterial OTUs were found to be ubiquitous across
all samples in this study and altogether represented 3.14% of the
relative abundance of the entire dataset. Of these, 21 belonged to
495
Propionate (mmol L−1 ) Butyrate (mmol L−1 ) A:P
31.90 ± 5.38A 12.15 ± 1.15 2.31 ± 0.26B
22.22 ± 1.67B 10.76 ± 1.86 3.07 ± 0.13A
19.04 ± 2.52B 12.29 ± 4.96 3.10 ± 0.70A
20.98 ± 1.42B 8.98 ± 0.29 3.15 ± 0.20A
phylum Firmicutes, eight to Bacteroidetes, two to TM7 and one to
Chloroflexi. At the genus level, the most abundant representatives
of this core bacterial microbiota included Prevotella, Ruminococcus,
Butyrivibrio, Clostridium, Coprococcus, Shuttleworthia, CF231 and
SHD-231. Within the archaeal community, two OTUs belonging to
the genera Methanobrevibacter and Methanosphaera were present
in all samples. None of the anaerobic fungi OTUs were present in
all samples.
Differences of rumen bacterial, archaeal and fungal diversity and
composition related to milk production groups
The number of species and the distribution of taxa across all
samples were evaluated using the Chao1 index for richness and
Shannon’s index for diversity. Bacterial richness (ANOVA P = 0.04)
but not diversity (ANOVA P = 0.26) varied by production group, with
LP cows showing higher richness (501.85 ± 58.44) than DC cows
(366.22 ± 79.86, TukeyHSD P = 0.02) (Table S5). The richness and
diversity of the archaeal and anaerobic fungal communities were
similar between groups (ANOVA P > 0.05).
Physiological and production traits were significant in cluster-
ing animals only in the bacterial and anaerobic fungal communities
(Fig. 2). Cows were observed to have similar bacterial community
composition (Jaccard, ANOSIM P = 0.58) but different bacterial com-
munity structure (Bray–Curtis, ANOSIM P = 0.02) (Table S6). Several
OTUs were found to influence beta-diversity (Tables S7 and S8), and
OTUs within the order Clostridiales were mainly responsible for the
observed differences (SIMPER, Table S7).
Archaeal community composition (Jaccard, ANOSIM P = 0.40)
and structure (Bray-Curtis, ANOSIM P = 0.23) did not differ between
groups (Table S6). Similarly, no differences were observed between
groups for the anaerobic fungal community composition (Jaccard,
ANOSIM P = 0.40), but the community structure was influenced by
the production group (Bray–Curtis, ANOSIM P = 0.01) (Table S6).
Differences in the relative abundances for specific taxa were
also determined, as shown in Table 3 and Table S9. The rela-
tive abundance of Firmicutes varied largely among the production
groups but the relative abundances of Bacteroidetes did not dif-
fer between groups (Kruskal–Wallis P > 0.05). DC cows had a
greater relative abundance of Methanobrevibacter (95.75 ± 1.89%)
and lower abundance of Methanosphaera (3.01 ± 1.47%) than HP
cows (90.04 ± 3.84%, 7.79 ± 3.81%, respectively, Kruskal–Wallis
P < 0.05). The relative abundances of the main fungal genera were
similar between production groups (Kruskal–Wallis P > 0.05).
Correlation between bacterial, archaeal and anaerobic fungal taxa
and physiological traits
The strongest negative correlations found between CH4 emis-
sions and bacteria were for sequences belonging to the family
Coriobacteriaceae (r = −0.86, Spearman’s test P = 0.01), as well as
the genera Butyrivibrio (r = −0.65, Spearman’s test P = 0.01) and
Schwartzia (r = −0.62, Spearman’s test P = 0.01) (Table S10). Fam-
ilies RF16 and Succinivibrionaceae showed positive correlations of
0.49 to 0.60 (Spearman’s test P < 0.05) with CH4 emissions, while
496 C.S. Cunha et al. / Systematic and Applied Microbiology 40 (2017) 492–499

Fig. 1. Percentage relative abundance of bacterial phyla (A), archaeal genera (B) and anaerobic fungal genera (C) in the rumen of Holstein cows at different production levels.
