Protein variant selection improves effect prediction with limited assays

TALK: One-Shot Batch Design for Protein Variant Effect Prediction

Computational Engineering, Finance, and ScienceMachine Learning

Summary

Predicting how changes to proteins affect their function is important but testing many variants in the lab can be expensive. The paper presents TALK, a method that helps decide which protein variants to test when there are rules limiting which ones can be measured. TALK picks variants so that learning from their test results will best improve predictions for a broader set of protein variants. This approach helps build better predictive models even with a small number of experiments.

What this means in practice

  • For protein engineers: Select optimal protein variants to assay within experimental constraints to improve prediction models for protein function.
  • For biotechnology researchers: Design experiments that effectively link measurable variants to target variants needing prediction despite protocol restrictions.

Authors

Qiao Huang, Ao Shen, Jie Du, Manning Wang

Abstract

Protein variant effect prediction is central to understanding protein function and guiding protein engineering, yet experimental labels remain costly. In many protein experiments, the variants that can be assayed are dictated by protocol constraints, while the variants that need accurate prediction may lie in a different or broader region of the landscape. This assay-prediction mismatch makes candidate-local criteria such as predicted function, uncertainty, or exploration value insufficient for measurement design. Given a protocol-defined accessible space and a limited budget, we ask which variants should be assayed so that the resulting data best improves landscape prediction. We propose TALK, which estimates each candidate's contribution through posterior coupling between accessible and target variants and reduces redundant information within the batch. Across retrospective ProteinGym and FLIP protocols, TALK generally improves target-space prediction under limited measurement budgets. These results point to a practical two-stage workflow: first, spend a limited assay budget on variants selected to be informative for the target landscape and train a predictive model from the measured labels; then use this model to score and prioritize high-potential variants. Code is available at https://github.com/joehuang618/TALK.