PuSH - Publication Server of Helmholtz Zentrum München

Kmicikiewicz, M. ; Fortuin, V. ; Szczurek, E.

PROSPERO: Active Learning for Robust Protein Design Beyond Wild-Type Neighborhoods.

In: (39th Conference on Neural Information Processing Systems, NeurIPS 2025, 02-07 December 2026, San Diego). 2025. 145223-145256 (Advances in Neural Information Processing Systems ; 38)
Designing protein sequences of both high fitness and novelty is a challenging task in data-efficient protein engineering. Exploration beyond wild-type neighborhoods often leads to biologically implausible sequences or relies on surrogate models that lose fidelity in novel regions. Here, we propose PROSPERO, an active learning framework in which a frozen pre-trained generative model is guided by a surrogate updated from oracle feedback. By integrating fitness-relevant residue selection with biologically-constrained Sequential Monte Carlo sampling, our approach enables exploration beyond wild-type neighborhoods while preserving biological plausibility. We show that our framework remains effective even when the surrogate is misspecified. PROSPERO consistently outperforms or matches existing methods across diverse protein engineering tasks, retrieving sequences of both high fitness and novelty.
Additional Metrics?
[➜Log in]
Edit extra informations Login
Publication type Article: Conference contribution
ISSN (print) / ISBN 1049-5258
Conference Title 39th Conference on Neural Information Processing Systems, NeurIPS 2025
Conference Date 02-07 December 2026
Conference Location San Diego
Quellenangaben Volume: 38, Issue: , Pages: 145223-145256 Article Number: , Supplement: ,
Institute(s) Institute of AI for Health (AIH)
Helmholtz Artificial Intelligence Cooperation Unit (HAI)