HP = high-producing; MP = medium-producing; LP = low-producing; DC = dry cows.

Table 3
Means and standard deviations for relative abundances of ruminal microbial groups related to production groups.

Microbial taxa HP MP LP DC

Firmicutes 47.87 ± 4.21B 50.62 ± 3.97A 56.38 ± 5.80A 42.15 ± 5.61B

Bacteroidetes 33.19 ± 10.45 34.10 ± 11.79 26.81 ± 7.53 31.15 ± 6.70
Coriobacteriaceae 1.10 ± 1.53A 0.32 ± 0.21B 0.87 ± 0.41B 0.12 ± 0.23C
Succinivibrionaceae 0.52 ± 0.51B 1.68 ± 2.46AB 0.25 ± 0.23B 3.63 ± 2.09A
OTU08 Ruminococcaceae 0.55 ± 0.18C 4.89 ± 2.38A 4.46 ± 2.45A 1.79 ± 1.73B
Methanobrevibacter 90.04 ± 3.84A 94.54 ± 0.74AB 94.07 ± 2.52AB 95.75 ± 1.89B
Methanosphaera 7.79 ± 3.81A 4.85 ± 0.75B 5.75 ± 2.39B 3.01 ± 1.47B
Vadin CA11 2.16 ± 2.60 1.13 ± 0.58 0.18 ± 0.17 1.24 ± 1.26
Caecomyces 23.85 ± 22.66 32.12 ± 14.76 16.17 ± 16.69 14.51 ± 13.59
Orpinomyces 0.38 ± 0.84 1.5 ± 1.55 7.08 ± 6.71 8.25 ± 12.53
Neocallimastix 1.54 ± 1.63 0.74 ± 1.65 0.00 ± 0.00 0.37 ± 0.82
OTU0003 Neocallimastigaceae 28.3 ± 11.57A 34 ± 18.47A 6.67 ± 8.68B 6.06 ± 6.77B

HP = high-producing; MP = medium-producing; LP = low-producing; DC = dry cows. Means followed by different letters in a row are different by the Kruskal-Wallis test
(P < 0.10).

Methanobrevibacter showed weak correlations with this variable The families Christensenellaceae, Mogibacteriaceae and S24-7, as
(r = 0.35, Spearman’s test P > 0.05) (Fig. 3; Table S10). well as the genera Butyrivibrio, Schwartzia and Treponema, were
negatively correlated with CH4 emissions, with correlations rang-
 C.S. Cunha et al. / Systematic and Applied Microbiology 40 (2017) 492–499 497

Fig. 2. Non-metric multidimensional scaling plot of Bray-Curtis dissimilarity indices of bacterial (A) and anaerobic fungal (B) community structure. Variables that had a
significant fit in envfit were fitted to the x-y plane as arrows. HP = high-producing; MP = medium-producing; LP = low-producing; DC = dry cows. CH4 = methane emissions;
dDMI = digestible dry matter intake; dOMI = digestible organic matter intake; A:P = acetate to propionate ratio.

Fig. 3. Correlation analysis between CH4 emissions, production parameters, and relative abundances of ruminal microbial taxa at the family and genus taxonomic levels.
Spearman’s linear correlation values for families/genera across all samples. Only significant correlations (P < 0.05) for at least one of the analyzed variables are shown. Names
in bold indicate families, whereas non-bold names indicate genera. Names in black and green indicate bacterial and archaeal taxa, respectively. CH4 = methane emissions;
dDMI = digestible dry matter intake; dOMI = digestible organic matter intake.

