Protein Language Models: Learning From Evolution, Designing Beyond It
- 1Helmholtz Munich
Abstract
Protein language models (PLMs) have transformed our ability to learn from evolutionary sequence space, but protein engineering ultimately asks a different question: not what evolution selected, but what we should build next. Zero-shot likelihoods therefore provide useful, but not universal, measures of fitness and can misalign with engineering objectives. Experimental supervision redirects these priors towards properties of interest, enabling target-specific prediction and closed-loop optimization. PLMs thereby complement structure-based design: structural methods provide geometric control, while PLMs integrate experimental feedback to optimize functional and developability properties. Realizing this potential requires evaluation beyond retrospective predictive accuracy, towards extrapolation, multi-objective optimization, and prospective experimental success. We argue for recurring competitions combining scalable predictive benchmarks with prospective experimental challenges. Ultimately, progress should be measured not by predicting existing experiments, but by enabling successful new ones.
Related
- Learning Constraints in Protein Evolution beyond Known Substitution Frequencies
- Linear Fitness Subspace in Protein Language Models Enables Sample-Efficient Directed Evolution
- SurfGraphPro: Integrating Protein Language Models with Geometric Deep Learning On Coarse Protein Surfaces for Binding Site Prediction
- AmyloCore-ML: AI/Machine Learning Enabled Identification of Amyloid Fibril Core Regions Using Protein Language Models
- EM3DFold: accurate de novo protein and nucleic acid model building for cryo-EM maps using language model-powered deep learning
BibTeX
@misc{brenner2026protein,
title = {{Protein Language Models: Learning From Evolution, Designing Beyond It}},
author = {M. Brenner and J. Schlensok and A. Plaikner and R. Schmirler and M. Heinzinger},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.23.753106},
url = {https://www.biorxiv.org/content/10.64898/2026.09.23.753106v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Protein Language Models: Learning From Evolution, Designing Beyond It
AU - M. Brenner
AU - J. Schlensok
AU - A. Plaikner
AU - R. Schmirler
AU - M. Heinzinger
PY - 2026
DA - 2026/09/23
PB - bioRxiv
DO - 10.64898/2026.09.23.753106
UR - https://www.biorxiv.org/content/10.64898/2026.09.23.753106v1
AB - Protein language models (PLMs) have transformed our ability to learn from evolutionary sequence space, but protein engineering ultimately asks a different question: not what evolution selected, but what we should build next. Zero-shot likelihoods therefore provide useful, but not universal, measures of fitness and can misalign with engineering objectives. Experimental supervision redirects these priors towards properties of interest, enabling target-specific prediction and closed-loop optimization. PLMs thereby complement structure-based design: structural methods provide geometric control, while PLMs integrate experimental feedback to optimize functional and developability properties. Realizing this potential requires evaluation beyond retrospective predictive accuracy, towards extrapolation, multi-objective optimization, and prospective experimental success. We argue for recurring competitions combining scalable predictive benchmarks with prospective experimental challenges. Ultimately, progress should be measured not by predicting existing experiments, but by enabling successful new ones.
ER -