Available via license: CC BY 4.0
Content may be subject to copyright.
Journal of Human Genetics
https://doi.org/10.1038/s10038-020-0808-9
ARTICLE
SARS-CoV-2 genomic variations associated with mortality rate of
COVID-19
Yujiro Toyoshima1●Kensaku Nemoto1●Saki Matsumoto1●Yusuke Nakamura1●Kazuma Kiyotani 1
Received: 9 July 2020 / Revised: 10 July 2020 / Accepted: 12 July 2020
© The Author(s) 2020. This article is published with open access
Abstract
The coronavirus disease 2019 (COVID-19) outbreak, caused by SARS-CoV-2, has rapidly expanded to a global pandemic.
However, numbers of infected cases, deaths, and mortality rates related to COVID-19 vary from country to country.
Although many studies were conducted, the reasons of these differences have not been clarified. In this study, we
comprehensively investigated 12,343 SARS-CoV-2 genome sequences isolated from patients/individuals in six geographic
areas and identified a total of 1234 mutations by comparing with the reference SARS-CoV-2 sequence. Through a
hierarchical clustering based on the mutant frequencies, we classified the 28 countries into three clusters showing different
fatality rates of COVID-19. In correlation analyses, we identified that ORF1ab 4715L and S protein 614G variants, which
are in a strong linkage disequilibrium, showed significant positive correlations with fatality rates (r=0.41, P=0.029 and
r=0.43, P=0.022, respectively). We found that BCG-vaccination status significantly associated with the fatality rates as
well as number of infected cases. In BCG-vaccinated countries, the frequency of the S 614G variant had a trend of
association with the higher fatality rate. We also found that the frequency of several HLA alleles, including HLA-A*11:01,
were significantly associated with the fatality rates, although these factors were associated with number of infected cases and
not an independent factor to affect fatality rate in each country. Our findings suggest that SARS-CoV-2 mutations as well as
BCG-vaccination status and a host genetic factor, HLA genotypes might affect the susceptibility to SARS-CoV-2 infection or
severity of COVID-19.
Introduction
The novel betacoronavirus, severe acute respiratory syn-
drome coronavirus 2 (SARS-CoV-2), which causes cor-
onavirus disease 2019 (COVID-19), was first reported in
Wuhan, China in December 2019 [1,2]. Soon after, the
virus caused an outbreak in China and has spread to the
world. According to the World Health Organization, the
current outbreak of COVID-19 has nearly 11.5 million
confirmed cases worldwide with more than 530,000 deaths,
as of July 6, 2020. The SARS-CoV-2 genome comprises of
around 30,000 nucleotides organized into specific genes
encoding structural proteins and nonstructural proteins
(Nsps) [1,2]. Structural proteins include spike (S), envelope
(E), membrane (M), and nucleocapsid (N) proteins. Surface
S glycoprotein is involved in the interaction with the host’s
angiotensin-converting enzyme 2 (ACE2) receptor and
plays an important role in rapid human to human trans-
mission. Nsps, which are generated as cleavage products of
the open reading frame 1ab (ORF1ab) viral polyproteins,
assemble to facilitate viral replication and transcription.
RNA-dependent RNA polymerase, also known as Nsp12, is
the key component that regulates viral RNA synthesis with
the assistance of Nsp7 and Nsp8 [3]. In addition, five
accessory proteins are encoded by ORF3a, ORF6, ORF7a
ORF8, and ORF10 genes.
SARS-CoV-2 has rapidly spread around the world
compared with SARS-CoV appeared in 2002 and Middle
East respiratory syndrome coronavirus (MERS-CoV) in
2012. Although the estimated fatality rate in the confirmed
cases is 6.6% in SARS-CoV-2, which is lower than those of
SARS-CoV and MERS-CoV, 9.6% and 34.3%, respectively
*Kazuma Kiyotani
kazuma.kiyotani@jfcr.or.jp
1Project for Immunogenomics, Cancer Precision Medicine Center,
Japanese Foundation for Cancer Research, Tokyo 135-8550, Japan
Supplementary information The online version of this article (https://
doi.org/10.1038/s10038-020-0808-9) contains supplementary
material, which is available to authorized users.
1234567890();,:
1234567890();,:
[4], there is an urgent need for its effective treatment based
on antivirals and vaccines that reduce the mortality and
morbidity rates of COVID-19. However, up to now, the
causes of the large country-by-country difference of the
mortality rates related to COVID-19 have not been clearly
understood. Although many studies were conducted, the
effects of SARS-CoV-2 genetic variations and host genetic
factors remain elusive.
