Jude Wells

This paper evaluates projected gradient descent (PGD) for hallucination-based protein binder design, finding it improves design efficiency and diversity compared to softmax-based methods.

The paper uses a PLM-guided energy landscape approach to design compact enzymes with preserved catalytic site structures, validated through structure prediction and molecular dynamics simulations.

ProFam is an open-source protein family language model that improves fitness prediction and design by leveraging evolutionary context through homologous sequences.

This paper improves molecular generation by steering chemical language models to control properties like solubility and lipophilicity using activation steering techniques.

Plixer generates drug-like molecules for protein pockets using a two-stage model that decouples spatial interaction learning from chemical grammar learning.

The paper introduces TED, a resource that segments and classifies over 370 million domains from the AlphaFold Protein Structure Database, expanding known structural domains and uncovering new structural interactions.

Chainsaw uses a fully convolutional neural network to predict protein domain boundaries, achieving higher accuracy than existing methods in domain segmentation.

VariPred uses a protein language model to predict missense variant pathogenicity, outperforming existing methods on the ClinVar dataset.