ing from −0.58 to −0.73 (Spearman’s test P < 0.01), but positively
correlated with the digestible fractions of DMI and OMI (r = 0.52
to 0.73, Spearman’s test P < 0.05) (Table S10). OTUs affiliated to
Coriobacteriaceae, RFP12, Clostridium, and Methanosphaera were
negatively correlated with CH4 emissions, with correlations rang-
ing from −0.86 to −0.55 (Spearman’s test P < 0.05) but did not
show significant correlations with dDMI and dOMI (Spearman’s test
P > 0.05). Correlations between CH4 emissions corrected for dDMI
or dOMI and anaerobic fungi were not found to be significant.
Finally, none of the bacteria, archaea or anaerobic fungi that cor-
related with CH4 emissions showed significant correlations (Mantel
test, P > 0.01) with total VFAs or the individual concentrations of
acetate, propionate or butyrate.
Discussion
In this study, the liquid fraction of the rumen microbiota of
Holstein cows and its relationship to both CH4 emissions and
production traits in a functioning, tropical dairy system were inves-
tigated. Since the aim was to model commercial dairy farms, the
feed strategy used reproduced common practices on tropical farms.
A recent study [7] showed that differences in methane emissions
from cows at similar physiological stages could be linked to dif-
ferences in the ruminal archaeal and bacterial communities. In the
present study, animals were considered at different physiological
states, including dry cows, and the focus was on archaea, in addi-
tion to bacteria and fungi, given their role in mediating the methane
production precursors used by archaea. The study sought to deter-
mine the interplay between the microbiota, methane and animal
production, since efforts to decrease methane have often resulted
in decreases in animal production. Understanding this complex
microbiota could lead to novel ways for mitigating methane pro-
duction without the concomitant losses in animal production.
Within the archaeal community, no significant correlations
were found between Methanobrevibacter and CH4 emissions. This
was unexpected since Methanobrevibacter has previously been
associated with higher CH4 emission in cattle [48]. The lack of
association found here was likely due to differences in abso-
lute versus relative Methanobrevibacter abundance, since it was
previously determined that the absolute abundance of particular
methanogens determines CH4 production in the rumen [8,51]. In
addition, there may have been species level differences within the
Methanobrevibacter that contributed to methane emissions [7]. The
findings also revealed that archaea in the genus Methanosphaera
had negative correlations with CH4 emissions, which corroborated
the findings of Kittelmann et al. [19]. Comparisons of the pathways
for CH4 production by Methanobrevibacter and Methanosphaera
have indicated that Methanobrevibacter produces one mole of CH4
for each mole of CO2 [14], while Methanosphaera requires four
moles of methanol to produce three moles of CH4 [12]. This shows
that, for any given carbon input, Methanobrevibacter produces more
CH4 than Methanosphaera. Thus, the findings from the current
study that Methanosphaera negatively correlated with CH4 may
reflect its lower CH4 production per unit of carbon, relative to
498 C.S. Cunha et al. / Systematic and Applied Microbiology 40 (2017) 492–499
Methanobrevibacter. An alternative explanation may be the avail-
ability of methylotrophic substrates (e.g. methanol), which would
strongly influence CH4 production by both Methanobrevibacter and
Methanosphaera.
In addition, archaea from the genus Methanobrevibacter were
found to be negatively correlated with dDMI and dOMI, indicat-
ing this genus’s association with lower production efficiency. This
was confirmed by the higher relative abundance of this genus in
high forage fed animals, such as DC cows. In contrast, archaea in
the genus Methanosphaera showed a moderate positive correlation
with dDMI and dOMI. Previous work found no association between
production traits and specific ruminal archaea [7]; thus, the corre-
lations identifying specific genera of archaea in this study provided
new information in this research area.
