bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022.

The copyright holder for this preprint

(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 Omicron BA.1 and BA.2 are antigenically distinct SARS-CoV-2 variants

2 Authors: Anna Z. Mykytyn1, Melanie Rissmann1*, Adinda Kok1*, Miruna E. Rosu1*, Debby
3 Schipper1, Tim I. Breugem1, Petra B. van den Doel1, Felicity Chandler1, Theo Bestebroer1,
4 Maurice de Wit2, Martin E. van Royen3, Richard Molenkamp1, Bas B. Oude Munnink1, Rory D.
5 de Vries1, Corine GeurtsvanKessel1, Derek J. Smith4, Marion P. G. Koopmans1, Barry Rockx1,
6 Mart M. Lamers1, Ron Fouchier1, Bart L. Haagmans1**
1
7 Viroscience Department, Erasmus Medical Center, Rotterdam, The Netherlands.
2
8 Department of Neurology, Erasmus University Medical Center, Rotterdam, The Netherlands
3
9 Department of Pathology, Erasmus University Medical Center, Rotterdam, The Netherlands.
4
10 Center for Pathogen Evolution, Department of Zoology, University of Cambridge, Cambridge,
11 CB2 3EJ, UK

12

13 *These authors contributed equally to this work.

14 **Correspondence: b.haagmans@erasmusmc.nl

15

16 Abstract: The emergence and rapid spread of SARS-CoV-2 variants may impact vaccine
17 efficacy significantly1. The Omicron variant termed BA.2, which differs genetically substantially
18 from BA.1, is currently replacing BA.1 in several countries, but its antigenic characteristics have
19 not yet been assessed2,3. Here, we used antigenic cartography to quantify and visualize antigenic
20 differences between SARS-CoV-2 variants using hamster sera obtained after primary infection.
21 Whereas early variants are antigenically similar, clustering relatively close to each other in
22 antigenic space, Omicron BA.1 and BA.2 have evolved as two distinct antigenic outliers. Our
23 data show that BA.1 and BA.2 both escape (vaccine-induced) antibody responses as a result of
24 different antigenic characteristics. Close monitoring of the antigenic changes of SARS-CoV-2
25 using antigenic cartography can be helpful in the selection of future vaccine strains.

26

27
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

28 Since its emergence in Wuhan, China, in 2019, SARS Coronavirus 2 (SARS-CoV-2) has caused
29 over 300 million cases and 5.5 million deaths4. The initial virus that spread globally was
30 characterized by a D614G change in the spike (S) protein5.Approximately one year into the
31 pandemic other variants with 6-12 mutations in the S protein started to become dominant in
32 various countries6. These variants included the Alpha variant in the United Kingdom, the Beta
33 variant in South Africa, and the Gamma variant in Brazil, of which Alpha became the dominant
34 variant globally by early 2021. In the summer of 2021, the Delta variant emerged first in India,
35 and replaced Alpha globally within several months 7-10. Emerging variants are termed Variants of
36 Concern (VOC) by the World Health Organization (WHO) if they are associated with a major
37 change in epidemiology and/or clinical presentation, increased virulence, increased
38 transmissibility, and/or decreased effectiveness of public health and social measures or available
39 diagnostics, vaccines or therapeutics 11. In addition, the WHO has designated other variants as
40 Variants of Interest (VOI), which possess mutations predicted or known to affect antibody
41 escape, virulence, or transmission. At the end of 2021 the VOC Omicron (sublineage BA.1)
42 emerged in Botswana and South Africa, carrying 30 mutations in S, raising concerns for
43 extensive immune evasion1. Whereas several earlier VOCs and VOIs exhibit some levels of
44 antibody escape (e.g. Beta, Gamma, Delta, and Mu), they were still neutralized well by
45 convalescent and post-vaccination sera12-16. In contrast, Omicron BA.1 almost completely
46 escapes neutralizing antibodies, leading to low levels of remaining protective antibodies in most
47 previously infected or vaccinated individuals, and a high frequency of breakthrough infections.
48 This antibody escape at least partly explains why this variant has become the dominant variant
49 globally over the span of a few weeks17-20. Fortunately, BA.1 appears to be less virulent
50 compared with earlier variants21,22. A second Omicron variant (sublineage BA.2), emerged in
51 South Africa around the same time as BA.1 and differs from BA.1 by 40 mutations (including 18
52 substitutions and 5 indels in S)2. Whereas BA.2 initially was only sporadically detected, it is
53 currently replacing BA.1 in several countries3. It is unclear why BA.2 is spreading so fast, but it
54 may be inherently more transmissible or considerably antigenically distinct from BA.1 and
55 earlier variants.

56 As SARS-CoV-2 continues to evolve and escape neutralizing antibody responses, it is becoming

57 increasingly important to understand the antigenic relationships among variants and the
58 substitutions that underlie antigenic changes. Antigenic cartography is a tool to quantitatively
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

59 analyze antigenic drift and visualize the emergence of new antigenic clusters, which is why it is
60 used to biannually inform influenza virus vaccine strain selection23,24. Here, we investigated the
61 neutralizing activity of human post-vaccination sera to both Omicron variants and applied
62 antigenic cartography to 11 SARS-CoV-2 variants and hamster sera elicited against 8 SARS-
63 CoV-2 variants by primary infection.

64

65 Neutralizing activity of human sera against Omicron BA.1 and BA.2

66 Multiple studies have shown that Omicron BA.1 efficiently evades antibody responses post-
67 infection and post-vaccination12-16. However, few studies have analyzed antibody responses to
68 Omicron BA.2 . Therefore, we investigated to what extent human post-BNT162b2 vaccination
69 sera neutralized Omicron BA.2. After a single BNT162b2 vaccination, on average low
70 neutralizing titers were detected against all variants with a 13 and 8-fold drop in neutralizing
71 titers against Omicron BA.1 and BA.2, respectively, compared with 614G. In line with previous
72 findings, Omicron BA.1 was 61-fold less efficiently neutralized after two BNT162b2
73 vaccinations18-20 (Fig. 1a-b). In comparison, Omicron BA.2 was neutralized somewhat more
74 efficiently with a 13-fold drop in neutralizing titers compared with 614G. A third vaccination
75 with BNT162b2 reduced the fold change to 614G to 11 and 7-fold for BA.1 and BA.2,
76 respectively. Titers against all variants increased with each dose (Fig. 1a-c). Combined, these
77 data show that Omicron BA.1 and BA.2 escape antibody responses, and suggest that the height
78 and breadth of the antibody response against SARS-CoV-2 can be increased by repeated
79 exposure to the original antigen. The differential effect of booster vaccination on BA.1 and BA.2
80 suggested that both variants are antigenically distinct, warranting further analysis. Sera from
81 primary infections with SARS-CoV-2 variants can be used to assess the antigenicity of different
82 variants, however human sera from primary variant infections are increasingly difficult to obtain.
83 We did attempt to obtain human sera post-primary Omicron infection and included 12
84 individuals who had not been vaccinated, of which 4 reported a previous SARS-CoV-2 infection.
85 6 had high neutralizing titers against 614G and Delta, suggesting that an additional 2 individuals
86 had been infected with another variant prior to their BA.1 infection (Extended Data Fig. 1a). The
87 remaining 6 sera had low neutralizing activity against all variants, including Omicron (geometric
88 mean titer of 14, 32, 63 and 41 against 614D, Delta, Omicron BA.1 and BA.2 respectively)
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

89 (Extended Data Fig. 1b). As primary antisera against future variants will become even more
90 difficult to obtain, we determined the antigenicity of SARS-CoV-2 variants using the hamster
91 model. Hamsters were used as they are highly susceptible to SARS-CoV-2 and are therefore
92 ideal for controlled infections and obtaining well defined sera25-27.

93

94 Antigenic cartography of SARS-CoV-2

95 To investigate the antigenic relationships between BA.1, BA.2 and other SARS-CoV-2 variants,
96 we used antigenic cartography28. We used the Syrian golden hamster model to generate antisera
97 by inoculating hamsters with SARS-CoV-2 variants (614G, Alpha, Beta, Gamma, Zeta, Delta,
98 Mu and Omicron (lineage BA.1)) (Fig. 2a, Extended Data Fig. 2). Virus stocks and the original
99 respiratory specimens were sequenced to confirm the absence of culture-acquired mutations that
100 may affect antigenic properties. In addition, at 7 days post-infection (dpi) swabs were collected
101 and sequenced to confirm that the virus did not change during the experiment. Apart from Delta
102 and Omicron BA.1-infected animals, whose swabs yielded too little virus to sequence, no
103 mutations in Spike were detected in swab sequences. Plaque reduction neutralization titers
104 resulting in a 50% reduction in infected cells (PRNT50) obtained with both authentic SARS-
105 CoV-2 and pseudotyped viruses were used to generate antigenic maps (Fig. 2b). All animals
106 were successfully infected, as shown by high viral RNA titers at 1 dpi in nasal washes and high
107 homologous antibody titers at 26 dpi (Fig. 2c-d).

