RFW captures species-level full profile of metagenomic functions via integrating genome annotation information

Abstract

Functional profiling on whole-metagenome shotgun sequencing (WMS) has made great contribution to the development of our understanding in microbe-host interactions. In this work, we revealed that severe microbial functional information loss of current functional profiling methods existed at both taxon-level and community-level. To correct the distortion brought by information incompleteness, we developed a new framework, RFW (Reference based functional profile inference on WMS), to infer microbial functional abundance on WMS through utilizing information from genome function annotation and WMS taxonomic profile. Furthermore, we built up a new algorithm for absolute abundance change quantification of microbial function between groups under RFW framework. By applying RFW to several datasets related to autism spectrum disorder and colorectal cancer, we revealed that RFW greatly renewed our knowledge in downstream analysis, including differential microbial function identification, association analysis between microbial function and host phenotype, etc. RFW are open-source and freely available at https://github.com/Xingyinliu-Lab/RFW.