Tools

Recent

Scaling down protein language modeling with MSA Pairformer

Fine-tuned models (ESM3ΔG, SaProtΔG) that predict per-residue protein stability (ΔG) and mutational effects from structure

Energy-based models of protein structure learned from sequence via denoising score matching

Exploiting structure hallucination within diffusion for protein design

Learning circular permutation-invariant representations to uncover putative protein homologs

One-shot design of functional protein binders

Visualizing protein, DNA, and RNA structures in 2D, for Google Colab and Jupyter

Making protein folding accessible to all via Google Colab

Design proteins using TrRosetta, RosettaFold and AlphaFold

Inverting All-Atom Structure Prediction Model for Generalized Biomolecular Binder Design

End-to-end learning of multiple sequence alignments with differentiable Smith-Waterman

End-to-end protein structures alignment

State-of-the-Art Estimation of Protein Model Accuracy using AlphaFold

Lightweight and fast prediction of ligand-binding sites

Web-server and database for predicting contacts. For source code see: C++, Python (Tensorflow), Python (Jax)

Categorical Jacobian to uncover pairwise relationships in sequence models. Implemented for: ESM2, ESM3, ProteinMPNN, Evo, gLM2

What do generative models learn from protein sequences?

Unified framework for modeling multivariate distributions in biological sequences

contact map alignment