How Trustworthy Are the Genomic Sequences of SARS-CoV-2 in GenBank?

Well-annotated gene and genomic sequences serve as a foundation for making inferences in molecular biology and evolution and can directly impact public health. The first SARS-CoV-2 genome was submitted to the GenBank database hosted by the U.S. National Center for Biotechnology Information and used to develop the two successful vaccines. Conserved protein domains are often chosen as targets for developing antiviral medicines or vaccines. Mutation and substitution patterns provide crucial information not only on functional motifs and genome/protein interactions but also for characterizing phylogenetic relationships among viral strains. These patterns, together with the collection time of viral samples, serve as the basis for addressing the question of when and where the host-switching event occurred. Unfortunately, viral genomic sequences submitted to GenBank undergo little quality control, and critical information in the annotation is frequently changed without being recorded. Researchers often have no choice but to hold blind faith in the authenticity of the sequences. There have been reports of incorrect genome annotation but no report that casts doubt on the genomic sequences themselves because it seems theoretically impossible to identify genomic sequences that may not be authentic. This paper takes an innovative approach to show that some SARS-CoV-2 genomes submitted to GenBank cannot possibly be authentic. Specifically, some SARS-CoV-2 genomic sequences deposited in GenBank with collection times in 2023 and 2024, isolated from saliva, nasopharyngeal, sewage, and stool, are identical to the reference genome of SARS-CoV-2 (NC_045512). The probability of such occurrence is effectively 0. I also compile SARS-CoV-2 genomes with changed sample collection times. One may be led astray in bioinformatic analysis without being aware of errors in sequences and sequence annotation.