I invented and implemented this method as a way to automatically construct mechanistic explanations of arbitrary sample metadata (e.g. patient disease status) from single-cell RNA sequencing data.

I discovered that a very natural way to solve this problem is to use a form of cell attention to pick out cell types for which the expression of some gene(s) is maximally covariant with the response variable.

While existing tensor decomposition methods allow the reduced-rank regression of regular (i.e. aligned) tensors (e.g. CP-PLSR) and there exist unsupervised tensor decomposition methods for unaligned tensors (e.g. PARAFAC2, which our lab also applied to scRNA), there is currently no way to perform reduced-rank regression on unaligned tensors, which is the structure of scRNA-seq data.

This method automatically answers a very typical type of question that researchers ask in scRNA, of the form “which genetic programs in which subsets of cells are responsible for this phenotype”.

This is a work in progress. For the math and preliminary results, see my slides.