Beyond SMILES: Evaluating Agentic Systems for Drug Discovery
Edward Wijaya · 2026
Read Beyond SMILES: Evaluating Agentic Systems for Drug Discovery by Edward Wijaya (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-10163.txt.
Description
Agentic systems now plan syntheses, mine literature, and design molecules on their own, but how far do they generalize? A 2026 paper by Edward Wijaya tests six frameworks against fifteen task classes spanning peptide therapeutics, in vivo pharmacology, and tight-resource settings. The verdict is five capability gaps, not one.
The frameworks lack protein language models and peptide-specific prediction, bridge nothing between in vivo and in silico data, and rely on language-model inference with no path to training or reinforcement learning. They assume large-pharma resources and optimize single objectives, ignoring safety-efficacy-stability trade-offs. A paired probing experiment suggests the bottleneck is architectural: frontier models reason about peptides well, but no framework exposes that ability.
Catalog record
| CREATOR | Edward Wijaya |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 96 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-10163