108 Pseudovirus neutralization assays are a safe and widely used tool to assess antibody
109 neutralization. We performed initial neutralization experiments on VeroE6 cells, which are the
110 most commonly used cell line for neutralization assays. We used the spike variants 614D, 614G,
111 Alpha, Beta, Delta, Kappa and Omicron BA.1. All homologous sera neutralized to high titers
112 (Extended Data Fig. 3a-h). The lowest cross-neutralizing titers were obtained against Omicron
113 for all sera with 9 to 43-fold reduction compared to homologous neutralization. Sera from
114 Omicron BA.1-infected hamsters poorly neutralized all other variants (2 to 81 fold reduction
115 compared to homologous neutralization). These data show that Omicron BA.1 induces different
116 antibody responses compared with all other variants. Next, we performed pseudovirus
117 neutralization assays on the human airway Calu-3 cell line. In Calu-3 cells SARS-CoV-2 enters
118 using the serine protease-mediated entry pathway that is also used in primary cells, whereas in
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

119 VeroE6 cells SARS-CoV-2 enters using the cathepsin-mediated endocytic entry pathway29. In
120 addition, the variability in infectivity between variants is lower for Calu-3 cells compared with
121 VeroE6, suggesting that Calu-3 cells may allow for more equal comparisons30. Neutralization
122 titers on Calu-3 cells were similar to VeroE6 and correlated well (Extended Data Fig. 3i-p,
123 Extended Data Fig. 4a-g).

124 Next, we constructed antigenic maps from the neutralization data using a multidimensional
125 scaling (MDS) algorithm described previously28 (Fig. 3a-b). These were verified to ensure fitted
126 distances correlated well with actual neutralizing titers and to confirm that the data was well
127 represented in two dimensions (Extended Data Fig. 5a-d). We found that the map constructed by
128 Calu-3 cells was very similar to the VeroE6 map, since the same antigens plotted within one 2-
129 fold dilution from each other in the two maps (Extended Data Fig. 6). Therefore, the choice of
130 the cell line for the neutralization assay did not affect the topology of the map substantially. In
131 order to assess whether the map generated with pseudovirus neutralization data would accurately
132 represent authentic SARS-CoV-2, we generated a map with the same antisera as in Fig 3a-b and
133 614G, Alpha, Beta, Delta and Omicron viruses on Calu-3 cells. (Fig. 3c). We confirmed that
134 there was a good correlation between the raw neutralizing titers of the 5 variants on Calu-3 and
135 VeroE6 cells (Extended Data Fig. 7-8). The map generated with authentic SARS-CoV-2 closely
136 resembled the maps generated with pseudovirus, the positions of the antigens differed no more
137 than one 2-fold dilution between maps (Fig. 3d). Antigenically, we found that all variants aside
138 from Omicron BA.1 grouped closely together. Omicron BA.1 formed a distinct antigenic outlier
139 in the map, 10 to 38-fold dilutions away from the nearest virus (Fig. 3a-c).
140 Next, we extended the authentic virus dataset to contain a larger set of variants: 614G, Alpha,
141 Beta, Gamma, Zeta, Delta, Delta AY.4.2, Lambda, Mu, Omicron BA.1 and Omicron BA.2.
142 As for pseudotyped virus, homologous sera neutralized to a high titer across all variants (Fig. 4a-
143 h, homologous titers per panel are a re-display from Fig. 1d). Similar to pseudovirus data, we
144 observed a reduction in neutralization titers of Omicron BA.1 sera against all other variants (2.4
145 to 9 fold compared to homologous) and poor neutralization of Omicron BA.1 by all non-
146 homologous sera (8 to 112 fold reduction). In addition, Omicron BA.2 was also poorly
147 neutralized by all sera (7 to 114 fold reduction), including Omicron BA.1 (8 fold reduction).
148 Although Omicron BA.1 and Omicron BA.2 possess many overlapping mutations in S, the
149 differences between the variants were sufficient to prevent efficient cross-neutralization. In
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

150 agreement, our study shows that antibodies elicited against the original SARS-CoV-2 cluster do
151 not neutralize Omicron BA.1 well, and vice-versa.

152 The generated antigenic map based on the extended neutralization dataset was verified by
153 ensuring fitted distances correlated well with actual neutralizing titers, that the data was well
154 represented in two dimensions and plotted the uncertainty of each antigen and antisera (Extended
155 Data Fig. 9). The antigenic map shows that, similarly to the maps in Fig. 3, all variants except
156 Omicron BA.1 and BA.2 group together in one antigenic cluster (Fig. 4i). In line with the limited
157 cross-neutralization of Omicron BA.2 with BA.1 sera, both variants were positioned distantly
158 from each other in the map, with BA.2 somewhat closer to the main cluster than BA.1, indicating
159 that Omicron BA.1 induced qualitatively different antibody responses and BA.1 and BA.2 are
160 antigenically distinct SARS-CoV-2 variants. The antisera corresponding to each virus grouped in
161 the same region of the map, indicating efficient homologous neutralization. As expected, based
162 on Fig. 3, 614G and Alpha are in the center of the cluster. Within this cluster, viruses grouped
163 together based on specific substitutions. Viruses containing E484K (Beta, Gamma, Zeta and Mu)
164 grouped to the top-right of the ancestral 614G virus, whereas viruses containing substitutions
165 L452R/Q grouped to the left of 614G. The Beta and Gamma variants, which in addition to
166 E484K both contain K417N/T also cluster together. The same clustering based on E484K and
167 L452R/Q was observed in the pseudovirus maps. A recently study by Wilks and colleagues used
168 human sera to generate an antigenic map31. Both the maps from our study and the map from
169 Wilks and colleagues show the same clustering of variants containing mutations at position 484,
170 417 and 452, indicating that hamsters and humans generate similar antibody responses.
171 However, differences were observed as well. In the map by Wilks and colleagues, there was
172 approximately a ~2 fold larger distance between 614G and variants Beta, Gamma and Mu,
173 compared to our map containing all authentic viruses. This may be caused by specific antigenic
174 relationships between the large set of variants, as including only 5 variants increased the distance
175 from 614G to Beta by ~2 fold (Fig. 3c). In addition, these differences may be caused by the
176 lower titers observed in naturally-infected humans compared with experimentally-infected
177 hamsters. Low titers against immunizing viruses may drop off more for viruses with only a few
178 immune-evasive substitutions (e.g. Beta, Gamma, and Delta), leading to relatively large
179 antigenic distances. On the other hand, low titers may underestimate the antigenic distance for
180 highly evasive viruses due to reaching the assay’s lower limit of detection. In agreement, the
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

181 distance of the main cluster to Omicron BA.1 was larger in our map compared with the map by
182 Wilks and colleagues. Nevertheless, these data indicate that human and hamster serum responses
183 to SARS-CoV-2 infection are similar as they lead to topologically similar maps. However, the
184 specific set of viruses and sera used to construct an antigenic map may influence antigenic
185 distances, e.g. due to omission of sera against Omicron. As human sera post-primary infection
186 are increasingly difficult to obtain, our data suggest that antigenic cartography using hamster sera
187 is a useful surrogate to assess antigenic relationships between SARS-CoV-2 variants.

188 The emergence and rapid spread of the heavily mutated Omicron BA.1 and BA.2 variants
189 suggests that population immunity is exerting strong selective pressure on SARS-CoV-2,
190 favoring the emergence of new antigenic variants. As the number of SARS-CoV-2 variants
191 increases it will become increasingly important to visualize and understand the antigenic
192 relationships between variants. Here, we used antigenic cartography to quantify and visualize
193 SARS-CoV-2 antigenic evolution and demonstrate that Omicron BA.1 and BA.2 have evolved as
194 two antigenically distinct variants, separate from an ancestral cluster with all earlier SARS-CoV-
195 2 variants.

196 The evolutionary history of SARS-CoV-2 in humans is relatively short compared with viruses
197 that have circulated in humans for decades, such as influenza viruses32. Before the emergence of
198 Omicron, most SARS-CoV-2 variants contained only few substitutions in the S protein and were
199 still recognized by convalescent and post-vaccination sera. These variants all grouped into the
200 same antigenic cluster in our antigenic maps, but within that cluster a grouping could be
201 observed based on S substitutions at position 417, 484 and 452, in line with previous data on
202 human sera31. The Omicron variants, however, positioned as distinct antigenic variants with
203 limited cross-neutralization to each other and the original cluster. These data were corroborated
204 by human post-vaccination antibody responses (Fig. 1). The ameliorating effect of the booster
205 vaccine on Omicron BA.2 responses suggests that population immunity against BA.2 may be
206 sufficient to prevent flooding of healthcare systems and high levels of mortality, seen before for
207 previous variants33-35. In addition, the widespread circulation of BA.1 may lead to the broadening
208 of the antibody response in previously infected or vaccinated individuals and subsequent
209 dampening of the intensity of spread. However, in regions with low access to vaccines, a wave of
210 primary infections with BA1 would potentially lead to low level cross-protection and continued
211 opportunity for further widespread circulation of BA2 or other variants.
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

212 The antigenic cartography of SARS-CoV-2 visualizes clearly how BA.1 and BA.2 can both
213 escape antibody responses without being antigenically similar. The emergence of both Omicron
214 variants indicates that population immunity is selecting for SARS-CoV-2 variants that efficiently
215 escape from neutralizing antibody responses, leading to the first signs of antigenic drift. SARS-
216 CoV-2 will eventually reach endemicity and likely cause annual or biannual infection waves as
217 seen for influenza and seasonal coronaviruses. Our study lays the foundation for the continuous
218 monitoring of the antigenic evolution of SARS-CoV-2, which may inform the selection of
219 vaccine strains in the future.
220

