Microbial abundances retrieved from sequencing data—automated NCBI taxonomy (MARS): a pipeline to create relative microbial abundance data for the microbiome modelling toolbox and utilising homosynonyms for efficient mapping to resources

Computational approaches to the functional characterisation of the microbiome, such as the Microbiome Modelling Toolbox, require precise information on microbial composition and relative abundances. However, challenges arise from homosynonyms—different names referring to the same taxon, which can hinder the mapping process and lead to missed species mapping when using microbial metabolic reconstruction resources, such as AGORA and APOLLO. We introduce the integrated MARS pipeline, a user-friendly Python-based solution that addresses these challenges. MARS automates the extraction of relative abundances from metagenomic reads, maps species and genera onto microbial metabolic reconstructions, and accounts for alternative taxonomic names. It normalises microbial reads, provides an optional cut-off for low-abundance taxa, and produces relative abundance tables apt for integration with the Microbiome Modelling Toolbox. A sub-component of the pipeline automates the task of identifying homosynonyms, leveraging web scraping to find taxonomic IDs of given species, searching NCBI for alternative names, and cross-reference them with microbial reconstruction resources. Taken together, MARS streamlines the entire process from processed metagenomic reads to relative abundance, thereby significantly reducing time and effort when working with microbiome data. MARS is implemented in Python. It can be found as an interactive application here: https://mars-pipeline.streamlit.app/along with a detailed documentation here: https://github.com/ThieleLab/mars-pipeline.