Identificação de sequências da subfamília Orthocoronavirinae em metagenomas e metatranscriptomas e análise da diversidade genômica do coronavírus humano NL63

Abstract

The development of high-throughput sequencing technologies in the last decade resulted in large repositories of raw sequencing data and metagenomic projects. Alongside, computational biology and bioinformatics became important allies and many taxonomic classification tools were created. The pandemic of COVID-19 (Coronavirus Disease 2019) has shown the importance of better investigating coronaviruses, specially human coronaviruses that impact health and public services. Among them, HCoV-NL63 is a seasonal human coronavirus that spreads worldwide, causes mild to moderate respiratory disease, and shares the same receptor for host entry as SARS-CoV and SARS-CoV-2. This study uses public databases of sequences to 1) compares two bioinformatic tools, Kaiju and Burrows-Wheeler Aligner (BWA), in identifying from Orthocoronavirinae viral sequences in human metagenomes and 2) conduct a comprehensive phylogenetic analysis of HCoV-NL63. 1169 human samples were randomly selected (from 3670), and taxonomic classification was performed with Kaiju (double-step analysis with different reference databases). 150 samples were found positive for CoVs using Kaiju, however, it was not possible to assembly any genomes. In addition, all 3670 samples were analyzed using BWA, with negative results for CoVs, except for one sample containing sequences from HCoV NL63. Our findings suggest that the investigated samples did not have coronaviruses, only conserved and similar sequences between organisms, which was overestimated by the Kaiju tool. On the other hand, BWA can be used to identify viromes in various sequencing products. In this study, we also used 173 publicly available spike genes sequences from HCoV-NL63 to perform phylogenetic analysis. Maximum likelihood analysis resulted in eight subgenotypes (A1, A2, A3, B1, B2, C1, C2, and C3), consolidating the lineage B division into B1 and B2. Subgenotypes B2 (20.1%), B1 (19.5%), and C3 (16.1%) were the most prevalent. Positive selection was found in S1 (V57, G96, N431) and S2 (V1177, E1206) residues. Seven non-synonymous substitutions were detected in the RBD region. Compared to other genotypes, genotype B has a remarkable number of amino acid substitutions in the S1 region. Our analysis also detected a high prevalence of recombination events within the S1 region (12–307 amino acids). Genotype A, which has the highest number of recombination events, shows a constricted period of emergence in both time and place, displaying a pattern of purifying selection. Our results highlight the importance of continuous epidemiologic and phylogenetic surveillance, which is strongly recommended in order to predict the evolution of future variants.