221 References
222 1 in CDC COVID-19 Science Briefs (2020).
223 2 Gallaher, W. R. Omicron is a Multiply Recombinant Set of Variants That Have Evolved Over Many
224 Months, < https://virological.org/t/omicron-is-a-multiply-recombinant-set-of-variants-that-have-
225 evolved-over-many-months/775> (2022).
226 3 e.V., F. v. G. GISAID - hCov19 Variants, <https://www.gisaid.org/hcov19-variants/> (2022).
227 4 WHO Coronavirus (COVID-19) Dashboard, <https://covid19.who.int/> (2022).
228 5 Korber, B. et al. Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity of
229 the COVID-19 Virus. Cell 182, 812-827 e819, doi:S0092-8674(20)30820-5 [pii]
230 10.1016/j.cell.2020.06.043 (2020).
231 6 Plante, J. A. et al. The variant gambit: COVID-19's next move. Cell Host Microbe 29, 508-515,
232 doi:10.1016/j.chom.2021.02.020 (2021).
233 7 Campbell, F. et al. Increased transmissibility and global spread of SARS-CoV-2 variants of concern as at
234 June 2021. Euro Surveill 26, doi:10.2807/1560-7917.ES.2021.26.24.2100509 (2021).
235 8 Frampton, D. et al. Genomic characteristics and clinical effect of the emergent SARS-CoV-2 B.1.1.7
236 lineage in London, UK: a whole-genome sequencing and hospital-based cohort study. Lancet Infect Dis,
237 doi:10.1016/S1473-3099(21)00170-5 (2021).
238 9 Mlcochova, P. et al. SARS-CoV-2 B.1.617.2 Delta variant replication and immune evasion. Nature 599,
239 114-119, doi:10.1038/s41586-021-03944-y
240 10.1038/s41586-021-03944-y [pii] (2021).
241 10 Tegally, H. et al. Detection of a SARS-CoV-2 variant of concern in South Africa. Nature 592, 438-443,
242 doi:10.1038/s41586-021-03402-9
243 10.1038/s41586-021-03402-9 [pii] (2021).
244 11 Tracking SARS-CoV-2 variants, <https://www.who.int/en/activities/tracking-SARS-CoV-2-variants/>
245 (2022).
246 12 Planas, D. et al. Sensitivity of infectious SARS-CoV-2 B.1.1.7 and B.1.351 variants to neutralizing
247 antibodies. Nat Med 27, 917-924, doi:10.1038/s41591-021-01318-5 (2021).
248 13 Hoffmann, M. et al. SARS-CoV-2 variants B.1.351 and P.1 escape from neutralizing antibodies. Cell 184,
249 2384-2393 e2312, doi:S0092-8674(21)00367-6 [pii]
250 10.1016/j.cell.2021.03.036 (2021).
251 14 Uriu, K. et al. Neutralization of the SARS-CoV-2 Mu Variant by Convalescent and Vaccine Serum. N Engl
252 J Med 385, 2397-2399, doi:10.1056/NEJMc2114706 (2021).
253 15 Planas, D. et al. Reduced sensitivity of SARS-CoV-2 variant Delta to antibody neutralization. Nature 596,
254 276-280, doi:10.1038/s41586-021-03777-9
255 10.1038/s41586-021-03777-9 [pii] (2021).
256 16 GeurtsvanKessel, C. H. et al. Divergent SARS CoV-2 Omicron-reactive T- and B cell responses in
257 COVID-19 vaccine recipients. Sci Immunol, eabo2202, doi:10.1126/sciimmunol.abo2202 (2022).
258 17 Planas, D. et al. Considerable escape of SARS-CoV-2 Omicron to antibody neutralization. Nature,
259 doi:10.1038/s41586-021-04389-z
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

260 10.1038/s41586-021-04389-z [pii] (2021).

261 18 Cele, S. et al. Omicron extensively but incompletely escapes Pfizer BNT162b2 neutralization. Nature,
262 doi:10.1038/s41586-021-04387-1
263 10.1038/s41586-021-04387-1 [pii] (2021).
264 19 Rossler, A., Riepler, L., Bante, D., von Laer, D. & Kimpel, J. SARS-CoV-2 Omicron Variant
265 Neutralization in Serum from Vaccinated and Convalescent Persons. N Engl J Med,
266 doi:10.1056/NEJMc2119236 (2022).
267 20 Dejnirattisai, W. et al. SARS-CoV-2 Omicron-B.1.1.529 leads to widespread escape from neutralizing
268 antibody responses. Cell 185, 467-484 e415, doi:S0092-8674(21)01578-6 [pii]
269 10.1016/j.cell.2021.12.046 (2022).
270 21 Wolter, N. et al. Early assessment of the clinical severity of the SARS-CoV-2 omicron variant in South
271 Africa: a data linkage study. Lancet 399, 437-446, doi:S0140-6736(22)00017-4 [pii]
272 10.1016/S0140-6736(22)00017-4 (2022).
273 22 Sigal, A. Milder disease with Omicron: is it the virus or the pre-existing immunity? Nat Rev Immunol 22,
274 69-71, doi:10.1038/s41577-022-00678-4
275 10.1038/s41577-022-00678-4 [pii] (2022).
276 23 Fouchier, R. A. & Smith, D. J. Use of antigenic cartography in vaccine seed strain selection. Avian Dis 54,
277 220-223, doi:10.1637/8740-032509-ResNote.1 (2010).
278 24 Lapedes, A. & Farber, R. The geometry of shape space: application to influenza. J Theor Biol 212, 57-69,
279 doi:10.1006/jtbi.2001.2347
280 S0022-5193(01)92347-1 [pii] (2001).
281 25 Halfmann, P. J. et al. SARS-CoV-2 Omicron virus causes attenuated disease in mice and hamsters. Nature,
282 doi:10.1038/s41586-022-04441-6
283 10.1038/s41586-022-04441-6 [pii] (2022).
284 26 Munoz-Fontela, C. et al. Animal models for COVID-19. Nature 586, 509-515, doi:10.1038/s41586-020-
285 2787-6 (2020).
286 27 Sia, S. F. et al. Pathogenesis and transmission of SARS-CoV-2 in golden hamsters. Nature 583, 834-838,
287 doi:10.1038/s41586-020-2342-5
288 10.1038/s41586-020-2342-5 [pii] (2020).
289 28 Smith, D. J. et al. Mapping the antigenic and genetic evolution of influenza virus. Science 305, 371-376,
290 doi:10.1126/science.1097211
291 1097211 [pii] (2004).
292 29 Mykytyn, A. Z. et al. SARS-CoV-2 entry into human airway organoids is serine protease-mediated and
293 facilitated by the multibasic cleavage site. Elife 10, doi:10.7554/eLife.64508 (2021).
294 30 Lamers, M. M. et al. SARS-CoV-2 Omicron efficiently infects human airway, but not alveolar epithelium.
295 bioRxiv, 2022.2001.2019.476898, doi:10.1101/2022.01.19.476898 (2022).
296 31 Wilks, S. H. et al. Mapping SARS-CoV-2 antigenic relationships and serological responses. bioRxiv,
297 2022.2001.2028.477987, doi:10.1101/2022.01.28.477987 (2022).
298 32 Potter, C. W. A history of influenza. Journal of Applied Microbiology 91, 572-579,
299 doi:https://doi.org/10.1046/j.1365-2672.2001.01492.x (2001).
300 33 Veneti, L. et al. Increased risk of hospitalisation and intensive care admission associated with reported
301 cases of SARS-CoV-2 variants B.1.1.7 and B.1.351 in Norway, December 2020 -May 2021. PLoS One 16,
302 e0258513, doi:10.1371/journal.pone.0258513
303 PONE-D-21-21292 [pii] (2021).
304 34 Lin, L., Liu, Y., Tang, X. & He, D. The Disease Severity and Clinical Outcomes of the SARS-CoV-2
305 Variants of Concern. Front Public Health 9, 775224, doi:10.3389/fpubh.2021.775224 (2021).
306 35 Funk, T. et al. Characteristics of SARS-CoV-2 variants of concern B.1.1.7, B.1.351 or P.1: data from seven
307 EU/EEA countries, weeks 38/2020 to 10/2021. Euro Surveill 26, doi:10.2807/1560-
308 7917.ES.2021.26.16.2100348 (2021).
309 36 Lamers, M. M. et al. SARS-CoV-2 productively infects human gut enterocytes. Science,
310 doi:10.1126/science.abc1669 (2020).
311 37 Corman, V. M. et al. Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR. Euro
312 Surveill 25, doi:10.2807/1560-7917.ES.2020.25.3.2000045 (2020).
313
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

314 Supplementary methods

315 Viruses and cell lines

316 HEK-293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented
317 with 10% FCS, sodium pyruvate (1 mM, Gibco), non-essential amino acids (1x, Lonza),
318 penicillin (100 IU/ mL), and streptomycin (100 IU/mL). VeroE6 cells were maintained in
319 DMEM supplemented with 10% FCS, HEPES (20 mM, Lonza) and sodium pyruvate (1 mM),
320 penicillin (100 IU/mL), and streptomycin (100 IU/mL). Calu-3 cells were maintained in Eagle’s
321 minimal essential medium (ATCC) supplemented with 10% FCS, penicillin (100 IU/mL), and
322 streptomycin (100 IU/mL). All cell lines were kept at 37˚C in a humidified CO 2 incubator.
323 Viruses propagated for infection in hamsters and neutralization assays were grown to passage 3
324 on Calu-3 cells (aside from 614G used in hamster inoculations, grown to passage 3 on Vero
325 cells), harvested 48-72 hours post-infection, cleared for 5 minutes at 1000 x g, aliquoted and
326 stored at -80˚C until use. All work with infectious SARS-CoV-2 was performed in a Class II
327 Biosafety Cabinet under BSL-3 conditions at Erasmus Medical Center.

