Research map / Probabilistic and scientific modeling / Molecules and proteins
Protein representation learning
Learn protein representations from sequence and 3D structure graphs.
40 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Molecules and proteins, keyword cluster “proteins · binding · sequence”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
14 papers in 2023–24 and 14 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 2.9% to 1.5%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Protein contact prediction from amino acid co-evolution using convolutional networks for graph-valued images (NeurIPS 2016 · 32 citations)
- 2017 · Protein Interface Prediction using Graph Convolutional Networks (NeurIPS 2017 · 557 citations)
- 2019 · Generative Models for Graph-Based Protein Design (NeurIPS 2019 · 209 citations)
- 2021 · Learning from Protein Structure with Geometric Vector Perceptrons (ICLR 2021 · 202 citations)
- 2022 · Iterative Refinement Graph Neural Network for Antibody Sequence-Structure Co-design (ICLR 2022 · 66 citations)
- 2023 · A Graph Neural Network Approach to Automated Model Building in Cryo-EM Maps (ICLR 2023 · 27 citations)
- 2024 · Prot2Text: Multimodal Protein’s Function Generation with GNNs and Transformers (AAAI 2024 · 41 citations)
- 2025 · Boosting Protein Graph Representations through Static-Dynamic Fusion (ICML 2025 · 3 citations)
- 2026 · Learning the Interaction Prior for Protein-Protein Interaction Prediction: A Model-Agnostic Approach (ICML 2026 · 0 citations)
Earliest papers here
- Protein contact prediction from amino acid co-evolution using convolutional networks for graph-valued images (NeurIPS 2016 · 32 citations)
- Protein Interface Prediction using Graph Convolutional Networks (NeurIPS 2017 · 557 citations)
- Generative Models for Graph-Based Protein Design (NeurIPS 2019 · 209 citations)
What has been tried
33 papers here report 343 measured changes (a component added, removed or swapped, compared with a baseline): helped 73%, no clear effect 5%, hurt 20%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Replace per-amino-acid fusion with concatenating the graph representation and sequence representations | SwissProt test set · BLEU Score | 32.11 -> 30.39 | Hurt | Prot2Text: Multimodal Protein’s Function Generation with GNNs and Transformers (AAAI 2024 · 41 citations) |
| replace neural clustering with GRACLUS | EC number prediction · Fmax | 0.866 -> 0.792 | Hurt | Clustering for Protein Representation Learning (CVPR 2024 · 11 citations) |
| add surface message passing to EquiPocket-LG | COACH420 · DCC | 0.212(0.016) -> 0.423(0.014) | Helped | EquiPocket: an E(3)-Equivariant Geometric Graph Neural Network for Ligand Binding Site Prediction (ICML 2024 · 7 citations) |
| Add sequence or structure denoising as an auxiliary task | Downstream tasks across models and primary tasks · task metric | 50 out of 60 combinations | Helped | Evaluating Representation Learning on the Protein Structure Universe (ICLR 2024 · 6 citations) |
| Add structure denoising guidance | GO-CC (TMscore ≥ 0.5) · Fmax | 0.429 -> 0.442 | Helped | Protein 3D Graph Structure Learning for Robust Structure-Based Protein Property Prediction (AAAI 2024 · 5 citations) |
| Align trajectory frames before computing correlation edges | Atomic adaptability · Pearson R | 0.5493 -> 0.6426 | Helped | Boosting Protein Graph Representations through Static-Dynamic Fusion (ICML 2025 · 3 citations) |
| Add batch support, Batch Normalization, and scale-informed diffusion-time initialization | RNA segmentation · test accuracy | 80.9% -> 84.1% | Helped | AtomSurf: Surface Representation for Learning on Protein Structures (ICLR 2025 · 3 citations) |
| Use ORF gene profiles instead of CRISPR knockout profiles | Downstream DTI prediction · performance | Helped | MOTIVE: A Drug-Target Interaction Graph For Inductive Link Prediction (NeurIPS 2024 · 3 citations) | |
| Add graph neural network layers instead of fully connected layers | AsEP · MCC | approximately 39.8 % | Helped | AsEP: Benchmarking Deep Learning Methods for Antibody-specific Epitope Prediction (NeurIPS 2024 · 3 citations) |
| add backbone-level representations beyond amino-acid-level representations | Fold classification — Fold · accuracy | 57.8 -> 60.6 | Helped | Learning Complete Protein Representation by Dynamically Coupling of Sequence and Structure (NeurIPS 2024 · 2 citations) |
| Add neighborhood prediction | Novel function prediction · Top-1 accuracy | 30.76 ± 0.43 -> 34.08 ± 0.76 | Helped | GoBERT: Gene Ontology Graph Informed BERT for Universal Gene Function Prediction (AAAI 2025 · 1 citation) |
| Increase the positive-to-negative training-pair ratio to 1:5 | Human intra-species PPI network construction, BFS · Network reconstruction performance | Helped | PRING: Rethinking Protein-Protein Interaction Prediction from Pairs to Graphs (NeurIPS 2025 · 1 citation) |
What reviewers will expect
Datasets and baselines named in the 9 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- GCN · 2 of 9 papers
- GAT · 2 of 9 papers
- GraphGPS · 2 of 9 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 28 papers; from OpenAlex and the papers' affiliations).
- Zhejiang University · 3 papers
- Tsinghua University · 3 papers
- Renmin University of China · 3 papers
- Westlake University · 2 papers
- Sun Yat-sen University · 2 papers
- University of Cambridge · 2 papers
Use it
Download BibTeX of 33 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Generative protein structure design (142 papers)
- Protein fitness and variant effect prediction (71 papers)
- Protein interface and docking prediction (72 papers)
- Biological sequence transformers (34 papers)
Elsewhere in the topic
- Prediction-powered and conformal inference (129 papers)
- Statistical analysis of ML practice (122 papers)
- Neuro-symbolic probabilistic layers (120 papers)
- Structure-based 3D molecule generation (107 papers)
- Sparse graphical and tensor models (97 papers)
- Flow matching generative modeling (97 papers)
- Machine learning for physical simulation (96 papers)
- Multimodal trajectory prediction (95 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).