The origin and underlying driving forces of the SARS-CoV-2 outbreak

Shu-Miaw Chaw; Jui-Hung Tai; Shi-Lun Chen; Chia-Hung Hsieh; Sui-Yuan Chang; Shiou-Hwei Yeh; Wei-Shiung Yang; Pei-Jer Chen; Hurng-Yi Wang

SARS-CoV-2 is the seventh coronavirus found to infect humans. Among the other six, 69 SARS-CoV and MERS-CoV can cause severe respiratory illness, whereas 229E, HKU1, 70 NL63, and OC43 produce mild symptoms [5] . Current evidence strongly suggests that all 71 human associated coronaviruses originated from other animals, such as bats and rodents [5, 72 6]. While SARS-CoV-2 shares similar genomic structure with other coronaviruses [7-10], its 73 sequence differs substantially from some of the betacoronaviruses that infect humans, such as 74 SARS-CoV (approximately 76% identity), MERS-CoV (43% identity), and HKU-1 (33% 75 identity), but exhibits 96% similarity to a coronavirus collected in Yunnan Province, China 76 from a bat, Rhinolophus affinis. Therefore, SARS-CoV-2 most likely originated from bats [2, 77 dN in the RBD region is 0.023, approximately one third of the estimate for the rest of the 105 spike gene (0.068), dS in the RBD (0.710) is actually slightly higher than in the rest of the 106 spike sequence (0.651). This argues against the recombination scenario. We noticed that the 107 dS of the whole spike and the RBD, are 2-and 3-fold, respectively, higher than the genome 108 average. Since synonymous sites are typically less influenced by selection, the increased 109 divergence in dS may reflect an underlying elevated mutation rate. skewed. While the former shows excess of both high and low frequency mutations, the latter 116 mainly exhibits an excess of low frequency changes (Fig. 1a) . The excess of low frequency 117 mutations is consistent with the recent origin of SARS-CoV-2 [19]. Both population 118 reduction and positive selection can increase high frequency mutations [20, 21] . However, 119 the first scenario is contradicted by the recent origin of the virus. If positive selection has 120 been operating, we would expect an excess of high frequency non-synonymous as well as 121 synonymous changes. Furthermore, the ratio of nonsynonymous to synonymous changes is 122 2.46 (138/56) among singleton variants, but only 1.42 (17/12) among non-singletons. Both of 123 these observations suggest that the majority of amino acid-altering mutations are selected 124 against, with no positive selection in evidence.