Publications

C3P: Contrastive promoter-protein pretraining yields representations capturing bacterial gene regulation
Cameron Dufault, Scott Xu, Alan M. Moses
Oral Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS 2026)
Zero-shot segmentation using embeddings from a protein language model identifies functional regions in the human proteome
Ami G. Sangster, Cameron Dufault, Haoning Qu, Denise Le, Julie D. Forman-Kay, Alan M. Moses
PLOS Computational Biology (November 2025)

We show that you can perform zero-shot segmentation of protein regions through applying change-point analysis of pLM embeddings!

Unsupervised Whole-Genome Representation Learning Captures Bacterial Phenotypes
Cameron Dufault, Alan M. Moses
Learning Meaningful Representations of Life (LMRL) Workshop at ICLR 2025

We show that by treating bacterial genomes as long sequences of conserved genomic elements (primarily Pfam domains), we can train a genomic language model which models genome-wide interactions in bacteria. Our learned whole-genome representations were more predictive of bacterial phenotypes than conventional manually engineered representations.

Transformers and genome language models
Micaela E. Consens, Cameron Dufault, Michael Wainberg, Duncan Forster, Mehran Karimzadeh, Hani Goodarzi, Fabian J. Theis, Alan Moses, Bo Wang
Nature Machine Intelligence (March 2025)

A review of the architectures and applications of genome language models.