Multi-assignment clustering: Machine learning from a biological perspective

A common approach for analyzing large-scale molecular data is to cluster objects sharing similar characteristics. This assumes that genes with highly expression profiles are likely participating in a process. Biological systems extremely complex and challenging understand, proteins having multiple functions sometimes need be activated or expressed time-dependent manner. Thus, the strategies applied clustering of these molecules into groups key importance translation biologically interpretable findings. Here we implemented multi-assignment (MAsC) allows assigned clusters, rather than single ones as commonly used techniques. When high-throughput transcriptomics data, MAsC increased power downstream pathway analysis allowed identification pathways high biological relevance experimental setting studied. Multi-assignment also reduced noise partition by excluding low correlation all resulting clusters. Together, findings suggest our methodology facilitates knowledge. The method made available an R package on GitLab (https://gitlab.com/wolftower/masc).