As part of CH4 production, archaea are responsible for removing
H2 produced during the degradation of feed by other microorgan-
isms. This also facilitates the reduction of available CO2 [43]. This
is of particular importance, as H2 accumulation is harmful to the
correct functioning of the rumen, since it inhibits upstream fermen-
tation and reduces overall production efficiency [28]. As a result,
depleting archaea from the rumen without an associated strategy
for reducing H2 production is not a desirable CH4 mitigation option
[28,49]. Given that the bacterial and fungal communities provide
the substrates for CH4 production by ruminal archaea, understand-
ing the interplay between all three ruminal kingdoms is necessary
for gaining insights into overall CH4 production.
The analysis of the ruminal bacterial microbiota indicated
that bacteria in the families Christensenellaceae, Mogibacteriaceae
and S24-7, as well as in the genera Butyrivibrio, Schwartzia, and
Treponema, were negatively correlated with CH4 emissions and
positively correlated with dDMI and dOMI. These attributes are
highly desirable, as they demonstrate the feasibility of high milk
production while lowering CH4 emissions. Therefore, it is specu-
lated that bacteria within these families and genera are likely to
be able to increase DMI because their members include known
degraders of fibrous feed. Given that fiber is primarily responsi-
ble for ruminal distention and filling, and that these attributes are
related to the satiety response in cows [45], one hypothesis is that
increasing the rate of fiber degradation could lead to more DM and
OM intake, which in turn could stimulate the flow of digesta and
the stimulus to eat.
In addition, genera such as Butyrivibrio are known for produc-
ing butyrate [32], and members of this genus can direct ruminal
fermentation towards more butyrate and less acetate, which leads
to lower levels of H2 and CH4 . Indeed, it was observed that among
cows that diverged the most in relation to CH4 emission (HP versus
DC), butyrate concentrations were greater for HP cows (Table 2),
which also emitted less CH4 than DC cows. Butyrate production
in the rumen consumes hydrogen, so these bacteria may serve
as competitors for hydrogen with methanogenic archaea. This is
in contrast to previous reports where low methane emitters had
lower butyrate proportions than high methane emitters on the
same diet [7]. However, this difference is unsurprising when con-
sidering that dietary differences existed in this study and the lowest
emitting group (LC, 175 g d−1 ) was much lower than that reported
by Danielsson et al. [7] (301 g d−1 ).
Bacterial taxa were also found that negatively correlated with
CH4 emission and had no significant associations to production
traits. These included the bacterial OTUs within the Coriobacteri-
aceae, RFP12, and Clostridium. Previous work confirms the findings
of the current study that Clostridium was negatively correlated with
CH4 emissions [47], although its actual role in CH4 is likely to be
variable given that this genus encompasses a wide variety of species
with numerous metabolic abilities. However, the other associated
taxa have been less well studied and their role in methane pro-
duction in the rumen is unknown. While these bacteria may not
increase production, their association with decreased CH4 emis-
sions is desirable, and future work pinpointing their role in CH4
emissions should be considered. Interestingly, none of the bacterial
taxa previously found to correlate with lower methane emissions,
namely those in the family Succinivibrionaceae or the genus Pre-
votella [7], were found to be associated with lower methane and
increased or unchanged production in this study. This could be due
to numerous factors, including diet and the impact of a tropical
management system on these animals.
Finally, the study also documented the ruminal anaerobic fungi
and their relationship to both production and CH4 emissions. Given
their role in degrading fibrous feed consumed by the host, fungi may
indirectly affect CH4 emissions by modulating fermentation pat-
terns. In line with other studies [23,25], the majority of the rumen
liquid-associated fungi were identified as unclassified (73.19%).
Due to their primary function, anaerobic fungi are predominantly
found attached to the particles of the solid phase of ruminal content
[1] and the findings of the present study may be due to the sam-
pling approach used, which only considered the liquid portion of
the rumen. This may explain why no significant associations were
found between the anaerobic fungal community and CH4 emis-
sions. Due to the non-significant correlations with fungi found in
the study, it is suggested that future work on the ruminal anaerobic
fungi should include those populations found in the solid portion
of the rumen in order to understand their role better in mediating
CH4 emissions.
