Benchmarking biochemical networks generated by large language models

Jeevan Tewari, Benjamin W Dahl, B Adam Bates, Jason A Papin, Jeffrey J Saucerman, Martin Graña, Jeevan Tewari, Benjamin W Dahl, B Adam Bates, Jason A Papin, Jeffrey J Saucerman · 2026

Read Benchmarking biochemical networks generated by large language models by Jeevan Tewari, Benjamin W Dahl, B Adam Bates, Jason A Papin, Jeffrey J Saucerman, Martin Graña, Jeevan Tewari, Benjamin W Dahl, B Adam Bates, Jason A Papin, Jeffrey J Saucerman (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/PMC13427342.txt.

Description

Tewari and colleagues tested in 2026 whether large language models, the text-based artificial intelligence systems in wide use, can build models of biochemical networks. Such networks map how signals and chemicals inside cells guide cell choices, but building them by hand from the literature is slow. The team asked general-purpose models to rebuild known signal networks for heart cell growth, scar-type cell activation, and force sensing, plus a core metabolic network from a gut bacterium.

Tewari found the models recovered between a quarter and about two-thirds of the reactions in the curated signal networks. Logic models built from those networks then predicted responses to changes with low accuracy, between 6 and 33 percent. For metabolism the models did better on coverage, reaching up to 91 percent of reactions, though accuracy at predicting food use varied a lot. The authors supply benchmarks to guide later gains.

Catalog record

CREATORJeevan Tewari, Benjamin W Dahl, B Adam Bates, Jason A Papin, Jeffrey J Saucerman, Martin Graña, Jeevan Tewari, Benjamin W Dahl, B Adam Bates, Jason A Papin, Jeffrey J Saucerman
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE58 KB
LICENSEhttps://creativecommons.org/licenses/by/4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/article/MED/PMC13427342

Browse the full library · About the Tower of Babel