esmfold

Protein 3D structure prediction pipeline

RNA, mRNA, DNA, Peptide, Protein, Antibody and Complex | Folding

AIRFold Protein Structure Prediction System

proStAMP is a Snakemake pipeline for protein structure-based annotation of microbial SNPs, integrating genomic, functional, and structural features

De novo therapeutic peptide design pipeline using RFdiffusion and ESMFold to hallucinate NaV1.7 pain-channel blockers locally on CPU

ESMFold2 on Apple silicon: MLX/Metal folding trunk and sampler, 8-bit ESMC-6B. 7.5x faster than Biohub's code on an M1 Pro (unofficial port)

De novo computational design of tumor-selective pH-conditional (pH 6.5 vs 7.4), cross-species EGFR miniprotein binders for the Anthropic × Adaptyv 2026 Challenge.

MutFold | Human breast cancer associated mutant protein structure prediction and alignment application

Cost-aware protein folding inference — routes cache → AlphaFold DB → ESMFold. Never compute the same structure twice.

Automated discovery pipeline and 3D screening studio mining novel antimicrobial peptides (AMPs) from extremophile genomes.

The protein-engineering design loop as one Rust binary: mutate, fold, QC folders of models into Parquet, ESM-2 variant scoring, a terminal viewer, and a GA4GH TES 1.1 server. Science validated against mdtraj, FreeSASA, cctbx and PLIP in CI.

Code for "Ancient amino acid sets enable stable protein folds": protein design with reduced amino acid alphabets (RFdiffusion + ProteinMPNN), mutation simulations with ESMFold, and analysis of the designs.

Unified Python library and CLI for protein structure prediction and inverse folding.

Lightweight AI drug virtual screening platform for biologic industry. Nearly 20 years of expertise in biopharma R&D, manufacturing, quality and regulatory affairs.

Pre-registered self-consistency benchmark of ProteinMPNN VHH redesigns (PDB 8YVJ): ESMFold/AF2-multimer refolding + Rosetta interface cross-check, with shuffle/random/native controls

Local-first multi-agent AI for precision oncology (TP53): AlphaFold/ESMFold structures, ESM-2 variant scoring, ΔΔG stability, AutoDock Vina molecular docking and a per-agent reasoning audit benchmarked against ClinVar/IARC. Offline on 8 GB RAM; HL7 FHIR R4; 566 tests. Research use only.

Demonstration of ESMFold2 created by me. Hope is I deploy this in tandem with Neuroscience projects, particularly in understanding the contraindications of drugs on synapses.

Code and frozen results for a phosphorylation structural-effect benchmark evaluating AlphaFold3 / ESMFold (data-purity gate P1-P5)

Rooted, dual-method (ML + Bayesian) phylogeny of the bacterial Hfq family, with archaeal SmAP outgroup rooting and ESMFold/Foldseek structural verification