Conclusions
By considering the ruminal bacteria, fungi, and archaea within
the context of CH4 emissions and animal production, the findings
identified a number of bacterial families and genera associated
with decreased CH4 emissions without reducing animal produc-
tion. The study specifically mirrored tropical commercial dairy
farms in an effort to understand more fully how such manage-
ment systems, which work to maximize production, impact CH4
emissions. Based on the results, it is suggested that attempts to
increase the abundances of microbial groups, such as Christensenel-
laceae, Mogibacteriaceae, S24-7, Butyrivibrio, Shwartzia, Treponema
and Methanosphaera, may result in a decrease in CH4 emissions
and an improvement in animal production traits, while decreasing
taxa affiliated to Coriobacteriaceae, RFP12 and Clostridium is likely
to reduce CH4 emissions without any impact on animal production.
Future work considering these taxa should be undertaken to assess
their impact more fully on increasing production, such as meat and
milk, while limiting CH4 emissions.
Acknowledgements
This work was supported by the Brazilian governmental
agencies CNPq, CAPES, FAPEMIG, INCT-CA, Embrapa, and PECUS-
RumenGases. The work was also supported, in part, by United
States Department of Agriculture, National Institute of Food and
Agriculture grant #WIS01729 to G.S. We thank all members of the
Veloso, Marcondes, Mantovani and Suen laboratories for their sup-
port and helpful discussions. We would also like to acknowledge
the CAPES Doctorate Sandwich Program that facilitated the collab-
oration enabling this study.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.syapm.2017.07.
008.
 C.S. Cunha et al. / Systematic and Applied Microbiology 40 (2017) 492–499
References
[1] Akin, D.E., Borneman, W.S. (1990) Role of rumen fungi in fiber degradation. J.
Dairy. Sci. 73, 3023–3032.
[2] Association of Official Analytical Chemists (AOAC), 1990 Official methods of
analyses of the AOAC, Fifteenth edition, AOAC International, Washington, pp. ,
1–1230.
[3] (2006) Nutrição de Ruminantes. In: Berchiellli, T.T., Pires, A.V., Oliveira, S.G.
(Eds.), FUNEP, Jaboticabal, 583p.
[4] Boadi, D., Benchaar, C., Chiquette, J., Massé, D. (2004) Mitigation strategies
to reduce enteric methane emissions from dairy cows: update review. Can.
J. Anim. Sci. 84, 319–335.
[5] Clarke, K.R. (1993) Non-parametric multivariate analyses of changes in com-
munity structure. Aust. J. Ecol. 18, 117–143.
[6] Clarke, K.R., Warwick, R.M. 1994 Change in marine communities: an approach
to statistical analysis and interpretation, Natural Environment Research Coun-
cil, UK.
[7] Danielsson, R., Dicksved, J., Sun, L., Gonda, H., Müller, B., Schnürer, A., Bertils-
son, J. (2017) Methane production in dairy cows correlates with rumen
methanogenic and bacterial community structure. Front. Microbiol. 8, 226.
[8] Danielson, R., Schnürer, A., Arthurson, V., Bertilsson, J. (2012) Methanogenic
population and methane production in Swedish dairy cows fed different levels
of forage. Appl. Environ. Microbiol. 78, 6172–6179.
[9] DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E.L., Keller, K., Huber,
T., Dalevi, D., Hu, P., Andersen, G.L. (2006) Greengenes, a chimera-checked 16S
rRNA gene database and 666 workbench compatible with ARB. Appl. Environ.
Microbiol. 72, 5069–5072.
[10] Detmann, E., Paulino, M.F., Valadares Filho, S.C., Huhtanen, P. (2014) Nutritional
aspects applied to grazing cattle in the tropics: a review based on Brazilian
results. Semina: Ciências Agrárias, Londrina, v.35, n.4, Suppl., pp. 2829–2854.
