Synteny

Synteny is a biotechnology company applying artificial intelligence to the problem of molecular recognition — how biological molecules selectively bind one another. Its platform, branded MANIFOLD, is described as unifying sequence, structure and physics-based representations to learn generalisable rules of molecular binding rather than relying only on observed examples, with the stated goal of making protein design programmable.

The company's applied focus is T cell receptor (TCR)-based therapeutics. Published research work covers fine-tuned protein language models for TCR sequence and repertoire stochasticity, training-free guidance and inference-time control of diffusion models for protein and molecule design, adaptation of the Boltz-2 structure-based affinity predictor to protein–protein affinity regression (evaluated on TCR3d and PPB-affinity), molecular dynamics with MM-PBSA for TCR–pMHC binding free energy prediction, and multiplexed library assays to estimate TCR affinity at scale. Synteny describes a "lab-in-the-loop" platform and an internal "Data Foundry" intended to generate large-scale, domain-specific experimental ground truth for model training, plus an agentic AI layer for autonomous scientific exploration. Work presented under an AACR25 heading describes combining generative AI with high-throughput screening to design TCRs for both selectivity and specificity.

Traction

Public traction consists mainly of research output: a series of technical articles and preprint-style write-ups on TCR language models, diffusion-model guidance, protein–protein affinity prediction, molecular dynamics for TCR–pMHC binding, and multiplexed affinity assays, plus work presented under an AACR25 heading on generative AI combined with high-throughput TCR screening. The company lists a leadership team, roughly a dozen named core team members, a board of directors and an advisory board, and is actively recruiting.