Abstract

Population-scale microbiome study poses specific challenges in data analysis, from enterotype analysis, identification of driver species, to microbiome-wide association of host covariates. Application of advanced data mining techniques to high-dimensional complex dataset is expected to meet the rapid advancement in large scale and integrative microbiome research. Here, we present tmap, a topological data analysis framework for population-scale microbiome study. This framework can capture complex shape of large scale microbiome data into a compressive network representation. We also develop network-based statistical analysis for driver species identification and microbiome-wide association analysis. tmap can be used for exploring variations in a population-scale microbiome landscape to study host-microbiome association.\n\nAvailability and implementationtmap is available at GitHub (https://github.com/GPZ-Bioinfo/tmap), accompanied with online documentation and tutorial (http://tmap.readthedocs.io).\n\nContacthttp://hk.zhou@siat.ac.cn