[11] Detmann, E., Souza, M.A., Valadares Filho, S.C., Queiroz, A.C., Berchielli, T.T.,
Saliba, E.O.S., Cabral, L.S., Pina, D.S., Ladeira, M.M., Azevêdo, J.A.G. 2012 Méto-
dos para análise de alimentos, first ed., Suprema, Visconde do Rio Branco, MG,
Brazil.
[12] Fricke, W.F., Seedorf, H., Henne, A., Kruer, M., Liesegang, H., Hedderich, R.,
Gottschalk, G., Thauer, R.K. (2006) The genome sequence of Methanosphaera
stadtmanae reveals why this human intestinal archaeon is restricted to
methanol and H2 for methane formation and ATP synthesis. J. Bacteriol. 188,
624–658.
[13] Henderson, G., Cox, F., Ganesh, S., Jonker, A., Young, W., Global Rumen Census
Collaborators, Hansen, P.H. (2015) Rumen microbial community composition
varies with diet and host, but a core microbiome is found across a wide geo-
graphical range. Sci. Rep. 5, 14567.
[14] Hook, S.E., Wright, A.-D.G., McBride, B.W. (2010) Methanogens: methane pro-
ducers of the rumen and mitigation strategies. Archaea, 945785.
[15] Intergovernmental Panel on Climate Change. Climatechange 2014 Synthesis
report 2014.
[16] Johnson, K.A., Huyler, M., Westberg, H.H., Lamb, B., Zimmerman, P. (1994)
Measurement of methane emissions from ruminant livestock using a sulphur
hexafluoride tracer technique. Environ. Sci. Technol. 28, 59–362.
[17] Johnson, K.A., Johnson, D.E. (1995) Methane emissions from cattle. J. Anim. Sci.
73, 2483–2492.
[18] Johnson, K.A., Westberg, H.H., Michal, J.J., Cossalman, M.W. (2007) The SF6
tracer technique: methane measurement from ruminants. In: Makkar, H.S.,
Vercoe, P. (Eds.), Measuring methane production from ruminants, Springer,
Netherlands, pp. 33–67.
[19] Kittelmann, S., Pinares-Patiño, C.S., Seedorf, H., Kirk, M.R., Ganesh, S., McEwan,
J.C., Janssen, P. (2014) Two different bacterial community types are linked with
the low-methane emission trait in sheep. PLoS One 9, e103171.
[20] Kittelmann, S., Seedorf, H., Walters, W.A., Clemente, J.C., Knight, R., Gordon, J.I.,
Janssen, P. (2013) Simultaneous amplicon sequencing to explore co-occurrence
patterns of bacterial, archaeal and eukaryotic microorganisms in rumen micro-
bial communities. PLoS One 8, e47879.
[21] Kõljalg, U., Nilsson, R.H., Abarenkov, K., Tedersoo, L., Taylor, A.F.S., Bahram,
M., Bates, S.T., Bruns, T.D., Bengtsson-Palme, J., Callaghan, T.M., Douglas, B.,
Drenkhan, T., Eberhardt, U., Dueñas, M., Grebenc, T., Griffith, G.W., Hartmann,
M., Kirk, P.M., Kohout, P., Larsson, E., Lindahl, B.D., Lücking, R., Martín, M.P.,
Matheny, P.B., Nguyen, N.H., Niskanen, T., Oja, J., Peay, K.G., Peintner, U., Peter-
son, M., Põldmaa, K., Saag, L., Saar, I., Schüßler, A., Scott, J.A., Senés, C., Smith,
M.E., Suija, A., Taylor, D.L., Telleria, M.T., Weiss, M., Larsson, K.-H. (2013)
Towards a unified paradigm for sequence-based identification of fungi. Mol.
Ecol. 22, 5271–5277.
