The study shows that mutant spectra of SARS-CoV-2 from diagnostic samples differed in point mutation abundance and complexity, and that significantly larger values were recorded in virus from patients who developed mild COVID-19 symptoms. The abundance of low frequency mutations marks a difference with other RNA viruses, and it may slow down the response of SARS-CoV-2 to escape specific selective constraints such as drugs or antibodies.

The possible relationship between coronavirus mutant spectrum complexity and disease associations has not been established. In the present study, we report an ultra-deep sequencing (UDS) analysis of the mutant spectrum of amplicons from the nsp12 (polymerase)- and spike (S)-coding region of thirty nasopharyngeal isolates of SARS-CoV-2 of the first COVID-19 pandemic wave (Madrid, Spain, April 2020), with a cut-off detection level of 0.5% for point mutations and deletions. The samples were classified according to the severity of ensuing COVID-19. Low frequency mutations and deletions, counted relative to the consensus sequence of the corresponding isolate, were overwhelmingly abundant. The number of different point mutations, mutations per haplotype and several diversity indices were significantly higher in those SARS-CoV-2 isolated from patients who developed mild disease than in those associated with moderate or severe disease. No such bias was observed with RNA deletions. The types and location of minority amino acids substitutions in nsp12 (polymerase) and S anticipate great potential for phenotypic diversification of SARS-CoV-2.