Selection of software and database for metagenomics sequence analysis impacts the outcome of microbial profiling and pathogen detection

Abstract

AimShotgun metagenomic sequencing analysis is widely used for microbial profiling of biological specimens and pathogen detection. However, very little is known about the technical biases caused by the choice of analysis software and databases. In this study, we evaluated shotgun metagenomics taxonomical profiling software to characterize the microbial compositions of biological samples collected from wild rodents.Method and ResultsUsing nine of the most widely used metagenomics software and four different databases, we analyzed shotgun metagenomic sequence data from three sets of wild rodent tissue samples. We demonstrated the discrepancies in results when different databases and software were used, which cause significant variation in microbial community characterizations. Our analysis also showed that these software differed in their ability to detect the presence of Leptospira, a major zoonotic pathogen of one health importance.ConclusionsSignificant differences in compositional profiles for the same dataset while using different databases and software combinations can result in confounding biological conclusions in microbial profiling.Significance and Impact of StudyThis study cautions that the selection of software and databases to analyze metagenomics data can influence the outcome and biological interpretation.

Impact of simulation and reference catalogues on the evaluation of taxonomic profiling pipelines

“…Although still widely used, MetaPhlAn3 [53][54][55] (along with its previous versions [35-37, 39-42, 50-52], see Supplementary Table S2), did not withstand the competition with more recent tools. Its primary handicap stemmed from its highly constricted native feature space, resulting in a substantial number of false negatives, thereby impairing all performance metrics.…”

Impact of simulation and reference catalogues on the evaluation of taxonomic profiling pipelines

“…Although still widely used, MetaPhlAn3 [51][52][53] (along with its previous versions [35-37, 39-42, 48-50], see Supplementary Table S2), did not withstand the competition with more recent tools. Its primary handicap stemmed from its highly constricted native feature space, resulting in a substantial number of false negatives, thereby impairing all performance metrics.…”

Taxonize-gb: A tool for filtering GenBank non-redundant databases based on taxonomy

“…Accordingly, the National Center for Biotechnology Information (NCBI) hosts GenBank, which contains a comprehensive collection of genetic sequence data, including DNA, RNA, and protein sequences submitted by researchers from around the globe (17). The NCBI offers up-to-date non-redundant protein and nucleotide databases (18), which seem to be the most suitable reference databases for analyzing shotgun sequences from environmental DNA samples (19). However, due to their comprehensiveness and regular updates, there are two major concerns.…”