In this study, we comprehensively analyzed 12,343
SARS-CoV-2 genome sequences isolated from patients/
individuals in six geographic areas, including Asia, North
America, South America, Europe, Oceania, and Africa, and
investigated their correlations to the fatality rates in 28
different countries. We also investigated the associations
with BCG-vaccination status as well as human leukocyte
antigen (HLA), which is an important molecule to recognize
virus by our host immune system.
Methods
Coronavirus sequences
Full-length viral nucleotide sequence of the reference
SARS-CoV-2 (accession number MN908947) [1] was
downloaded from the NCBI GenBank. We used a total of
12,343 SARS-CoV-2 sequences isolated in 50 different
countries of six geographic areas, including 1062 sequences
from Asia, 4060 from North America, 99 from South
America, 6012 from Europe, 1028 from Oceania, and 82
from Africa regions, which were deposited in the Global
Initiative on Sharing Avian Influenza Data as of 7 May
2020 [5]. To analyze mutations based on countries, we used
the data of 28 countries in which more than 30 SARS-CoV-
2 sequences are available, among the 50 countries.
Mutation analysis
We analyzed mutations of SARS-CoV-2 as described pre-
viously [6]. Briefly, we first aligned each of the SARS-
CoV-2 sequences to the reference sequence SARS-CoV-
2_Wuhan-Hu-1 (accession number MN908947) using
BLAT software [7]. After the alignment, we extracted
nucleotide sequences corresponding to individual proteins
of SARS-CoV-2, translated them into amino acid sequen-
ces, and then compared them to reference amino acid
sequences of SARS-CoV-2_Wuhan-Hu-1 (accession num-
bers QHD43415-QHD43423, QHI42199).
Data acquisition
Data on numbers of confirmed cases and deaths related to
COVID-19 were obtained from the Worldometer
(https://www.worldometers.info/coronavirus/) on 7 May
2020 (Supplementary Table 1). Data of confirmed cases and
deaths in each state in the United States were obtained on 3
July 2020. Fatality rate in infected individuals was calcu-
lated from total infected cases and total deaths in each
country. The allelic frequencies of HLA genes were
obtained from The Allele Frequency Net Database [8]. Data
on BCG-vaccination status in each country were obtained
from the previous reports [9–11].
Statistical analyses
Continuous variables were compared using the Student’st
test. Fisher’s exact test was used to analyze differences of
mutation rates of SARS-CoV-2 among the different geo-
graphic areas. A hierarchical clustering was performed to
identify clusters corresponding to distinct subgroups with
the selected mutations using R package stats. Global maps
of clusters or mutations were drawn using R package
rworldmap. Pearson’s correlation was used to evaluate
correlations among mutant frequencies, HLA allele fre-
quencies and fatality rates. Haploview software was used to
analyze and visualize the haplotypes of SARS-CoV-2
mutations [12]. Multiple regression analysis was used to
test for an independent contribution of identified factors to
fatality rates of COVID-19. All statistical analyses were
carried out using the R statistical environment version 3.6.1.
Results
All replicating viruses, including coronavirus, continuously
accumulate genomic mutations that persist due to natural
selections. These mutations contribute to enhancement of
ability of viral proliferation and infection as well as an
escape from host immune attack. We firstly investigated
mutations in 12,343 SARS-CoV-2 genome sequences iso-
lated from patients/individuals in six different regions,
including Asia, North America, South America, Europe,
Oceania, and Africa. We identified a total of 1234 mutations
detected in at least two independent samples, including 131
mutations found at a frequency of more than 10% (Sup-
plementary Table 2). A hierarchical clustering using 16
common amino acid mutations classified 28 countries into
three clusters (Fig. 1a). The cluster 1 includes most of the
Asian countries we analyzed, whereas the cluster 2 includes
European and South American countries, and the cluster 3
includes European, North American, Oceania, African and a
few Asian countries (Fig. 1b). Comparing the mutations
among the three clusters, the average frequency of an L
variant of an ORF1ab P4715L in the countries classified as
the cluster 1 was 14.7%, which is significantly lower than
81.3% and 73.2%, respectively, in the countries classified as
Y. Toyoshima et al.
the clusters 2 and 3 (P=1.3 × 10−6and P=2.5 × 10−5,
respectively; Supplementary Fig. 1A). The ORF1ab 4715L
variant was detected at the significantly low frequency in
Asian countries compared with the other areas (20.8% vs.