328

329 Pseudovirus production

330 Pseudoviruses were produced as described previously29. Briefly, HEK-293T cells were
331 transfected with pseudotyping vectors from InvivoGen (Original D614, D614G, Alpha, Beta,
332 Kappa and Delta Spike) or kindly provided by Dr. Berend Jan Bosch (Omicron Spike). All Spike
333 expressing plasmids contained a deletion of the last 19 amino acids containing the Golgi
334 retention signal of the SARS-CoV-2 Spike protein. Twenty four hours post-infection cells were
335 infected with pseudoviruses expressing VSV-G. Two hours posti-infection, cells were washed
336 three times with Opti-MEM I (1×) + GlutaMAX and replaced with medium containing anti-
337 VSV-G neutralizing antibody (Absolute Antibody). The supernatant was collected after 24 and
338 48 hr, cleared by centrifugation at 2000 x g for 5 min, and stored at 4°C.

339

340 Virus titrations

341 Stock titers were determined by preparing 10-fold serial dilutions in Opti-MEM I (1X) +
342 GlutaMAX. One hundred ml of each dilution was added to monolayers of Calu-3 cells in the
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

343 same medium in a 12 well plate. After 4 hours at 37˚C, cells were overlaid with 1.2% Avicel
344 (FMC biopolymers) in Opti-MEM I (1X) + GlutaMAX (Gibco) for 72 hr. Avicell was then
345 removed and plates were fixed in formalin, permeabilized in 70% ethanol and washed in PBS.
346 Cells were blocked in 3% BSA (bovine serum albumin; Sigma) in PBS, followed by rabbit anti-
347 nucleocapsid (Sino biological; 1:2000) in PBS containing 0.1% BSA. Plates were washed in PBS
348 then stained with goat anti-mouse Alexa Fluor 488 (Invitrogen; 1:4000) in PBS containing 0.1%
349 BSA. Plates were then washed again in PBS and scanned on the Amersham Typhoon
350 Biomolecular Imager (channel Cy2; resolution 10 mm; GE Healthcare). Eight hour titers of
351 SARS-CoV-2 and pseudoviruses were determined by preparing 10-fold serial dilutions in Opti-
352 MEM I (1X) + GlutaMAX. Thirty ml of each dilution was added to monolayers of Calu-3 or
353 VeroE6 cells in the same medium in a 96 well plate. After 16 hours at 37˚C, Pseudovirus
354 infected plates were fixed in paraformaldehyde and washed in PBS. After 8 hours at 37˚C SARS-
355 CoV-2 infected cells were fixed in formalin, permeabilized in 70% ethanol and washed in PBS.
356 Cells were blocked in 3% BSA (bovine serum albumin; Sigma) in PBS, followed by rabbit anti-
357 nucleocapsid (Sino biological; 1:2000) in PBS containing 0.1% BSA. Plates were washed in PBS
358 then stained with goat anti-mouse Alexa Fluor 488 (Invitrogen; 1:4000) in PBS containing 0.1%
359 BSA. SARS-CoV-2 and Pseudovirus infected cells were next stained with Hoechst
360 (ThermoFisher) and washed with PBS. Cells were imaged using the Opera phenix spinning disk
361 confocal HCS system (Perkin Elmer) equipped with a 10x air objective (NA 0.3) and 405 nm
362 and 488 nm solid state lasers. Hoechst and GFP were detected using 435-480 nm and 500-550
363 nm emission filters, respectively. Nine fields per well were imaged covering approximately 50%
364 of the individual wells. The number of green fluorescent protein (GFP) positive/ Alexa Fluor 488
365 positive infected cells were quantified using the Harmony software (version 4.9, Perkin Elmer).

366

367 Hamster infections

368 Female Syrian golden hamsters (Mesocricetus auratus; 6 weeks old; Janvier, France) were
369 handled in an ABSL-3 biocontainment laboratory. Groups of animals (n=4) were inoculated
370 intranasally with 1.0x10^5 PFU (614G, Alpha, Beta, Gamma, Zeta, Mu) or 5.0x10^4 PFU
371 (Delta, Omicron) in a total volume of 100µl per animal. Nasal washes (250µL) were taken at 7
372 dpi. All animals were sacrificed 26 dpi. Research involving animals was conducted in
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

373 compliance with the Dutch legislation for the protection of animals used for scientific purposes
374 (2014, implementing EU Directive 2010/63) and other relevant regulations. The licensed
375 establishment where this research was conducted (Erasmus MC) has an approved OLAW
376 Assurance # A5051-01. Research was conducted under a project license from the Dutch
377 competent authority and the study protocol (#17-4312) was approved by the institutional Animal
378 Welfare Body. Animals were housed in groups of 2 animals in filter top cages (T3, Techniplast),
379 in Class III isolators allowing social interactions, under controlled conditions of humidity,
380 temperature and light (12-hour light/12-hour dark cycles). Food and water were available ad
381 libitum. Animals were cared for and monitored (pre- and post-infection) daily by qualified
382 personnel. All animals were allowed to acclimatize to husbandry for at least 7 days. For unbiased
383 experiments, all animals were randomly assigned to experimental groups. The animals were
384 anesthetized (3-5% isoflurane) for all invasive procedures. Hamsters were euthanized by cardiac
385 puncture under isoflurane anesthesia and cervical dislocation.

386

387 Viral RNA quantification using qRT-PCR

388 RNA extraction was performed as described previously36. Briefly, 60 μL of sample was lysed in
389 90 μL of MagnaPure LC Lysis buffer (Roche) followed by a 15 minute incubation with 50 μL
390 Agencourt AMPure XP beads (Beckman Coulter). Beads were washed twice with 70% ethanol
391 on a DynaMag-96 magnet (Invitrogen) and eluted in 50 μL diethylpyrocarbonate treated water.
392 qRT-PCR was performed using primers targeting the E gene37 and comparing the Ct values to a
393 standard curve derived from a virus stock titrated on Calu-3 cells.

394

395 Plaque reduction neutralization assay

396 Post-vaccination and post-Omicron BA.1 infection sera was kindly obtained in the scope of the
397 healthcare worker study performed at the Erasmus MC. This study was approved by the
398 institutional review board of the Erasmus MC (medical ethical committee, MEC-2020-0264).
399 PRNT50 assays were performed as described previously. Briefly, sera was heat inactivated for
400 30 min at 56˚C. Sera were 3-fold serially diluted in 60 μL Opti-MEM I (1X) + GlutaMAX
401 (Gibco). Four hundred PFU (based on 8 hr titrations) in 60 μL were added per well to a final
402 volume of 120 μL and a serum dilution of 1:20 in the first well. Plates were incubated for 1 hr at
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

403 37˚C. Next 100 μL of virus and serum mix was added to confluent monolayers of Calu-3 or
404 VeroE6 cells. SARS-CoV-2 infected plates were incubated for 8 hr at 37˚C before fixing in
405 formalin and permeabilizing in ethanol. Plates were then washed in PBS and stained as described
406 for virus titrations. Pseudovirus infected plates were incubated for 16 hr at 37˚C before fixing in
407 paraformaldehyde and washing in PBS. Nuclei were stained with Hoechst for 30 min. Cells were
408 imaged using the Opera phenix spinning disk confocal HCS system and the number of GFP-
409 positive/ Alexa Fluor 488 positive infected cells were quantified using the Harmony software as
410 described above. The PRNT50 was calculated using Graphpad Prism 9 based on non-linear
411 regression.

