Elucidating the Space of Enzymatic Reaction: A Unified Benchmark and Pretrained Model

Key points

• Existing reaction models primarily learn molecular transformations, whereas enzy- matic reactions depend jointly on molecular structure and catalytic function.1
• We formulate this problem as learning an enzymatic reaction space linking reactants, products, and Enzyme Commission (EC) annotations.1
• Benchmarking representative chemical and enzymatic models reveals a clear chemical-to-enzymatic domain gap, driven by limited domain data, catalytic-context dependency, and the difficulty of modeling large biomolecular structures.1
• To bridge this gap, we develop VenusRX, a unified T5-style sequence-to-sequence model for enzymatic reactions.1

Sources (1)

• [1]Elucidating the Space of Enzymatic Reaction: A Unified Benchmark and Pretrained Model
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 11:05 AM

  To characterize this space, we introduce VenusRX-Bench, a unified benchmark for forward reaction prediction, single-step retrosynthesis, and EC-number prediction.

  Existing reaction models primarily learn molecular transformations, whereas enzy- matic reactions depend jointly on molecular structure and catalytic function.

Extractive summary: sentences quoted from the sources.