Assignment of operational taxonomic units for metagenomic experiments

Abstract

Species characterization is a challenging task in metagenomics mainly due to data complexity and sequence fragmentation. Reliance on broadly conserved genes as phylotyping markers for assigning sequences to their operational taxonomic units is limited by high cost of sequencing and the fact that some markers do not span the entire phylogenetic range. To expand the phylotyping candidate loci for both microbial and viral communities, a set of markers including narrowly conserved genes was assessed against genes used in automated pipeline for phylogenomic analysis. The study assessed Clustering of bacteria and archae (microbial) and viral genomes‟ orthologs, identified suitable phylotyping markers in microbial and viral genomes and and compared the identified markers against AMPHORA. OrthoMCL analysis was employed for clusters generation followed by selection of suitable phylotyping candidates on the basis of sequence identity. Sequences with at least 70 and 55 percent identities for the microbial and viral databases respectively were selected and their phylotyping accuracy determined on simulated pyrosequencing datasets. Up to 4 times increase in sensitivity was achieved with increased number of markers from 31 to 145 and high specificity of 0.99 was recorded at all taxonomic ranks assessed. Nevertheless, there was notable increase in number of misclassified reads with increased reference markers. Viral markers on the other hand showed average sensitivity and high specificity of 0.97. Recorded increase in sensitivity indicates that use of narrowly conserved genes will improve the accuracy with which metagenomics reads are placed into their respective taxa. The results will hence equip medical practioners with disease outbreak preparedness and will facilitate pathogen management. In addition, accurate large-scale analysis of environmental samples to determine composition of microbial and viral communities in an ecosystem will be enhanced.