Russ joined the Wyss Institute in 2022 as a computational biologist and has since worked across the Institute’s various labs and programs. His work centers on assessing which analytical and machine learning methods apply to a given biological question and the data actually available for it, implementing those methods, and building alternatives where published approaches don’t fit.
He has worked with various data types: bulk and single-cell RNA-seq and multiome (paired RNA and ATAC); mass-spectrometry proteomics, including thermal proteome profiling; and brightfield microscopy images. Alongside standard analytics such as differential testing, data normalization and quality control, and pathway and enrichment interpretation, he has implemented HLA typing from single-cell RNA-seq data, MSA- and HMM-based protein mining methods, network models for drug repurposing prediction, and topic modeling and other similar dimensionality reduction methods. He has also worked extensively on predicting targets of age reversal in hematopoietic stem cells via a novel prediction pipeline with single cell multiomic data.
A constant factor in his work is the increasingly large gap between our ability to generate predictions and our ability to rigorously test them for feasibility and correctness and incorporate that knowledge back into a model. A broader theoretical interest of his is dimensionality reduction and the interpretation of embedding spaces: what do the resulting coordinates represent, and what do the abstractions discard? He welcomes conversations on these and other questions. He is also, inevitably, becoming dumber as AI does more of our work for us, so you’d better hurry up while he can still hold a conversation.
Russ holds a B.S. in Mathematics and an M.S. in Machine Learning and Bioinformatics from Oregon State University.