Microbial network inference and analysis has become a successful approach to generate biological hypotheses from microbial sequencing data. Network clustering is a crucial step in this analysis. Here, we present a novel heuristic flow-based network clustering algorithm, which equals or outperforms existing algorithms on noise-free synthetic data and performs well when a large percentage of the data are shuffled. manta comes with unique strengths such as the ability to identify nodes that represent an intermediate between clusters, to exploit negative edges and to assess the robustness of cluster membership. We demonstrate in two case studies how these properties help to gain a better understanding of the microbial community under study. manta does not require parameter tuning, is straightforward to install and run, and can easily be combined with existing microbial network inference tools. We therefore expect it to be useful in a wide range of microbial network applications. This repository contains an archived version of manta, in addition to all scripts and raw data used for the manuscript.

The latest version of manta can be found at its Github repository: https://github.com/ramellose/manta. manta has been tested for Python 3.5.