others 54.9–86.8%, P=1.1 × 10−118; Supplementary
Fig. 2). Similarly, the frequency of a G variant of S protein
D614G was significantly lower in the cluster 1 than the
other two clusters (P=1.2 × 10−6and P=1.7 × 10−5,
respectively, for the clusters 2 and 3; Supplementary
Fig. 1B). In the cluster 2, K/R variants of N protein R203K/
G204R mutations were significantly enriched at 43.1%,
compared with the other clusters (5.2%, P=0.00011 for the
cluster 1 and 11.8%, P=5.6 × 10−7for the cluster 3;
Supplementary Fig. 1C). In addition, in the cluster 1, L and
F variants of N P13L and ORF1ab L3606F were pre-
dominantly enriched. The L variant of N P13L was found at
17.8%, which was significantly higher than 0.2% and 1.4%,
respectively, in the clusters 2 and 3 (P=0.012 and P=
0.0079; Supplementary Fig. 1D). The F variant of ORF1ab
L3606F was detected at a higher frequency of 40.1% than
10.0% and 7.9% in the clusters 2 and 3, respectively (P=
0.0035 and P=0.00050; Supplementary Fig. 1E). To fur-
ther analyze the mutational profile, we performed a haplo-
type analysis by drawing a linkage disequilibrium (LD) map
for SARS-CoV-2 viral genomes (Supplementary Fig. 3).
We found that ORF1ab 4715L and S protein 614G variants
were in a nearly complete LD (r2of LD =0.98 and D’=
1.00). N protein 203K/204R variants were additionally
acquired in the S protein 614G type of virus genome as
indicated as r2of LD =0.11 and D’=0.99. These results
indicate that S protein 614G-N protein 203K/204R haplo-
type characterizes the cluster 2.
We then investigated the association with the fatality
rates among confirmed cases in the 28 countries. In the
analysis comparing the fatality rates in the countries clas-
sified as either of the three clusters, average fatality rate of
the countries belonging to the cluster 2 was 9.3%, which
was higher than 3.0% and 5.8% of averages of the countries
(A)
(B) (C)
5
10
15
Fatality rate (%)
0
123
Mutation cluster
P= 0.026
P= 0.095
P= 0.19
Frequency of
mutants
(%)
75
50
25
0
123
Mutation cluster
D
P
L
L
P
T
A
G
P
Y
V
T
R
G
T
Q
Ref
614
4715
3606
84
13
2016
4489
251
5828
5865
378
175
203
204
265
57
Position
G
L
F
S
L
K
V
V
L
C
I
M
K
R
I
H
Mut
S
ORF1ab
ORF1ab
ORF8
N
ORF1ab
ORF1ab
ORF3a
ORF1ab
ORF1ab
ORF1ab
M
N
N
ORF1ab
ORF3a
Protein
Fig. 1 Clustering analysis of SARS-CoV-2 among 28 countries.
aHeatmap for the frequencies of SARS-CoV-2 mutants. The 28
countries were classified into three clusters based on the mutational
signature by a hierarchical clustering. Protein sequence based on the
SARS-CoV-2_Wuhan-Hu-1 sequence (GenBank accession number
MN908947) is used as a reference. Ref; amino acid in reference
SARS-CoV-2 sequence, Mut, amino acid in mutant SARS-CoV-2. bA
global mapping of the three clusters. cFatality rates according to the
clusters. Horizontal lines represent the means. The Student’sttest was
used to evaluate statistical significance
SARS-CoV-2 genomic variations associated with mortality rate of COVID-19
belonging to the clusters 1 and 3, respectively (P=0.026
and P=0.095; Fig. 1c). Among the mutations we analyzed,
the frequencies of ORF1ab 4715L-type and S 614G-type
viruses showed significant positive correlations with fatality
rates (Pearson’s correlation coefficient (r)=0.41, P=0.029
and r=0.43, P=0.022, respectively; Fig. 2a, b). Since the
clusters 2 and 3 were separated mainly by the frequency of
N 203K/204R, we also examined the correlations of this
variant or S 614G-N 203R/204G haplotype with fatality
rates; however, the correlations were not statistically sig-
nificant (r=0.31, P=0.11; r=0.27, P=0.17, respec-
tively; Supplementary Fig. 4A, B).
