Generative Retrieval for Reverse Target Fishing with a Proteome-Scale Benchmark

Key points

• We reformulate the task as generative retrieval and introduce GenFish.1
• Given a molecular fingerprint, GenFish autoregressively models codes conditioned on the query and preceding code, ranking proteins by identifier likelihood without explicitly labeling unrecorded pairs as negatives.1
• We introduce ProTargetBench, the first proteome-scale benchmark for reverse target fishing to jointly support measured-panel discrimination and full-catalogue recovery of experimentally recorded positives.1
• These results establish generative retrieval as a principled paradigm for reverse target fishing under incomplete bioactivity records.1

Sources (1)

• [1]Generative Retrieval for Reverse Target Fishing with a Proteome-Scale Benchmark
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 17, 04:56 AM

  Reverse target fishing identifies protein targets for a query molecule and is central to drug discovery and repurposing.

  We reformulate the task as generative retrieval and introduce GenFish.

Extractive summary: sentences quoted from the sources.