Research map / Probabilistic and scientific modeling / Molecules and proteins
Exploring evolution-aware & -free protein language models as protein function predictors
Mingyang Hu, Fajie Yuan, Kevin K Yang, Fusong Ju, Jin Su, Hui Wang, Fei Yang, Qiuyang Ding
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add AlphaFold pre-training to Evoformer (ablation) | Evoformer · Secondary Structure | score: 0.614 -> 0.785 | Improved |
| add pre-training (ablation) | ESM-1b · Secondary Structure | score: 0.500 -> 0.703 | Improved |
| add pre-training (ablation) | Evoformer · Metal Ion Binding | accuracy: 0.645 -> 0.794 | Improved |
| add pre-training (ablation) | Evoformer · Antibiotic Resistance | accuracy: 0.920 -> 0.979 | Improved |
| add pre-training (ablation) | Evoformer · Fluorescence | Spearman ρ: 0.36 -> 0.67 | Improved |
| add pre-training (ablation) | Evoformer · Stability | Spearman ρ: 0.52 -> 0.79 | Improved |
| add MSA input (ablation) | Evoformer · Secondary Structure | score: 0.716 -> 0.785 | Improved |
| add MSA input (ablation) | Evoformer · Metal Ion Binding | accuracy: 0.724 -> 0.794 | Improved |
| add MSA input (ablation) | MSA-Transformer · Antibiotic Resistance | accuracy: 0.908 -> 0.961 | Improved |
| replace Jackhmmer MSA with ESM-MSA (ablation) | Evoformer · Secondary Structure | score: 0.785 -> 0.776 | Hurt |
| replace Jackhmmer MSA with ESM-MSA (ablation) | MSA-Transformer · Metal Ion Binding | accuracy: 0.715 -> 0.774 | Improved |
| use ESM-MSA instead of HHblits for MSA retrieval | MSA retrieval · MSA retrieval in 12h with 12-core CPU | Number of MSAs: 1509 -> 8800 | Improved |
Closest papers by meaning
- LEMON-ZEST: Evolution-Informed Tokenization for Efficient Protein Language Modeling (NeurIPS 2026 · similarity 0.90)
- DePLM: Denoising Protein Language Models for Property Optimization (NeurIPS 2024 · similarity 0.90)
- From Mechanistic Interpretability to Mechanistic Biology: Training, Evaluating, and Interpreting Sparse Autoencoders on Protein Language Models (ICML 2025 · similarity 0.89)
- SaProt: Protein Language Modeling with Structure-aware Vocabulary (ICLR 2024 · similarity 0.89)
- Training Compute-Optimal Protein Language Models (NeurIPS 2024 · similarity 0.89)
- DPLM-2: A Multimodal Diffusion Protein Language Model (ICLR 2025 · similarity 0.89)
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.