It is reported that fatality rates are different among the
areas or states in the United States [13]. When we compared
fatality rates among the three different areas, Western,
Central and Eastern, in the United States, an Eastern area
showed a higher fatality rate of 6.5% than that of 2.2% in a
Western area (P=0.010) and that of 3.9% in a Central area
(P=0.10; Fig. 3a). Therefore, we further investigated the
correlations of the variants with fatality rates in the
17 states. The frequencies of ORF1ab 4715L- and S protein
614G-types tended to show positive correlations with the
fatality rates (r=0.49, P=0.047; r=0.45, P=0.070,
respectively; Fig. 3b, c). Even when integrating the data of
17 states and the remaining 27 countries, the significant
correlations kept significant (r=0.38, P=0.014; r=0.39,
P=0.011, respectively; Supplementary Fig. 5A, B).
Several other factors are investigated in association with
mortality related to COVID-19. Ecological studies have
suggested that countries that mandate BCG vaccination for
the population have a lower number of infections and a
reduced mortality from COVID-19, although the associa-
tion is still controversial and the underlying mechanism has
not been clarified [9,14,15]. We classified 28 countries into
two groups according to the BCG-vaccination status as the
routine vaccine schedules. As a result, the mean of fatality
rates was significantly lower in 11 BCG-vaccinated coun-
tries than in 17 BCG-non-vaccinated countries (4.1% vs.
8.1%, P=0.031; Fig. 4a). When we divided BCG-
vaccinated countries into subgroups according to the
strains of BCG vaccine, we observed some differences in
the fatality rates among the countries by different strains of
BCG vaccine, but sample sizes of subgroups are too small
to evaluate statistical significance (Supplementary Fig. 6).
We also found the frequencies of S 614G variant showed a
trend of positive correlation with fatality rates (r=0.54,
P=0.090; Fig. 4b) in BCG-vaccinated countries, but such
correlation was not observed in BCG-non-vaccinated
countries (r=0.19, P=0.47; Fig. 4b). In addition, the
number of confirmed cases per million population was
significantly lower in BCG-vaccinated countries than in
BCG-non-vaccinated countries (710 vs. 2912, P=0.0012;
Fig. 4c). These results suggest that BCG-vaccination may
protect from SARS-CoV-2 infection by potentiation of
innate immune response; however, ORF1ab 4715L-type
and S protein 614G-type SARS-CoV-2 variants may escape
from the immune response.
Host genetic differences, especially in HLA loci, are
well-known to contribute to individual variations in the
immune responses to pathogens. We finally searched pep-
tide epitopes with a high binding affinity to HLA molecules,
which we previously reported [6], involving the two SARS-
CoV-2 mutations, ORF1ab P4715L and S D614G, to
investigate the association with host immune responses. We
found that several epitopes, which include the position of
ORF1ab P4715L or S protein D614G, are possibly bind to
HLA molecules, including HLA-A*02:06, HLA-A*11:01,
HLA-B*07:02, and HLA-B*54:01, although the mutated
epitopes from variant SARS-CoV-2 also predicted to bind
to HLA molecules at similar affinities (Supplementary
Table 3). Using the information of 21 countries in which
allele frequency data are available, we examined a
Fatality rate (%)
100
80
60
40
20
0
51015
Frequency of
ORF1ab 4715L variant (%)
Japan
Singapore
Korea China
England
Belgium
Netherlands
Spain
India
Thailand
Canada
France
Italy
Hungary
Sweden
USA
Greece
Brazil
Australia
Taiwan Germany
Finland
Switzerland
Luxembourg
Iceland
Portugal
Congo
Denmark
r= 0.41
P= 0.029
Frequency of
S 614G variant (%)
Fatality rate (%)
100
80
60
40
20
0
51015
Japan
Singapore
Korea China
India
Thailand
Australia
Taiwan England
Netherlands
Spain
Canada
USA
Belgium
France
Italy
Hungary
Sweden
Switzerland
Greece
Brazil
Germany
Luxembourg
Iceland
Congo
Denmark
Finland
Portugal
r= 0.43
P= 0.022
00
)
B
()
A
(
Fig. 2 Correlation analysis of variant frequencies of SARS-CoV-2 ORF1ab 4715L (a) or S 614G (b) with fatality rates of COVID-19 among 28
countries. Pearson’s correlation coefficients (r) were calculated. Colors of each dot were corresponding to the mutational clusters shown in Fig. 1a
Y. Toyoshima et al.
relationship between allele frequency of HLA-A*11:01 and
the fatality rates. Consequently, we found a significant
negative correlation (r=−0.61, P=0.0031; Fig. 5a).