[22] Kozich, J.J., Westcott, S.L., Baxter, N.T., Highlander, S.K., Schloss, P.D. (2013)
Development of a dual-index sequencing strategy and curation pipeline for
analyzing amplicon sequence data on the MiSeq Illumina sequencing platform.
Appl. Environ. Microbiol. 79, 5112–5120.
[23] Kumar, S., Indugu, N., Vecchiarelli, B., Pitta, D.W. (2015) Associative patterns
among anaerobic fungi, methanogenic archaea, and bacterial communities in
response to changes in diet and age in the rumen of dairy cows. Front. Microbiol.
6, 781.
[24] Lana, R.P. 2005 Nutrição e Alimentação animal: mitos e realidades, UFV, Viçosa,
pp. 344.
[25] Liggenstoffer, A.S., Youssef, N.H., Couger, M., Elshahed, M.S. (2010) Phylo-
genetic diversity and community structure of anaerobic gut fungi (phylum
499
Neocallimastigomycota) in ruminant and non-ruminant herbivores. ISME J. 4,
1225–1235.
[26] Lodge-Ivey, S.L., Browne-Silva, J., Horvath, M.B. (2009) Technical note: Bacterial
diversity and fermentation end products in rumen fluid samples collected via
oral lavage or rumen cannula. J. Anim. Sci. 87, 2333–2337.
[27] Mantel, N. (1967) The detection of disease clustering and a generalized regres-
sion approach. Cancer Res. 27, 209–220.
[28] Martin, C., Morgavi, D.P., Doreau, M. (2010) Methane mitigation in ruminants:
from microbe to the farm scale. Animal 4, 351–365.
[29] McCaughey, W.P., Wittenberg, K., Corrigan, D. (1999) Impact of pasture type on
methane production by lactating beef cows. Can. J. Anim. Sci. 79, 221–226.
[30] Mendiburu, F., (2016) Agricolae: Statistical Procedures for Agricul-
tural Research. R package version 1. 2–4. http://CRAN.R-project.org/
package=agricolae.
[31] Menezes, A.B., Lewis, E., O’Donovan, M., O’Neill, B.F., Clipson, N., Doyle, E.M.
(2011) Microbiome analysis of dairy cows fed pasture or total mixed ration
diets. FEMS Microbiol. Ecol. 78, 256–265.
[32] Mohammed, R., Brink, G.E., Stevenson, D.M., Neumann, A.P., Beauchemin, K.A.,
Suen, G., Weimer, P.J. (2014) Bacterial communities in the rumen of Holstein
heifers differ when fed orchardgrass as pasture vs hay. Front. Microbiol. 5, 689.
[33] Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B.,
Simpson, G.L., Solymos, P., Stevens, M.H.H., Wagner, H. 2015 Vegan: community
ecology package http://vegan.r-forge.r-project.org/.
[34] Oss, D.B., Marcondes, M.I., Machado, F.S., Pereira, L.G.R., Tomich, T.R., Ribeiro,
G.O., Jr, Chizzotti, M.L., Ferreira, A.L., Campos, M.M., Maurício, R.M., Chaves, A.V.,
McAllister, T.A. (2016) An evaluation of the face mask system based on short-
term measurements compared with the sulfur hexafluoride (SF6 ) tracer, and
respiration chamber techniques for measuring CH4 emissions. Anim. Feed Sci.
Technol. 216, 49–57.
[35] Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., Peplies, J., Glockner,
F.O. (2007) SILVA: a comprehensive online resource for quality checked and
660 aligned ribosomal RNA sequence data compatible with ARB. Nucl. Acids
Res. 35, 7188–7196.
[36] R Core Team, 2015 R: a language and environment for statistical computing, R
Foundation for Statistical Computing, Vienna, Austria.
