aiqbench analysis
TadA-Bench introduces a million-variant benchmark for agentic protein engineering that prioritizes future wet-lab discovery over static data fitting. While the model performs well on random splits, it struggles significantly with future-round ranking and finite-budget candidate selection, highlighting that evolutionary coverage is more critical than local data density. The release provides a reproducible substrate for testing AI agents in directed-evolution campaigns, with data and code available on Hugging Face and GitHub.
TadA-Bench: A Million-Variant Benchmark for Future-Round Discovery Toward Agentic Protein Engineering
From arXiv
Full textarXiv:2606.02624v1 Announce Type: cross Abstract: AI for scientific discovery is entering an agentic era, where protein-engineering systems are expected to prioritize future wet-lab experiments rather than merely fit static measurements. We introduce TadA-Bench, a million-variant wet-lab replay benchmark from 31 TadA directed-evolution rounds for future-round discovery toward agentic protein engineering. TadA-Bench preserves the campaign chronology and defines a fixed-data replay task: given earlier experimental rounds, models rank variants that appear only in later rounds.
It provides aligned DNA, RNA, and protein views, and uses Seq2Graph, a graph-based label-unification pipeline, to reconcile noisy enrichment measurements into consistent cross-round activity labels. Random-split controls show strong interpolation, but future-round ranking and finite-budget candidate selection are much weaker. Controlled analyses suggest that evolutionary coverage is more informative than local data density, positioning TadA-Bench as a reproducible wet-lab replay substrate for future-round discovery toward agentic protein engineering; the data and code are released on Hugging Face and GitHub.
Abstract:AI for scientific discovery is entering an agentic era, where protein-engineering systems are expected to prioritize future wet-lab experiments rather than merely fit static measurements. We introduce TadA-Bench, a million-variant wet-lab replay benchmark from 31 TadA directed-evolution rounds for future-round discovery toward agentic protein engineering. TadA-Bench preserves the campaign chronology and defines a fixed-data replay task: given earlier experimental rounds, models rank variants that appear only in later rounds. It provides aligned DNA, RNA, and protein views, and uses Seq2Graph, a graph-based label-unification pipeline, to reconcile noisy enrichment measurements into consistent cross-round activity labels.
Random-split controls show strong interpolation, but future-round ranking and finite-budget candidate selection are much weaker. Controlled analyses suggest that evolutionary coverage is more informative than local data density, positioning TadA-Bench as a reproducible wet-lab replay substrate for future-round discovery toward agentic protein engineering; the data and code are released on Hugging Face and GitHub.
Accepted at the 43rd International Conference on Machine Learning (ICML 2026). Data: this https URL . Code: this https URL
Quantitative Methods (q-bio.QM); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2606.02624 [q-bio.QM]
(or arXiv:2606.02624v1 [q-bio.QM] for this version)
https://doi.org/10.48550/arXiv.2606.02624 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 268 words