412

413 Illumina sequencing

414 Deep sequencing was performed as described previously16. Briefly, RNA was extracted as
415 described above followed by cDNA synthesis and PCR using the QIAseq® SARS-CoV-2 Primer
416 Panel kit (Qiagen) according to the manufacturer’s protocol. Omicron samples were amplified
417 with an additional 11 primers (as described by ARTIC V4.1 primer set). Amplicons were
418 purified using 0.8x AMPure XP beads followed by converting 100ng of DNA to paired-end
419 Illumina sequencing libraries using the KAPA HyperPlus library preparation kit (Roche).
420 Samples were pooled and analyzed on an Illumina sequencer V3 MiSeq flowcell (2x300 cycles).
421 The 614G virus used for hamster infections was cultured to passage 3 on Vero cells and its S
422 protein did not contain additional mutations. All other variant were propagated to passage 3 on
423 Calu-3 cells. For neutralization assays another passage 3 614G isolate (Bavpat-1) was used with
424 an identical S amino acid sequence (European Virus Archive Global #026 V-03883) and another
425 passage 3 Beta isolate with an identical S amino acid sequence. The 614G Bavpat-1 passage 3
426 sequence was identical to the passage 1 (kindly provided by Dr. Christian Drosten). The Alpha
427 (B.1.1.7), Gamma (P.1), Delta (B.1.617.2), Delta AY.4.2 (B.1.617.2 AY.4.2), Lambda (C.37),
428 Mu (B.1.621), Omicron BA.1 (B.1.1.529 BA.1) and Omicron BA.2 (B.1.1.529 BA.2) variant
429 passage 3 sequences were identical to the original respiratory specimens. Low coverage regions
430 in Spike were confirmed by Sanger sequencing. The Beta variant (B.1.351, used for hamster
431 inoculations) passage 3 sequence contained one synonymous mutation compared to the original
432 specimen: A26449C (Wuhan-Hu-1 position) in E. The Beta variant (B.1.351, used in
433 neutralization assays) passage 3 sequence contained two mutations compared the original
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

434 respiratory specimen: one synonymous mutation C13860T (Wuhan-Hu-1 position) in ORF1ab
435 and a L71P change in the E gene (T26456C, Wuhan-Hu-1 position). The Spike changes of all
436 variants compared to Wuhan-Hu-1 are indicated in Extended data Fig. 2. All isolate sequences
437 were submitted to Genbank. Hamster nasal wash sample sequences were identical to the input
438 viruses.
439

440 Antigenic cartography

441 Antigenic map construction was performed as described previously28. Briefly, antigenic
442 cartography is a method to quantify and visualize neutralization data. In an antigenic map, the
443 distance between antiserum point S and antigen point A corresponds to the difference between
444 the log2 of the maximum titer observed for antiserum S against any antigen and the log2 of the
445 titer for antiserum S against antigen A. Thus, each titer in a cross-titration can be thought of as
446 specifying a target distance for the points in an antigenic map. Modified multidimensional
447 scaling methods are then used to arrange the antigen and antiserum points in an antigenic map to
448 best satisfy the target distances specified by the neutralization data. The result is a map in which
449 the distance between the points represents antigenic distance as measured by the neutralization
450 assay in which the distances between antigens and antisera are inversely related to the log2 titer.
451 Because antisera are tested against multiple antigens, and antigens tested against multiple
452 antisera, many measurements can be used to determine the position of the antigen and antiserum
453 in an antigenic map, thus improving the resolution of the data. The antigenic maps were
454 computed with the Racmacs package (https://acorg.github.io/Racmacs/, version 1.1.18.) in R. A
455 web-based version of the software is available from https://www.antigenic-cartography.org/ .
456 The maps were constructed using 1000 optimizations, with the minimum column basis parameter
457 set to “none”.

458 Acknowledgments:

459 The present manuscript was part of the research program of the Netherlands Centre for One
460 Health

461 Funding

462 This work was financially supported by the Netherlands Organization for Health Research and
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

463 Development (ZONMW) grant agreement 10150062010008 to B.L.H. , the Health~Holland

464 grants EMCLHS20017 to R.D.d.V. and LSHM19136 to B.L.H co-funced by the PPP Allowance

465 made available by the Health~Holland, Top Sector Life Sciences & Health, to stimulate public

466 –private partnerships, and the European Union’s Horizon 2020 research and innovation program

467 under grant no. 101003589 (RECoVER: M.P.G.K.) and EU funding grant agreement number

468 874735(VEO). B.L.H., R.A.M. F., B.R. and M.P.G. are supported by the NIH/NIAID Centers of

469 Excellence for Influenza Research and Response (CEIRR) under contract 75N93021C00014
470 -Icahn School of Medicine at Mt. Sinai.

471 Author contributions:

472 Conceptualization: A.Z.M, M.M.L, R.F, B.R, B.L.H

473 Formal analysis: A.Z.M, M.R, A.K, M.E.R, M.M.L

474 Funding acquisition: B.H, M.P.K, R.R.d.V

475 Investigation: A.Z.M, M.R, D.S, T.B, PvD, F.C, T.B, M.d.W, M.E.R

476 Resources: B.B.O.M, C.G, R.R.d.V

477 Supervision: M.M.L, R.F, B.R, B.L.H

478 Writing—original draft: A.Z.M, M.M.L, B.L.H

479 Writing—review and editing: all authors reviewed and edited the final version.

480 Competing interests:

481 The authors declare that they have no competing interests.

482 Data and materials availability:

483 The sequences generated in this study have been deposited in GenBank. All data needed to

484 evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.

485 Figure Legends:

486 Fig. 1. Neutralizing activity of human post-vaccination sera against Omicron BA.1 and
487 BA.2. a-c, Neutralization titers against 614G, Delta, Omicron BA.1 and Omicron BA.2 of
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

488 vaccinated individuals after vaccination with one (a), two (b) or three (c) doses of BNT162b2.
489 Geometric mean is displayed above each graph. PRNT50: plaque reduction neutralization titers
490 resulting in 50% plaque reduction. Dotted lines indicate limit of detection.

491 Fig. 2. SARS-CoV-2 variants efficiently infect hamsters, inducing high neutralizing
492 antibody titers. a, Hamsters were inoculated with the indicated SARS-CoV-2 variants. Nasal
493 washes were collected 7 days post-infection and sequenced. At 26 days post-infection blood was
494 collected for serological analysis. b, Hamster sera was assessed for neutralizing antibodies
495 against pseudotyped and authentic SARS-CoV-2. PRNT50 values were used to generate
496 antigenic maps using a multidimensional scaling algorithm. c, RNA titers of nasal washes
497 collected one day post-infection. d, Homologous PRNT50 titers were determined using authentic
498 SARS-CoV-2. Geometric mean is displayed above graph. PRNT50: plaque reduction
499 neutralization titers resulting in 50% plaque reduction. Dotted line indicates limit of detection.
500 Error bars indicate SEM. Panels a and b were created with BioRender.com.

501 Fig. 3. Antigenic maps comparing neutralizations with SARS-CoV-2 pseudoviruses and
502 authentic SARS-CoV-2. a-b, MDS was used to create an antigenic map from the PRNT50 titers
503 generated against 614D, 614G, Alpha, Beta, Delta, Kappa and Omicron pseudoviruses on either
504 VeroE6 (a) or Calu-3 (c) cells. c, MDS was used to create an antigenic map from the PRNT50
505 titers generated against 614G, Alpha, Beta, Delta and Omicron authentic SARS-CoV-2. d, Re-
506 display of antigenic map in c with lilac arrows indicating antigen positions in map a and black
507 arrows indicating antigen positions in map b. Viruses are shown as coloured circles and antisera
508 as squares with the same outline colour as the matching viruses. Viruses and antisera are
509 positioned in the map so that the distances between them are inversely related to the antibody
510 titers, with minimized error. The grid in the background scales to a 2-fold dilution of antisera in
511 the titrations. MDS: multidimensional scaling. PRNT50: plaque reduction neutralization titers
512 resulting in 50% plaque reduction.

513 Fig. 4. Antigenic cartography using authentic SARS-CoV-2. a-h, Neutralizing titers of
514 hamsters infected with either (a) 614G, (b) Alpha, (c) Beta, (d) Gamma, (e) Zeta, (f) Delta, (g)
515 Mu or (h) Omicron BA.1 viruses. i, Multidimensional scaling was used to create an antigenic
516 map utilizing PRNT50 titers generated from authentic SARS-CoV-2 on Calu-3 cells. See legend
517 to Fig. 3 for details. Subdivided by dotted ellipses are variants possessing overlapping
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

518 substitutions as indicated. Geometric mean is displayed above each graph. PRNT50: plaque
519 reduction neutralization titers resulting in 50% plaque reduction. Dotted lines indicate limits of
520 detection. Error bars indicate SEM.

521 Extended Data Fig. 1. Neutralizing activity of human sera post-Omicron BA.1 infection. a,
522 Neutralization titers against 614G, Delta, Omicron BA.1 and Omicron BA.2 after primary
523 Omicron BA.1 infection. b, Individual sera neutralization titers against 614G, Delta, Omicron
524 BA.1 and Omicron BA.2 from 6 lowest responders in a. PRNT50: plaque reduction
525 neutralization titers resulting in 50% plaque reduction. Dotted lines indicate limits of detection.

526 Extended Data Fig. 2. Spike substitutions in SARS-CoV-2 variants. Indicated are Spike
527 substitutions relative to Wuhan-Hu-1 present in all variants assessed. Indicated in red are
528 substitutions at the 484 position, in pink substitutions at the 501 position and in dark blue
529 substitutions at the 417 position. Differential substitutions between Omicron BA.1 and Omicron
530 BA.2 are indicated in light blue and purple respectively.

531 Extended Data Fig. 3. Pseudotyped SARS-CoV-2 neutralization titers. a-h, Neutralizing
532 titers against pseudotyped SARS-CoV-2 on VeroE6 cells of hamsters infected with either (a)
533 614G, (b) Alpha, (c) Beta, (d) Gamma, (e) Zeta, (f) Delta, (g) Mu or (h) Omicron BA.1. i-p,
534 Neutralizing titers against pseudotyped SARS-CoV-2 on Calu-3 cells of hamsters infected with
535 either (i) 614G, (j) Alpha, (k) Beta, (l) Gamma, (m) Zeta, (n) Delta, (o) Mu or (p) Omicron
536 BA.1. Geometric mean is displayed above each graph. PRNT50: plaque reduction neutralization
537 titers resulting in 50% plaque reduction. Dotted lines indicate limits of detection. Error bars
538 indicate SEM.