[37] Roehe, R., Dewhurst, R.J., Duthie, C.A., Rooke, J.A., McKain, N., Ross, D.W., Hylop,
J.J., Waterhouse, A., Freeman, T.C., Watson, M., Wallace, R.J. (2016) Bovine host
genetic variation influences rumen microbial methane production with best
selection criterion for low methane emitting and efficiently feed converting
hosts based on metagenomic gene abundance. PLoS Genet. 12, e1005846.
[38] Ross, E.M., Moate, P.J., Marett, L., Cocks, B.G., Hayes, B.J. (2013) Investigating
the effect of two methane-mitigating diets on the rumen microbiome using
massively parallel sequencing. J. Dairy Sci. 96, 6030–6046.
[39] Schloss, P.D., Handelsman, J. (2005) Introducing DOTUR, a computer program
for defining operational taxonomic units and estimating species richness. Appl.
Environ. Microbiol. 71, 1501–1506.
[40] Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister,
E.B., Lesniewski, R.A., Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W.,
Stres, B., Thallinger, G.G., Van Horn, D.J., Weber, C.F. (2009) Introducing
Mothur: open-source, platform-independent, community supported software
for describing and comparing microbial communities. Appl. Environ. Microbiol.
75, 7537–7541.
[41] Siegfried, B.R., Ruckemann, H., Stumpf, G. 1984 Method for the determination
of organic acids in silage by high performance liquid chromatography, Vol. 37,
Landwirtsch. Forsch, p. 298.
[42] Stevenson, D.A., Weimer, P.J. (2007) Dominance of Prevotella and low abun-
dance of classical ruminal bacterial species in the bovine rumen revealed by
relative quantification real-time PCR. Appl. Microbiol. Biotechnol. 75, 165–174.
[43] Stewart, C.S., Flynt, H.J., Bryant, M.P. (1997) The rumen bacteria. In: Hobson,
P.N., Stewart, C.S. (Eds.), The rumen microbial ecosystem, Blackie Academic
and Professional, New York, NY, pp. 10–72.
[44] St-Pierre, N.R., Thraen, C.S. (1999) Animal grouping strategies, sources of vari-
ation, and economic factors affecting nutrient balance on dairy farms. J. Anim.
Sci. 77, 72–83.
[45] Van Soest, P.J. 1982 Gastrointestinal fermentations. In: Nutritional ecology of
the ruminant. Ruminant metabolism, nutritional strategies, the cellulolytic fer-
mentation and the chemistry of forages and plant fibers, Cornell University
Press, Ithaca, NY, pp. , 152–177.
[46] Vlaming, J.B., Lopez-Villalobos, N., Brookes, I.M., Hoskin, S.O., Clark, H. (2008)
Within- and between-animal variance in methane emissions in non-lactating
dairy cows. Aust. J. Exp. Agric. 48, 124–127.
[47] Wallace, R.J., Rooke, J.A., Duthie, C., Hyslop, J.J., Ross, D.W., McKain, N., Souza,
S.M., Snelling, T.J., Waterhouse, A., Roehe, R.1 (2014) Archaeal abundance in
post-mortem ruminal digesta may help predict methane emissions from beef
cattle. Sci. Rep. 4, 5892.
[48] Wallace, R.J., Rooke, J.A., McKain, N., Duthie, C.A., Hyslop, J.J., Ross, D.W., Water-
house, A., Watson, M., Roehe, R. (2015) The rumen microbial metagenome
associated with high methane production in cattle. BMC Genomic 16, 839.
[49] Wedlock, D.N., Janssen, P.H., Leahy, S., Shu, D., Buddle, B.M. (2013) Progress in
the development of vaccines against rumen methanogens. Animal 7, 244–252.
[50] Wei, T. (2012) Package ‘Corrplot’.
[51] Zhou, M., Chung, Y.-H., Beauchemin, K.A., Holtshausen, L., Oba, M., McAllister,
T.A., Guan, L.L. (2011) Relationship between rumen methanogens and methane
production in dairy cows fed diets supplemented with a feed enzyme additive.
J. Appl. Microbiol. 111, 1148–1158.