Research map / Graph learning / Graph algorithms and clustering
RepBin: Constraint-Based Graph Representation Learning for Metagenomic Binning
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 RepBin instead of the strongest listed binning baseline | metagenomic contig binning · Sim-5G | F1: 63.22 -> 99.69$\pm$0.10 | Helped |
| Use RepBin instead of the strongest listed binning baseline | metagenomic contig binning · Sim-10G | F1: 59.17 -> 99.37$\pm$0.04 | Helped |
| Use RepBin instead of the strongest listed binning baseline | metagenomic contig binning · Sim-20G | F1: 91.86$\pm$1.25 -> 97.15$\pm$0.61 | Helped |
| Use RepBin instead of the strongest listed binning baseline | metagenomic contig binning · CAMI | F1: 87.11 -> 87.41$\pm$0.60 | Helped |
| Use RepBin instead of the strongest listed binning baseline | metagenomic contig binning · Sharon | F1: 53.50 -> 74.97$\pm$0.16 | Helped |
| Add constraint-based learning to RepBin (ablation) | RepBin · CAMI | F1: 82.10$\%$ -> 87.41$\%$ | Helped |
| Add constraint-based binning instead of using K-Means for final bins (ablation) | RepBin · CAMI | F1: 77.84$\%$ -> 87.41$\%$ | Helped |
| Increase the proportion of incorrect constraints | RepBin · Sim-20G | F1: 97.15 -> 96.94 -> 94.98 -> 93.15 | Hurt |
| Increase the proportion of incorrect constraints | RepBin-Learning · Sim-20G | F1: 91.81 -> 88.68 -> 84.48 -> 83.07 | Hurt |
| Vary diffusion, constraint-weight, and embedding-dimension parameters across the tested ranges (ablation) | RepBin · parameter analysis | performance | No clear effect |
Closest papers by meaning
- Encoding Unitig-level Assembly Graphs with Heterophilous Constraints for Metagenomic Contigs Binning (ICLR 2024 · similarity 0.92)
- Metagenomic Binning using Connectivity-constrained Variational Autoencoders (ICML 2023 · similarity 0.89)
- Revisiting K-mer Profile for Effective and Scalable Genome Representation Learning (NeurIPS 2024 · similarity 0.88)
- Modelling Microbial Communities with Graph Neural Networks (ICML 2024 · similarity 0.85)
- THESAURUS: Contrastive Graph Clustering by Swapping Fused Gromov-Wasserstein Couplings (AAAI 2025 · similarity 0.83)
- Unsupervised Gene-Cell Collective Representation Learning with Optimal Transport (AAAI 2024 · similarity 0.83)
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.