TIGER: Text-Informed Generalized Enzyme-Reaction Retrieval
The article presents TIGER, a framework designed for enzyme-reaction retrieval in computational biology. It addresses challenges in existing methods, such as poor generalization and asymmetry in retrieval directions. TIGER utilizes a Dynamic Gating Network and a Structure-Shared Feature Projector to enhance enzyme representations and improve retrieval performance.
- ▪TIGER stands for Text-Informed Generalized Enzyme-Reaction Retrieval.
- ▪The framework improves enzyme-reaction mapping by leveraging protein-to-text generation models.
- ▪Extensive experiments show that TIGER outperforms state-of-the-art methods across various distributions.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.24489 (cs) [Submitted on 23 May 2026] Title:TIGER: Text-Informed Generalized Enzyme-Reaction Retrieval Authors:Yuhang Zhang, Keyan Ding, Peilin Chen, Han Liu, Can Lin, Ruixi Chen, Shiqi Wang, Qi Song View a PDF of the paper titled TIGER: Text-Informed Generalized Enzyme-Reaction Retrieval, by Yuhang Zhang and 6 other authors View PDF HTML (experimental) Abstract:Enzyme-reaction retrieval is a fundamental problem in computational biology, underpinning enzyme characterization, reaction mechanism elucidation, and the rational design of metabolic pathways and biocatalysts. As a bidirectional task, it entails both enzyme-to-reaction and reaction-to-enzyme mapping.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
