Summary
Structure-conditioned inverse folding selects more evolutionarily conserved design positions than sequence-only protein language models (1). When selecting positions for site-saturation mutagenesis of Rubisco, [[ESM-IF|ESM-IF1[[ recommendations were more conserved than those from ESM-1b.
See also
- Inverse-folding-guided site-saturation libraries outperform random mutagenesis
- ESM-IF, but not other inverse folding models, has learned some evolutionary constraints from sequence databases
- Zero-shot performance of PLMs, but not inverse folding models, correlates with number of homologs available for training
- Structure-based methods outperform sequence-based methods on protein stability prediction of point mutants, but not full sequences
1.
McDonald JL, Lin J, Zhao Y, Hie BL, Birch R, Gehring M, et al. Machine learning-assisted directed evolution of plant Rubisco. 2026. Available from: https://doi.org/10.64898/2026.07.25.740226