Research map / Point clouds / Point cloud classification
ProteinShake: Building datasets and benchmarks for deep learning on protein structures
Tim Kucera, Carlos Oliver, Dexiong Chen, Karsten Borgwardt
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 |
|---|---|---|---|
| use point cloud instead of graph representation | PointNet++ vs GIN · Protein-Protein Interface | task metric: 0.883 ± 0.050 -> 0.974 ± 0.003 | Improved |
| use voxel instead of graph representation | 3D CNN vs GIN · Structural Similarity | task metric: 0.598 ± 0.018 -> 0.645 ± 0.020 | Improved |
| use voxel instead of graph representation | 3D CNN vs GIN · Ligand Affinity | task metric: 0.670 ± 0.019 -> 0.689 ± 0.013 | Improved |
| use point cloud instead of graph representation | PointNet++ vs GIN · Enzyme Class | task metric: 0.790 ± 0.007 -> 0.712 ± 0.016 | Hurt |
| use voxel instead of graph representation | 3D CNN vs GIN · Structural Class | task metric: 0.495 ± 0.012 -> 0.337 ± 0.011 | Hurt |
| use point cloud instead of graph representation | PointNet++ vs GIN · Binding Site | task metric: 0.721 ± 0.010 -> 0.609 ± 0.006 | Hurt |
| move from random to sequence to structure split | all models · all tasks | task metric | Hurt |
| pre-train on AlphaFoldDB with residue masking (ablation) | graph, point, voxel models · all tasks | relative improvement | Mixed |
Closest papers by meaning
- Evaluating Representation Learning on the Protein Structure Universe (ICLR 2024 · similarity 0.90)
- ProteinBench: A Holistic Evaluation of Protein Foundation Models (ICLR 2025 · similarity 0.89)
- Protein Structure Tokenization: Benchmarking and New Recipe (ICML 2025 · similarity 0.87)
- PSBench: a large-scale benchmark for estimating the accuracy of protein complex structural models (NeurIPS 2025 · similarity 0.87)
- Proteina: Scaling Flow-based Protein Structure Generative Models (ICLR 2025 · similarity 0.86)
- PepSpecBench: A Unified Evaluation Benchmark for Peptide Tandem Mass Spectrometry Prediction (NeurIPS 2026 · similarity 0.86)
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.