Similarly, a trend of negative correlations was observed
between allele frequencies of HLA-A*02:06 or HLA-
B*54:01 and the fatality rates (r=−0.39, P=0.14, N=16
and r=−0.60, P=0.017, N=15; Fig. 5b, c). However,
the significant correlations became not statistically sig-
nificant after adjusted by the frequency of S 614G variant in
multiple regression (P=0.13 for HLA-A*11:01,P=0.73
for HLA-A*02:06 and P=0.45 for HLA-B*54:01). We also
found negative correlations between allele frequencies of
the HLAs and the number of confirmed cases per million
population (r=−0.43, P=0.054 for HLA-A*11:01, r =
−0.44, P=0.086 for HLA-A*02:06 and r=−0.52, P=
0.047 for HLA-B*54:01; Fig. 5d–f). Together, these
results suggest that differences in HLA allele frequencies
may explain different susceptibilities to SARS-CoV-2
infection among the countries, although there are
many other potential confounding factors needed to be
considered.
Discussion
The current outbreak of COVID-19 has rapidly spread
worldwide. Most patients with COVID-19 exhibit no or
mild to moderate symptoms, but ~15% progress to severe
pneumonia and about 5% eventually develop acute
respiratory distress syndrome, septic shock, and multiple
organ failures. The mortality rates related to COVID-19
vary among countries, generally known to be significantly
higher in European and North American countries than
those of Asian countries. Although several possibilities to
explain the differences in the mortality rates are demon-
strated, including the difference of age distribution, BCG-
vaccination status, virus genomic types, and genetic back-
grounds, nothing is clear at this moment. In this study, we
)B()A(
(C)
0
10
5
Fatality rate (%)
P= 0.15
P= 0.10
P= 0.010
New York
Connecticut
Virginia
New Jersey
Massachusetts
Pennsylvania
Florida
Texas
Louisiana
Wisconsin
Illinois
Ohio
Washington
California
Arizona
Utah
Oregon
0
20
40
60
80
100
0510
Fatality rate (%)
Frequency of
ORF1ab 4715L variant (%)
r= 0.49
P= 0.047
New York
Connecticut
Virginia
New Jersey
Massachusetts
Pennsylvania
Florida
Texas
Louisiana
Wisconsin
Illinois
Ohio
Washington
California
Arizona
Oregon
0
20
40
60
80
100
0510
Fatality rate (%)
Frequency of
S 614G variant (%)
r= 0.45
P= 0.070
Utah
Fig. 3 Association of variant frequencies of SARS-CoV-2 with fatality
rates of COVID-19 among 17 states in the United States. aFatality
rates in three different areas in the United States, Western, Central, and
Eastern. Horizontal lines represent the means. The Student’sttest was
used to evaluate statistical significance. b,cCorrelation analysis
between frequencies of SARS-CoV-2 ORF1ab 4715L (b) or S 614G
variants (c) and fatality rates. Pearson’s correlation coefficients (r)
were calculated
SARS-CoV-2 genomic variations associated with mortality rate of COVID-19
investigated the SARS-CoV-2 virus mutations and found
that the frequencies of S protein 614G variant and its highly
linked variant, ORF1ab 4715L, were significantly correlated
with fatality rates in the 28 countries and 17 states of the
United States.
The D614G spike mutation is the mutation detected in
Europe in the early phase and has widely spread around the
globe, especially to European and North American countries
[16–19]. Spike glycoprotein is essential for interaction with
ACE2 expressed in host cells and is important for viral
transmission [20,21]. Therefore, spike glycoprotein is most
vital hotspot of amino acid mutations when viruses acquire
mutations to enhance the virus-cell entry to adapt environ-
ments. Structural analyses indicated that S protein having a
D614G substitution is located on the surface of the virus and
interacts with ACE2. Concordant to our results, a few
reports demonstrated that S 614G variant was associated
with the mortality related to COVID-19 [13,22]. ORF1ab
P4715L is located in Nsp12, which is important for viral
RNA replication. We found significant associations between
these mutations and the fatality rates; however, the func-
tional significance of these mutations has not clarified yet.
