Research map / Probabilistic and scientific modeling / Molecules and proteins
ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use autoregressive protein language models instead of masked language models | zero-shot fitness prediction · ProteinGym zero-shot substitution DMS | avg Spearman | Helped |
| combine alignment-based and language-model scoring (TranceptEVE vs Tranception L) | Tranception L · ProteinGym zero-shot substitution DMS | avg Spearman: 0.434 -> 0.456 | Helped |
| use ProteinNPT axial-attention model instead of embedding-based baselines | supervised fitness prediction · ProteinGym supervised substitution DMS | avg Spearman: 0.571 -> 0.613 | Helped |
| replace one-hot features with PLM embeddings (ESM-1v) | supervised baseline · ProteinGym supervised substitution DMS | avg Spearman: 0.224 -> 0.542 | Helped |
| choose model by assay type in zero-shot indels | indel models · ProteinGym zero-shot indel DMS | Spearman by assay type | Mixed |
Closest papers by meaning
- MutaPLM: Protein Language Modeling for Mutation Explanation and Engineering (NeurIPS 2024 · similarity 0.89)
- Augmenting Biological Fitness Prediction Benchmarks with Landscapes Features from GraphFLA (NeurIPS 2025 · similarity 0.89)
- ProteinBench: A Holistic Evaluation of Protein Foundation Models (ICLR 2025 · similarity 0.88)
- Improving protein optimization with smoothed fitness landscapes (ICLR 2024 · similarity 0.88)
- Steering Generative Models with Experimental Data for Protein Fitness Optimization (NeurIPS 2025 · similarity 0.88)
- DePLM: Denoising Protein Language Models for Property Optimization (NeurIPS 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.