PRX Life

A new approach for antibody design uses generative diffusion models to guide the discovery of complementarity-determining regions with properties suitable for therapeutic applications.

AFEXplorer tailors AlphaFold’s predictions by incorporating user-defined constraints to generate alternative protein conformations that match specific functional states – a tool for exploring dynamic protein structures, such as kinase activation states and membrane transporter configurations.

Combining machine learning for structure prediction with quantum-inspired optimization for sequence selection, a new algorithm enables efficient and stable protein design.

Residue Level Alignment integrates protein sequence and structure information in a self-supervised model, improving speed and precision in predicting protein binding and structural stability.

KnotFold, a method that mutates conserved cysteine residues in multiple sequence alignments to simulate protein coevolution, enables AlphaFold2 to correct mispredicted disulfide connectivity patterns.

EMPOT uses unbalanced optimal transport to compute rigid-body transformations and align density maps, an essential step for comparing protein structures solved by cryo-EM, which outperforms existing methods in both map registration and model fitting for partial overlaps.

By integrating experimental data, this diffusion-based model guides the generation of protein ensembles and outperforms traditional structure-based approaches in producing realistic, functional conformations—particularly for flexible and intrinsically disordered proteins.

This method adds sequence homology information to protein language models, improving amino acid prediction and enabling efficient, controlled sequence generation and motif scaffolding with minimal extra computation.

A scalable approximation of pseudo-perplexity from language models enables fast, accurate prediction of mutation effects on protein function and stability.

This model uses AlphaFold2-like architecture to perform flexible docking of small molecules into protein targets, predicting both structure and affinity.