Since immune responses through HLA and T cells are
important to protect from virus infections and also known to
be involved in the progression of COVID-19, we screened
epitopes around the mutations associated with fatality rates
(Supplementary Table 3). ORF1ab P4715L is located in the
epitope sequences of ORF1ab 4713–4721, FPPTSFGPL,
ORF1ab 4713–4722, FPPTSFGPLV, and ORF1ab
4715–4724, PTSFGPLVRK, which were predicted to have
strong binding affinities of 44, 41, and 45 nM to HLA-
B*07:02, HLA-B*54:01, and HLA-A*11:01, respectively.
In a computational prediction, corresponding mutated pep-
tides show higher binding affinities of 11, 12, and 23 nM.
Similarly, S D614G is located in the epitope sequences of
S606-615, NQVAVLYQDV, and S612-620, YQDVNC-
TEV. Both of wild-type and mutated epitopes were pre-
dicted to bind to HLA-A*02:06 at similar affinities. Among
them, the countries where the proportion of individuals with
HLA-A*11:01,HLA-A*02:06, and HLA-B*54:01 alleles are
-G
CB+
G
CB
Fatality rate (%)
5
10
15
0
P= 0.031
(A)
(C)
0
2,000
4,000
6,000
BCG+ BCG-
P= 0.0012
Cases per million population
Fatality rate (%)
100
80
60
40
20
0
Frequency of
S 614G variant (%)
51015
Japan
Singapore
Korea China
India
Thailand
Hungary
Brazil
Taiwan
Portugal
Congo
r= 0.54
P= 0.090
0
(B)
100
80
60
40
20
0
51015
England
Belgium
Netherlands
Spain
Canada
France
Italy
Sweden
USA
Greece
Australia
Germany
Finland
Switzerland
Luxembourg
Iceland
Denmark
r= 0.19
P= 0.47
0
Fatality rate (%)
BCG+ BCG-
Fig. 4 Association of BCG-vaccination status with fatality rates and
infected cases of COVID-19 among 28 countries. aFatality rates in
BCG-vaccinated (BCG+) and BCG-non-vaccinated countries (BCG−).
Horizontal lines represent the means. The Student’sttest was used to
evaluate statistical significance. bCorrelation analysis between
frequencies of S 614G variant of SARS-CoV-2 and fatality rates
in BCG+and BCG−countries. Pearson’s correlation coefficients (r)
were calculated. cNumber of infected cases in BCG+and
BCG−countries. Horizontal lines represent the means. The Student’st
test was used to evaluate statistical significance
Y. Toyoshima et al.
relatively high showed lower fatality rates as well as num-
ber of confirmed cases (Fig. 5). However, the significant
correlations with fatality rates became not significant after
adjusted by the frequency of S protein 614G-type virus in
multiple regression analysis. These results suggest that
individuals with HLA-A*11:01,HLA-A*02:06,orHLA-
B*54:01 might be protected from infection of SARS-CoV-
2, although further studies are needed to investigate the
effects of other potential confounding factors, such as dif-
ferent phases of outbreak, age of infected population,
management of the pandemic. In SARS-CoV and MERS-
CoV, several HLA genotypes have been reported to
associate with susceptibility or resistance, including HLA-
B*07:03, HLA-B*46:01, HLA-C*08:01, HLA-C*15:02,
HLA-DRB1*03:01, HLA-DRB1*11:01, and HLA-
DRB1*12:02 [23–26]. Although further studies are required
to elucidate whether such cytotoxic T lymphocytes targeting
the epitopes are present in peripheral blood in patients,
especially in severe patients, and also large scale case-
control association studies are needed to confirm the asso-
ciation of HLA genotype with susceptibility or disease
progression of SARS-CoV-2 infection, these findings in the
current study provide an important insight into treatment of
the current SARS-CoV-2 and prevention of the second
SARS-CoV-2 pandemic.
In summary, we comprehensively investigated SARS-
CoV-2 genome mutations, BCG-vaccination status, and HLA
genotypes in the 28 different countries and identified
significant associations of some virus genome variants
with the fatality rates. These results may explain, at least a
part of the differences of the SARS-CoV-2 infection or the
mortality rates related to COVID-19 among various countries.
Acknowledgements The super-computing resource was provided by
Human Genome Center, the Institute of Medical Science, the Uni-
versity of Tokyo (http://sc.hgc.jp/shirokane.html).
