Boolean matrix factorisation for scRNA-seq

Ellen Visscher, Michael Forbes, Christopher Yau

Keywords: Boolean matrix factorisation, scRNA-seq, combinatorial optimisation

Track: Findings

Abstract: We present bfact, a Python package for performing accurate low-rank Boolean or binary matrix factorisation (BMF). bfact uses a hybrid combinatorial optimisation/machine learning (ML) approach based on a priori candidate factors generated from clustering algorithms. It selects the best disjoint factors before optionally performing a second ML algorithm to recover the BMF. We verify bfact in simulated settings and show it achieves strong signal recovery with a much lower rank on single-cell RNA-sequencing datasets from the Human Lung Cell Atlas.

General Area: Models and Methods

Specific Subject Areas: Unsupervised Learning

Data And Code Availability: Yes

Ethics Board Approval: No

Entered Conflicts: I confirm the above

Anonymity: I confirm the above

Submission Number: 94