TACFormer generates full PROTACs conditioned on target protein sequences
Original title: TACFormer: Transformer-Based Generative Model for Target-Conditioned De Novo PROTAC Design
A GPT-style decoder uses ESM-2 embeddings of the protein of interest instead of designing only the linker.
Researchers introduce TACFormer, a transformer generative model for de novo design of proteolysis-targeting chimeras. Unlike linker-only methods that fix the warhead and E3-binding pieces, it conditions a GPT-style molecular decoder on pretrained ESM-2 representations of the protein-of-interest sequence. The authors report novel, chemically diverse molecules and POI-dependent shifts toward chemical neighborhoods seen for held-out targets, including new chemotypes.
Unconfirmed: Novelty, diversity, and target-dependent chemical shifts are author-reported and not independently confirmed.
- TACFormer generates the full PROTAC rather than only a linker.
- Conditioning uses pretrained ESM-2 embeddings of POI sequences.
- Authors report POI-dependent shifts in the generated chemical space.
- Generated molecules are described as novel relative to held-out data.
From the original article
1 min read at arxiv.orgTargeted protein degradation (TPD) represents a fundamental shift in drug discovery, moving beyond traditional target inhibition toward the event-driven removal of disease-causing proteins. A prominent strategy for TPD is the use of proteolysis-targeting chimeras (PROTACs), heterobifunctional molecules that selectively route a protein of interest (POI) to the ubiquitin-proteasome system for degradation.
By Yossra Gharbi · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.