Compliance with ethical standards
Conflict of interest YN is a stockholder and a scientific advisor of
OncoTherapy Science, Inc. KK is a scientific advisor of Cancer Pre-
cision Medicine, Inc. This study is unrelated to the activity in these
companies.
Publisher’s note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as
long as you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons license, and indicate if
changes were made. The images or other third party material in this
article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not
included in the article’s Creative Commons license and your intended
use is not permitted by statutory regulation or exceeds the permitted
use, you will need to obtain permission directly from the copyright
holder. To view a copy of this license, visit http://creativecommons.
org/licenses/by/4.0/.
HLA-A*11:01 allele frequency (%)
15
10
5
00102030
Belgium
Taiwan
China
India Thailand
Singapore
Finland Japan
Korea
Australia
Sweden
Netherlands
England
France
Italy
Spain
Germany
Portugal
Greece USA
Brazil
r= -0.61
P= 0.0031
(B)
Fatality rate (%)
HLA-A*02:06 allele frequency (%)
0510
15
10
5
0
Taiwan
China
India
Thailand
Singapore
Australia
Sweden
Netherlands
Italy
Spain
Germany
Portugal
USA
Brazil
Japan
Korea
r= -0.39
P= 0.14
(A)
HLA-B*54:01 allele frequency (%)
0510
15
10
5
0
r= -0.60
P= 0.017
Taiwan
Thailand
France
Germany
Netherlands
Spain
USA China
Korea
India Japan
Singapore
Portugal
Brazil
Italy
(C)
(E)
(D) (F)
Cases per million
population
0
2,000
4,000
6,000
03010 20
Taiwan Thailand
Finland
France Germany
Sweden
Spain
USA
Australia China
Korea
India
Japan
Singapore
Portugal
Brazil
Belgium
Greece
Italy
England
Netherlands
r= -0.43
P= 0.054
HLA-A*11:01 allele frequency (%)
0
2,000
4,000
6,000
0510
Taiwan
Thailand
Germany
Sweden
Spain
USA
Australia China Korea
India Japan
Singapore
Portugal
Brazil
Italy
Netherlands
r= -0.44
P= 0.086
HLA-A*02:06 allele frequency (%)
0
2,000
4,000
6,000
0510
Taiwan
Thailand
France
Germany
Spain
USA
ChinaKorea
India Japan
Singapore
Portugal
Brazil
Italy
Netherlands
r= -0.52
P= 0.047
HLA-B*54:01 allele frequency (%)
Fig. 5 Association of HLA allele frequency with fatality rates and
infected cases of COVID-19 among countries. a–cCorrelation
between HLA-A*11:01 (a), HLA-A*02:06 (b), and HLA-B*54:01
(c) allelic frequencies and fatality rates of COVID-19. Numbers of
analyzed countries are 21, 16, and 15, respectively, for HLA-A*11:01,
HLA-A*02:06, and HLA-B*54:01. Pearson’s correlation coefficient (r)
was calculated. d–fCorrelation between HLA-A*11:01 (d), HLA-
A*02:06 (e), and HLA-B*54:01 (f) allelic frequency and number of
infected cases of COVID-19. Pearson’s correlation coefficient (r) was
calculated
SARS-CoV-2 genomic variations associated with mortality rate of COVID-19
References
1. Wu F, Zhao S, Yu B, Chen YM, Wang W, Song ZG, et al. A new
coronavirus associated with human respiratory disease in China.
Nature. 2020;579:265–9.
2. Zhou P, Yang XL, Wang XG, Hu B, Zhang L, Zhang W, et al. A
pneumonia outbreak associated with a new coronavirus of prob-
able bat origin. Nature. 2020;579:270–3.
3. Subissi L, Posthuma CC, Collet A, Zevenhoven-Dobbe JC, Gor-
balenya AE, Decroly E, et al. One severe acute respiratory syn-
drome coronavirus protein complex integrates processive RNA
polymerase and exonuclease activities. Proc Natl Acad Sci USA.
2014;111:E3900–9.
4. Wang C, Horby PW, Hayden FG, Gao GF. A novel coronavirus
outbreak of global health concern. Lancet. 2020;395:470–3.
5. Shu Y, McCauley J. GISAID: Global initiative on sharing all
influenza data—from vision to reality. Eur Surveill. 2017;22:30494.
6. Kiyotani K, Toyoshima Y, Nemoto K, Nakamura Y. Bioinfor-
matic prediction of potential T cell epitopes for SARS-Cov-2. J
Hum Genet. 2020;65:569–75.
