Birnbaum and Keating present PottsMPNN, a sequence design model that learns a sequence-energy landscape from multiple sequence alignments. PottsMPNN outperforms other sequence design models and is a drop-in alternative to ProteinMPNN. Here’s the code: github.com/KeatingLab/P....

Show Post

Here’s a preprint from the Keating Lab: www.biorxiv.org/content/10.6.... Foster Birnbaum and Amy E. Keating demonstrate that sequence design models, such as ProteinMPNN, are limited because they were trained only on native sequences.

Show Post

We introduce epiplexity, a new measure of information that provides a foundation for how to select, generate, or transform data for learning systems. We have been working on this for almost 2 years, and I cannot contain my excitement! arxiv.org/abs/2601.03220 1/7

Show Post

A dataset of 40 million protein families and an autoregressive model of protein families. Great to see other protein Atlases popping up after Dayhoff!! @judewells.bsky.social @dmmiller597.bsky.social www.biorxiv.org/content/10.6...

Show Post

Built by CATH, TÜM and NVIDIA, ProFam-1 is our new open-source protein family language model (pfLM) designed to generate functional protein variants and predict fitness using in-context example sequences.

Show Post