539 Extended Data Fig. 4. SARS-CoV-2 pseudovirus neutralization titers generated on VeroE6
540 cells correlate with titers generated on Calu-3 cells. a-e, Correlation of PRNT50 titers
541 obtained with pseudotyped (a) 614D, (b) 614G, (c) Alpha, (d) Beta, (e) Delta, (f) Kappa and (g)
542 Omicron BA.1 variants on VeroE6 cells compared to Calu-3 cells. PRNT50: plaque reduction
543 neutralization titers resulting in 50% plaque reduction.
544 Extended Data Fig. 5. Validation of representing neutralization data in two dimensional
545 antigenic maps. a-b, Scatter plot of detectable neutralizing titers on the x-axis and fitted titers
546 from the antigenic maps on the y-axis of maps generated in Fig. 3 and 4. Scatter plots generated
547 on data obtained with pseudovirus neutralizations on (a) VeroE6 cells (Fig. 3a) and (b) Calu-3
 bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

548 cells. c-d, Dimensionality tests indicating the mean RMSE of detectable neutralization titers in 1
549 to 5 dimensions for maps generated with data from neutralization on (c) VeroE6 cells or (d)
550 Calu-3 cells. For each dimension 100 antigenic maps with 1000 optimizations each are generated
551 while randomly excluding 10% of the titers. The mean RMSE is calculated by comparing the
552 predicted titers in each run to the actual measured titers on a log2 scale. RMSE: root mean square
553 prediction error.

554 Extended Data Fig. 6. Comparison of the antigenic map generated with pseudovirus on
555 VeroE6 cells to Calu-3 cells. Antigenic map from Fig. 3a with arrows indicating position of
556 antigens in antigenic map in Fig. 3b. See legend to Fig. 3 for details.

557 Extended Data Fig. 7. SARS-CoV-2 pseudovirus neutralization titers on VeroE6 cells
558 correlate with authentic SARS-CoV-2 neutralization titers. a-e, Correlation of PRNT50 titers
559 obtained with pseudotyped and authentic SARS-CoV-2 (a) 614G, (b) Alpha, (c) Beta, (d) Delta
560 and (e) Omicron variants on VeroE6 cells. PRNT50: plaque reduction neutralization titers
561 resulting in 50% plaque reduction.

562 Extended Data Fig. 8. SARS-CoV-2 pseudovirus neutralization titers on Calu-3 cells
563 correlate with authentic SARS-CoV-2 neutralization titers. a-e, Correlation of PRNT50 titers
564 obtained with pseudotyped and authentic SARS-CoV-2 (a) 614G, (b) Alpha, (c) Beta, (d) Delta
565 and (e) Omicron variants on Calu-3 cells. PRNT50: plaque reduction neutralization titers
566 resulting in 50% plaque reduction.

567 Extended Data Fig. 9. Uncertainty in antigen and antisera positions of antigenic map
568 generated with authentic SARS-CoV-2. a, Scatter plot of detectable neutralizing titers on the
569 x-axis and fitted titers from the antigenic map on the y-axis of map generated in Fig. 4i. b,
570 Dimensionality test indicating the mean RMSE of detectable neutralization titers in 1 to 5
571 dimensions for antigenic map generated in Fig. 4i. See Extended Fig. 4 for details. c, Antigenic
572 map from Fig. 4i with depicted regions (triangulation blobs) in which a particular antigen or
573 antisera can be positioned without increasing the total stress of the map above one unit. See Fig.
574 3 legend for details.
 A BNT162b2 (1x) B BNT162b2 (2x) C BNT162b2 (3x)
81920 225 238 17 29 1216 783 20 97 81920 3243 2388 289 493
81920
40960 40960 40960
20480 20480 20480
10240 10240 10240
5120 5120 5120
2560 2560 2560

PRNT50
PRNT50

1280 1280

PRNT50
1280
640 640 640
320 320 320
160 160 160
80 80 80
40 40 40
20 20 20
10 10 10
5 5 5
4G

ro lta

.1

.2

4G

ta

.1

.2
r o l ta
4G

ro A.1

.2
BA

BA

el

BA

BA
e

BA
e
61

61
D

61

D
m nB
D
n

n
n
ro

ro

ro
ic

ic

ic

ic
ic

ic
m

m
m
O

O
O

Fig. 1. Neutralizing activity of human post-vaccination sera against Omicron BA.1 and BA.2. a-c, Neutralization
titers against 614G, Delta, Omicron BA.1 and Omicron BA.2 of vaccinated individuals after vaccination with one (a),
two (b) or three (c) doses of BNT162b2. Geometric mean is displayed above each graph. PRNT50: plaque reduction
neutralization titers resulting in 50% plaque reduction. Dotted lines indicate limit of detection.
A

C 10 D 40960 3816 2966 2544 951 1613 5728 1326 1589

Log10 PFU equivalent / ml

20480
10240
8 5120
2560
1280
PRNT50

6 640
320
160
80
4 40
20
10
2 5
G ta

u
ta
a
4G

ta
a

.1
G ta

on u
ta
a

ta
4G

.1

M
ph

m
M
ph

Be

Ze

el

BA
Be

Ze

el

BA

61

am
61

am

D
Al
D
Al

n
ro
icr

ic
m
m

O
O

Fig. 2. SARS-CoV-2 variants efficiently infect hamsters, inducing high neutralizing antibody titers. a, Hamsters were inoculated with the indicated SARS-
CoV-2 variants. Nasal washes were collected 7 days post-infection and sequenced. At 26 days post-infection blood was collected for serological analysis. b,
Hamster sera was assessed for neutralizing antibodies against pseudotyped and authentic SARS-CoV-2. PRNT50 values were used to generate antigenic maps
using a multidimensional scaling algorithm. c, RNA titers of nasal washes collected one day post-infection. d, Homologous PRNT50 titers were determined
using authentic SARS-CoV-2. Geometric mean is displayed above graph. PRNT50: plaque reduction neutralization titers resulting in 50% plaque reduction.
Dotted line indicates limit of detection. Error bars indicate SEM. Panels a and b were created with BioRender.com.
 A B
Alpha

Beta

614D
614G

Delta

Kappa
Omicron
BA.1
Pseudovirus on VeroE6 cells Pseudovirus on Calu-3 cells

C D

Authentic SARS-CoV-2 on Calu-3 cells

Fig. 3. Antigenic maps comparing neutralizations with SARS-CoV-2 pseudoviruses and authentic SARS-CoV-2. a-b, MDS was used to create an
antigenic map from the PRNT50 titers generated against 614D, 614G, Alpha, Beta, Delta, Kappa and Omicron pseudoviruses on either VeroE6 (a) or Calu-3
(c) cells. c, MDS was used to create an antigenic map from the PRNT50 titers generated against 614G, Alpha, Beta, Delta and Omicron authentic SARS-CoV-
2. d, Re-display of antigenic map in c with lilac arrows indicating antigen positions in map a and black arrows indicating antigen positions in map b. Viruses
are shown as coloured circles and antisera as squares with the same outline colour as the matching viruses. Viruses and antisera are positioned in the map so
that the distances between them are inversely related to the antibody titers, with minimized error. The grid in the background scales to a 2-fold dilution of
antisera in the titrations. MDS: multidimensional scaling. PRNT50: plaque reduction neutralization titers resulting in 50% plaque reduction.
A 40960
9
614G
3816 3827 137 2252 1144 7597 38274658 227
0
34 137
B 40960
Alpha
753 2966 758 530 164 1764 1048 4358 922 27 26
C 40960
Beta
294 705 2544 699 151 268 255 529 505 48 82
D 40960
20480
Gamma
1714 3420 2435 951 203 10091 5721 4229 432 63 142
20480 20480 20480
10240 10240 10240 10240
5120 5120 5120 5120
2560 2560 2560 2560
1280 1280 1280 1280

PRNT50
PRNT50

PRNT50

PRNT50
640 640 640 640
320 320 320 320
160 160 160 160
80 80 80 80
40 40 40 40
20 20 20 20
10 10 10 10
5 5 5 5
a a a a a 2 a u 1 2
4G lph et m Zet elt .4. bd M A. A.
4G

ta
G e ta

O r on u
AY a
a

4G

ta

O r on u
La 4.2

n .1
.2

G e ta

AY ta
a

a
La 4.2

ro A.1

.2
4G

ta

O ron u
G e ta

AY ta
a

a
La 4.2

ro A.1

.2
M
ta elt

bd
ph

M
bd
ph

m
61

M
Ze

bd
ph

m
B m

ic BA
BA

Ze

BA
B B

Ze

el De
D AY m

BA
61

el De
A

61
B
am

B
am

.
61
a

B
D

B
m
Al

am

m
Al
B
La on ron

m
Al
G

n
ta r

n
ro
l ic ic

ta
ta
e
ic

ic

ic
el

ic

ic
D m m
m
m

m
D

D
O O

D
O

O
O
E 40960 2
Zeta
353 3518 3080 2238 1613 3316 183 2861 2902 49 172
2
F 40960
Delta
1822 7014 1652 1481 931 5728 9579 3710 1734 387 200
G 40960
Mu
2269 3920 697 6678 768 133 112
6 0 4
322 1326 174 404
9
H 40960
20480
Omicron BA.1
234 652 270 210 174 153 377 662 342 1589 197
20480 20480 20480
10240 10240 10240 10240
5120 5120 5120 5120
2560 2560 2560 2560
1280 1280 1280 1280

PRNT50
PRNT50

PRNT50

PRNT50
640 640 640 640
320 320 320 320
160 160 160 160
80 80 80 80
40 40 40 40
20 20 20 20
10 10 10 10
5 5 5 5
a a a a a 2 a u 1 2 a a a a a 2 a u 1 2
4G lph et m Zet elt .4. bd M A. A. 4G lph et m Zet elt .4. bd M A. A.

