Seidr: a gene meta-network calculation toolkit

Abstract Summary Gene network analysis is a powerful tool to identify and prioritize candidate genes, especially from data sets where experimental design renders other approaches, such as differential expression analysis, limiting or infeasible. Numerous gene network inference algorithms have been published and are commonly individually applied in transcriptomics studies. It has, however, been shown that every algorithm is biased towards identifying specific types of gene interaction and that an ensemble of inference methods can reconstruct more accurate networks. This approach has been hindered by lack of an implementation to run and combine such combinations of inference algorithms. Here, we present Seidr: a toolkit to perform multiple gene network inferences and combine their results into a unified meta-network. Availability and implementation Seidr code is open-source, available from GitHub and also compiled in docker and singularity containers. It is implemented in C++ for fast computation and supports massive parallelisation through MPI. Documentation, tutorials and exemplary use are available from https://seidr.readthedocs.io. Contact bastian.schiffthaler@umu.se, nicolas.delhomme@slu.se