Retracing Phylogenetic, Host and Geographic Origins of Coronaviruses with Coloured Genomic Bootstrap Barcodes: SARS-CoV and SARS-CoV-2 as Case Studies

Abstract

Phylogenetic trees of coronaviruses are difficult to interpret because they undergo frequent genomic recombination. Here, we propose a new method, coloured genomic bootstrap (CGB) barcodes, to highlight the polyphyletic origins of human sarbecoviruses and understand their host and geographic origins. The results indicate that SARS-CoV and SARS-CoV-2 contain genomic regions of mixed ancestry originating from horseshoe bat (Rhinolophus) viruses. First, different regions of SARS-CoV share exclusive ancestry with five Rhinolophus viruses from Southwest China (RfYNLF/31C: 17.9%; RpF46: 3.3%; RspSC2018: 2.0%; Rpe3: 1.3%; RaLYRa11: 1.0%) and 97% of its genome can be related to bat viruses from Yunnan (China), supporting its emergence in the Rhinolophus species of this province. Second, different regions of SARS-CoV-2 share exclusive ancestry with eight Rhinolophus viruses from Yunnan (RpYN06: 5.8%; RaTG13: 4.8%; RmYN02: 3.8%), Laos (RpBANAL103: 3.3%; RmarBANAL236: 1.7%; RmBANAL52: 1.0%; RmBANAL247: 0.7%), and Cambodia (RshSTT200: 2.3%), and 98% of its genome can be related to bat viruses from northern Laos and Yunnan, supporting its emergence in the Rhinolophus species of this region. Although CGB barcodes are very useful in retracing the origins of human sarbecoviruses, further investigations are needed to better take into account the diversity of coronaviruses in bats from Cambodia, Laos, Myanmar, Thailand and Vietnam.

Phylogeographic evolution of horseshoe bat sarbecoviruses in Vietnam and implications for the origins of SARS-CoV and SARS-CoV-2

“…2-7). Such results were predictable as previous studies on sarbecoviruses have shown that their genomes evolve through frequent recombination 13,23, 24 , which means that different regions of a genome or gene alignment can bring multiple phylogenetic signals. To better summarize evolutionary relationships among sarbecoviruses, we have therefore chosen to perform phylogenetic analyses based on the sliding window bootstrap (SWB) and SuperTRI methods (Supplementary Table 5).…”

“…This approach was designed to reveal the phylogenetic relationships supported by the largest proportion(s) of the genome 13, 24 . The five SuperTRI bootstrap consensus (SB) trees reconstructed from the SWB analyses based on five window sizes (400, 500, 600, 1000, and 2000 nt) are available in Supplementary Fig.…”

“…A method based on coloured genomic bootstrap (CGB) barcode has been recently developed to better interpret the multiplicity of phylogenetic signals present in the genome alignment of viruses undergoing frequent recombination 24 . A phylogenetic CGB barcode is a small image representing the genome of a virus in which the different colours show the best phylogenetic signals i.e., the GRPS containing the fewest number of closely-related viruses.…”

“…2-7). To better interpret the phylogeography of sarbecoviruses, phylogenetic CGB barcodes were transformed into geographic CGB barcodes using specific colours for geographic areas where closely-related viruses were collected 24 For SARS-CoV, the two phylogeographic indicators were statistically significant (at 0.01) and convergent to designate Yunnan as the province of origin: CTG = 96% (against 58% for China except Yunnan, 55% for northern Vietnam and 43% for southern Vietnam) and CEG = 39% (against 3% for China except Yunnan; 0.7% for northern Vietnam; and 0% for southern Vietnam).…”

“…Two methods previously described 13, 24 were applied on our alignment of 111 Sarbecovirus genomes to construct the coloured genomic bootstrap (CGB) barcodes of 15 Sarbecovirus species found in Vietnam, SARS-CoV, SARS-CoV-2 and all other sarbecoviruses closely-related to SARS-CoV-2. A phylogenetic CGB barcode is a small image representing the genome of a virus (or a specific gene) in which the different colours show the best robust phylogenetic signals i.e., containing the fewest number of closely-related viruses, detected in different regions of the alignment.…”