4G

ta

O ron u
G eta

AY ta
a

4G

ta

O ron Mu
La 4.2

ro A.1

.2

G eta

AY ta
a

a
La 4.2

n .1
.2
61 61

M
bd
ph

m
B m B m

bd
ph

m
B B B B

Ze

l
D AY m D AY m

BA

Ze

ic BA
BA
el De
A A

el De
61
a a

61
am

B
am

.
B
La n n La n n

Al

m
Al
G ta ro ro G ta ro ro

n
el ic ic el ic ic

ro
ta

ta
ic

ic
m m m m

ic
D D

m
O O O O

m
m
D

D
O

O
I
614G

N501Y
Beta
K417N/T Gamma
Lambda
Mu

Zeta
Delta
E484K

Omicron BA.2

L452R/Q
Delta AY.4.2
Alpha

Omicron BA.1

Fig. 4. Antigenic cartography using authentic SARS-CoV-2. a-h, Neutralizing titers of hamsters infected with either (a) 614G, (b) Alpha, (c)
Beta, (d) Gamma, (e) Zeta, (f) Delta, (g) Mu or (h) Omicron BA.1 viruses. i, Multidimensional scaling was used to create an antigenic map
utilizing PRNT50 titers generated from authentic SARS-CoV-2 on Calu-3 cells. See legend to Fig. 3 for details. Subdivided by dotted ellipses are
variants possessing overlapping substitutions as indicated. Geometric mean is displayed above each graph. PRNT50: plaque reduction
neutralization titers resulting in 50% plaque reduction. Dotted lines indicate limits of detection. Error bars indicate SEM.
 1280

A B 640
320
Sera 1 Sera 2 Sera 3

160

PRNT50
20480 80
10240 40
5120 20
2560 10 10 10
1280 204805
640 10240
PRNT50

10240 10240

4G

ta
4G

ro lta

.1

.2
.1

.2

4G

ta

.1

.2
el

BA

BA
BA

BA

el

BA

BA
e
320 5120

61
61

61
D
D

D
n

n
n

n
ro

ro
ro

ro

ro
160 2560

ic

ic
ic

ic

ic

ic
m

m
O

O
80 1280
40 640 Sera 4 Sera 5 Sera 6
20 320
10 160
PRNT50

5 80
40
4G

ta

.1

.2
el

BA

BA

20
61

n
ro

ro

10 10 10
ic

ic
m

5
O

4G

ta

4G
.1

.2

ta

4G
.1

.2

ro lta

.1

.2
el

el
BA

BA

BA

BA

BA

BA
e
61

61

61
D

D
n

n
ro

ro

ro

ro

ro
ic

ic

ic

ic

ic

ic
m

m
O

O
Extended Data Fig. 1. Neutralizing activity of human sera post Omicron BA.1 infection. a, Neutralization titers against
614G, Delta, Omicron BA.1 and Omicron BA.2 after primary Omicron BA.1 infection. b, Individual sera neutralization titers
against 614G, Delta, Omicron BA.1 and Omicron BA.2 from 6 lowest responders in a. PRNT50: plaque reduction
neutralization titers resulting in 50% plaque reduction. Dotted lines indicate limits of detection.
5’ UTR 6 7a 7b 8 10
ORF1a 3a
ORF1b S M N
E 3’ UTR

FP TM
614G NTD RBD HR1 HR2 IC

D614G TM
FP
Alpha NTD RBD HR1 HR2 IC

bioRxiv preprint doi: https://doi.org/10.1101/2022.02.23.481644; this version posted February 24, 2022. The copyright holder for this preprint
(which Δ69-70
was not certified by peer review) isN501Y
the author/funder.
D614G
All rights reserved. No reuse allowed
T716I D1118H
without permission.
Δ144 A570D P681H S982A TM
FP
Beta NTD RBD HR1 HR2 IC

L18F D215G K417N N501Y D614G

D80A Δ241-243 E484K A701V
FP TM
Gamma NTD RBD HR1 HR2 IC

P26S
T20N N501Y D614G V1176F
L18F D138Y R190S K417T E484K H655Y FP T1027I TM
Zeta NTD RBD HR1 HR2 IC

D614G V1176F
E484K FP TM

Delta NTD RBD HR1 HR2 IC

A222V
T19R G142D R158G T478K D614G
Δ156-157 L452R P681R FP D950N TM

Delta AY.4.2 NTD RBD HR1 HR2 IC

Y145H A222V
T19R R158G T478K D614G D1260N
Δ156-157 L452R P681R FP D950N TM

Kappa NTD RBD HR1 HR2 IC

E484Q D614G
E154K L452R P681R FP Q1071H TM

Lambda NTD RBD HR1 HR2 IC

T76I D253N L452Q D614G

G75V Δ246-252 F490S T859N TM
FP
Mu NTD RBD HR1 HR2 IC

Y145N
Y144S R346K N501Y D614G
T95I E484K P681H D950N TM
FP
Omicron BA.1 NTD RBD HR1 HR2 IC
G339D T547K
A67V Y145D S371L Y505H L981F
Δ69-70 ins215EPE S373P N501Y D614G P681H D796Y N969K
T95I L212I S375F Q498R N679K N764K N856K Q954H
Δ211 K417N G496S H655Y
Δ142-144 N440K Q493R
G446S E484A
T478K
S477N TM
FP
Omicron BA.2 NTD RBD HR1 HR2 IC
G339D
A27S S371F Y505H
Δ24-26 V213G S373P N501Y D614G P681H D796Y N969K
T19I G142D S375F Q498R N679K N764K Q954H
T376A Q493R H655Y
D405N E484A
R408S T478K
K417N S477N
N440K

Extended Data Fig. 2. Spike substitutions in SARS-CoV-2 variants. Indicated are Spike substitutions present in all variants assessed.
Indicated in red are substitutions at the 484 position, in pink substitutions at the 501 position and in dark blue substitutions at the 417 position.
Differential substitutions between Omicron BA.1 and Omicron BA.2 are indicated in light blue and purple respectively.
 A 614G B Alpha C Beta D Gamma
40960 826 1043 1620 421 985 643 53 40960 535 1589 1561 670 292 248 59 40960 116 99 614 626 157 302 70 40960 1639 2850 1941 1354 985 643 53

20480 20480 20480 20480

10240 10240 10240 10240
5120 5120 5120 5120
2560 2560 2560 2560
1280 1280 1280 1280

PRNT50

PRNT50
PRNT50
PRNT50

640 640 640 640

320 320 320 320
160 160 160 160
80 80 80 80
40 40 40 40
20 20 20 20
10 10 10 10
5 5 5 5

ta

ta
4D

4D
a

m K lta

m K lta
4G

n a

4G

n a
ta
4D
4G

m K lta
n a

ta
4D

.1

.1
a

m K l ta
4G

n a
.1

.1

ph

ro p

ph

ro p
ph

ro p

ph

ro p

Be

Be
Be

BA

BA
Be
BA

BA

e
e

i c ap

i c ap
i c ap

ic ap

61

61
61

61

61

61
61

61

D
D

Al

Al
D
Al

Al
O

O
O
E 40960
Zeta
1089 1959 1481 3242 1304 1797 602
F 40960
Delta
1490 3126 2288 1650 5545 3585 130
G 40960
Mu
1391 1675 2250 3916 405 471 311
H 40960 36
Omicron BA.1
69 22 200 108 840 1773
20480 20480 20480 20480
10240 10240 10240 10240
5120 5120 5120 5120
2560 2560 2560 2560
1280 1280
PRNT50

1280 1280
PRNT50

PRNT50

PRNT50
640 640 640 640
320 320 320 320
160 160 160 160
80 80 80 80
40 40 40 40
20 20 20 20
10 10 10 10
5 5 5 5
ta
4D
4G

m K lta
n a
.1

ta
4D

ta
4D
a

m K lta
ta

4G

n a

m K lta
n a
4D

4G
a

m K lta
n a
4G

.1

.1
.1
ph

ro p
Be

ph

ro p

ph

ro p
BA

ph

ro p
e

Be

Be
ic ap

BA
Be

BA
BA
61
61

e
e

ic ap

ic ap
ic ap

61

61
D

61

61

61
Al

61

D
Al

Al
D
Al
O

O
O

I 614G
J Alpha
K Beta
L Gamma
40960 1650 3068 2278 787 1742 1834 93 40960 328 612 1246 355 715 449 31 40960 129 220 530 1372 248 332 56 40960 833 1016 1021 601 3482 1279 181