7. Kent WJ. BLAT-the BLAST-like alignment tool. Genome Res.
2002;12:656–64.
8. Gonzalez-Galarza FF, McCabe A, Santos E, Jones J, Takeshita L,
Ortega-Rivera ND, et al. Allele frequency net database (AFND)
2020 update: gold-standard data classification, open access genotype
data and new query tools. Nucleic Acids Res. 2020;48:D783–8.
9. Ozdemir C, Kucuksezer UC, Tamay ZU. Is BCG vaccination
affecting the spread and severity of COVID-19? Allergy.
2020;75:1824–7.
10. Ritz N, Curtis N. Mapping the global use of different BCG vac-
cine strains. Tuberculosis. 2009;89:248–51.
11. Zwerling A, Behr MA, Verma A, Brewer TF, Menzies D, Pai M.
The BCG World Atlas: a database of global BCG vaccination
policies and practices. PLoS Med. 2011;8:e1001012.
12. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and
visualization of LD and haplotype maps. Bioinformatics.
2005;21:263–5.
13. Becerra-Flores M, Cardozo T. SARS-CoV-2 viral spike G614
mutation exhibits higher case fatality rate. Int J Clin Pract.
2020;00:e13525.
14. Gursel M, Gursel I. Is global BCG vaccination-induced trained
immunity relevant to the progression of SARS-CoV-2 pandemic?
Allergy. 2020;75:1815–9.
15. Hamiel U, Kozer E, Youngster I. SARS-CoV-2 rates in BCG-
vaccinated and unvaccinated young adults. JAMA.
2020;323:2340–1.
16. Forster P, Forster L, Renfrew C, Forster M. Phylogenetic network
analysis of SARS-CoV-2 genomes. Proc Natl Acad Sci USA.
2020;117:9241–3.
17. Koyama T, Weeraratne D, Snowdon JL, Parida L. Emergence of
drift variants that may affect COVID-19 vaccine development and
antibody treatment. Pathogens. 2020;9:E324.
18. Gonzalez-Reiche AS, Hernandez MM, Sullivan MJ, Ciferri B,
Alshammary H, Obla A, et al. Introductions and early spread of
SARS-CoV-2 in the New York City area. Science.
2020;369:297–301.
19. Deng X, Gu W, Federman S, du Plessis L, Pybus OG,
Faria N, et al. Genomic surveillance reveals multiple introductions
of SARS-CoV-2 into Northern California. Science. 2020. In press.
20. Letko M, Marzi A, Munster V. Functional assessment of cell entry
and receptor usage for SARS-CoV-2 and other lineage B beta-
coronaviruses. Nat Microbiol. 2020;5:562–9.
21. Hoffmann M, Kleine-Weber H, Schroeder S, Kruger N, Herrler T,
Erichsen S, et al. SARS-CoV-2 cell entry depends on ACE2 and
TMPRSS2 and is blocked by a clinically proven protease inhi-
bitor. Cell. 2020;181:271–80.e8.
22. Eaaswarkhanth M, Al Madhoun A, Al-Mulla F. Could the D614G
substitution in the SARS-CoV-2 spike (S) protein be associated with
higher COVID-19 mortality? Int J Infect Dis. 2020;96:459–60.
23. Lin M, Tseng HK, Trejaut JA, Lee HL, Loo JH, Chu CC, et al.
Association of HLA class I with severe acute respiratory syn-
drome coronavirus infection. BMC Med Genet. 2003;4:9.
24. Ng MH, Lau KM, Li L, Cheng SH, Chan WY, Hui PK, et al.
Association of human-leukocyte-antigen class I (B*0703) and
class II (DRB1*0301) genotypes with susceptibility and resistance
to the development of severe acute respiratory syndrome. J Infect
Dis. 2004;190:515–8.
25. Chen YM, Liang SY, Shih YP, Chen CY, Lee YM, Chang L,
et al. Epidemiological and genetic correlates of severe acute
respiratory syndrome coronavirus infection in the hospital with the
highest nosocomial infection rate in Taiwan in 2003. J Clin
Microbiol. 2006;44:359–65.
26. Hajeer AH, Balkhy H, Johani S, Yousef MZ, Arabi Y. Associa-
tion of human leukocyte antigen class II alleles with severe
Middle East respiratory syndrome-coronavirus infection. Ann
Thorac Med. 2016;11:211–3.
Y. Toyoshima et al.