20480 20480 20480 20480

10240 10240 10240 10240
5120 5120 5120 5120
2560 2560 2560 2560
1280 1280
PRNT50

1280 1280
PRNT50
PRNT50

PRNT50

640 640 640 640

320 320 320 320
160 160 160 160
80 80 80 80
40 40 40 40
20 20 20 20
10 10 10 10
5 5 5 5
a a a a a a a a
4D 4G ph et elt pp A.1 a a a a
4D 4G ph et elt pp A.1 4D 4G ph et elt pp A.1
ta
4D
4G

m K lta
n a
.1
ph

ro p

6 1 6 1 Al 6 1 6 1 Al
Be

BA
B D a B 6 1 6 1 Al B D a B
e
ic ap

B D a B
61
61

D
Al

K on K on K on
r r r
ic ic ic
m m m
O O O
O

M Zeta N Delta O Mu P Omicron BA.1

40960 955 2631 1898 1667 1158 1600 159 40960 1168 1508 1639 639 4937 2379 365 40960 992 897 1638 2467 834 1073 196 40960 177 135 229 90 325 598 3516

20480 20480 20480 20480

10240 10240 10240 10240
5120 5120 5120 5120
2560 2560 2560 2560
1280 1280 1280 1280
PRNT50
PRNT50
PRNT50

PRNT50

640 640 640 640

320 320 320 320
160 160 160 160
80 80 80 80
40 40 40 40
20 20 20 20
10 10 10 10
5 5 5 5
a a a a
a a a
4D 4G ph et elt ppa A.1 4D 4G ph et elt pp A.1
ta
4D

m K lta
4G

n a
.1
ta
4D

m K lta
4G

n a
.1

6 1 6 1 Al 6 1 6 1 Al
ph

ro p

B D a B
ph

ro p

B D a B
Be

BA
Be

BA

e
ic ap
e
i c ap

61
61
61

K on
61

K on
Al
D
Al

r r
ic ic
m m
O O
O
O

Extended Data Fig. 3. Pseudotyped SARS-CoV-2 neutralization titers. a-h, Neutralizing titers against
pseudotyped SARS-CoV-2 on VeroE6 cells of hamsters infected with either (a) 614G, (b) Alpha, (c) Beta, (d)
Gamma, (e) Zeta, (f) Delta, (g) Mu or (h) Omicron BA.1. i-p, Neutralizing titers against pseudotyped SARS-CoV-
2 on Calu-3 cells of hamsters infected with either (i) 614G, (j) Alpha, (k) Beta, (l) Gamma, (m) Zeta, (n) Delta, (o)
Mu or (p) Omicron BA.1. Geometric mean is displayed above each graph. PRNT50: plaque reduction
neutralization titers resulting in 50% plaque reduction. Dotted lines indicate limits of detection. Error bars indicate
SEM.
 A 10000 2
614D
B 10000 2
614G
C 10000 2
Alpha
D 10000 2
Beta

R =0.3062 R =0.1748 R =0.5721 R =0.5845

Pseudovirus VeroE6

Pseudovirus VeroE6

Pseudovirus VeroE6

Pseudovirus VeroE6
1000 1000 1000 1000

100 100 100 100

100 1000 10000 100 1000 10000 100 1000 10000 100 1000 10000
Pseudovirus Calu-3 Pseudovirus Calu-3 Pseudovirus Calu-3 Pseudovirus Calu-3

E 10000
Delta
F 10000
Kappa
G 10000
Omicron BA.1

R2=0.8126 R2=0.8300 R2=0.6649

Pseudovirus VeroE6

Pseudovirus VeroE6

Pseudovirus VeroE6
1000 1000 1000

100 100 100

100 1000 10000 100 1000 10000 100 1000 10000

Pseudovirus Calu-3 Pseudovirus Calu-3 Pseudovirus Calu-3

Extended Data Fig. 4. SARS-CoV-2 pseudovirus neutralization titers generated on VeroE6 cells correlate with titers generated on Calu-3 cells. a-e,
Correlation of PRNT50 titers obtained with pseudotyped (a) 614D, (b) 614G, (c) Alpha, (d) Beta, (e) Delta, (f) Kappa and (g) Omicron BA.1 variants on
VeroE6 cells compared to Calu-3 cells. PRNT50: plaque reduction neutralization titers resulting in 50% plaque reduction.
A 8 B 8

6 6
Fitted log titers

Fitted log titers

4 4

2 2

0 0

0 2 4 6 8 0 2 4 6 8
Measured log titers Measured log titers

C D
Mean RMSE detectable titers

Mean RMSE detectable titers

1.7 1.7

1.6 1.6

1.5 1.5

1.4 1.4

1.3 1.3

1.2 1.2
0 1 2 3 4 5 0 1 2 3 4 5
Dimension Dimension
Extended Data Fig. 5. Validation of representing neutralization data in two dimensional antigenic maps. a-b, Scatter plot of
detectable neutralizing titers on the x-axis and fitted titers from the antigenic maps on the y-axis of maps generated in Fig. 3 and 4. Scatter
plots generated on data obtained with pseudovirus neutralizations on (a) VeroE6 cells (Fig. 3a) and (b) Calu-3 cells. c-d, Dimensionality
tests indicating the mean RMSE of detectable neutralization titers in 1 to 5 dimensions for maps generated with data from neutralization on
(c) VeroE6 cells or (d) Calu-3 cells. For each dimension 100 antigenic maps with 1000 optimizations each are generated while randomly
excluding 10% of the titers. The mean RMSE is calculated by comparing the predicted titers in each run to the actual measured titers on a
log2 scale. RMSE: root mean square prediction error.
 A
Alpha

Beta

614D
614G

Delta

Kappa
Omicron
BA.1

Extended Data Fig. 6. Comparison of the antigenic map generated with

pseudovirus on VeroE6 cells to Calu-3 cells. Antigenic map from Fig. 3a with
arrows indicating position of antigens in antigenic map in Fig. 3b. See legend to Fig. 3
for details.
A 10000 2
614G
B 10000
Alpha
C 10000
Beta
2 2
R =0.3392 R =0.3919 R =0.6453
Authentic SARS-CoV-2

Authentic SARS-CoV-2
Authentic SARS-CoV-2
1000 1000 1000

100 100 100

100 1000 10000 100 1000 10000 100 1000 10000

Pseudovirus Pseudovirus Pseudovirus

D 10000
Delta E 10000
Omicron BA.1

R2=0.5231 R2=0.4604
Authentic SARS-CoV-2

Authentic SARS-CoV-2

1000 1000

100 100

100 1000 10000 100 1000 10000

Pseudovirus Pseudovirus

Extended Data Fig. 7. SARS-CoV-2 pseudovirus neutralization titers on VeroE6 cells correlate with authentic SARS-CoV-2
neutralization titers. a-e, Correlation of PRNT50 titers obtained with pseudotyped and authentic SARS-CoV-2 (a) 614G, (b)
Alpha, (c) Beta, (d) Delta and (e) Omicron variants on VeroE6 cells. PRNT50: plaque reduction neutralization titers resulting in
50% plaque reduction.
A 614G B Alpha C Beta
10000 10000 10000
R2=0.8958 R2=0.5129 R2=0.8703

Authentic SARS-CoV-2
Authentic SARS-CoV-2

Authentic SARS-CoV-2
1000 1000 1000

100 100 100

100 1000 10000 100 1000 10000 100 1000 10000

Pseudovirus Pseudovirus Pseudovirus

D 10000
Delta
E 10000
Omicron BA.1

R2=0.4860 R2=0.6534
Authentic SARS-CoV-2

Authentic SARS-CoV-2

1000 1000

100 100

100 1000 10000 100 1000 10000

Pseudovirus Pseudovirus

Extended Data Fig. 8. SARS-CoV-2 pseudovirus neutralization titers on Calu-3 cells correlate with authentic SARS-CoV-2
neutralization titers. a-e, Correlation of PRNT50 titers obtained with pseudotyped and authentic SARS-CoV-2 (a) 614G, (b)
Alpha, (c) Beta, (d) Delta and (e) Omicron variants on Calu-3 cells. PRNT50: plaque reduction neutralization titers resulting in 50%
plaque reduction.
A 10.0
B

Mean RMSE detectable titers

1.7

7.5 1.6
Fitted log titers

1.5
5.0

1.4

2.5
1.3

1.2
0.0
0 1 2 3 4 5
0.0 2.5 5.0 7.5 10.0
Measured log titers Dimension

C
614G
Beta
Lambda Gamma
Mu

Zeta
Delta
Omicron BA.2

Delta AY.4.2
Alpha

Omicron BA.1

Extended Data Fig. 9. Uncertainty in antigen and antisera positions of antigenic map generated with authentic
SARS-CoV-2. a, Scatter plot of detectable neutralizing titers on the x-axis and fitted titers from the antigenic map on
the y-axis of map generated in Fig. 4i. b, Dimensionality test indicating the mean RMSE of detectable neutralization
titers in 1 to 5 dimensions for antigenic map generated in Fig. 4i. See Extended Fig. 4 for details. c, Antigenic map
from Fig. 4i with depicted regions (triangulation blobs) in which a particular antigen or antisera can be positioned
without increasing the total stress of the map above one unit. See Fig. 3 legend for details.