AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

[Re] Learning Unknown from Correlations: Graph Neural Network for Inter-novel-protein Interaction Prediction

Urša Zrimšek

NeurIPS 2022 · Probabilistic and scientific modeling › Molecules and proteins · approach: Protein representation learning

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 changedWhereMeasuredEffect
model PPI correlations with a GIN graph over proteins (GNN-PPI vs PIPR)PIPR · STRING BFSmicro-F1: 55.65 -> 78.37Improved
graph-based GNN-PPI vs PIPRPIPR · STRING DFSmicro-F1: 67.45 -> 91.07Improved
graph-based GNN-PPI vs PIPRPIPR · SHS27k BFSmicro-F1: 44.48 -> 63.81Improved
graph-based GNN-PPI vs PIPR on random splitPIPR · SHS148k Randommicro-F1: 90.05 -> 92.26Improved
replace random split with BFS/DFS splitPIPR · STRINGmicro-F1: 94.43 -> 55.65Hurt
replace random split with BFS splitGNN-PPI · STRINGmicro-F1: 95.43 -> 78.37Hurt
evaluate on trainset-homologous test vs unseen STRING test, Random splitPIPR · SHS148kmicro-F1: 92.42 -> 53.85Hurt
evaluate with BFS split instead of Random split as generalization proxyPIPR · SHS148k to STRINGmicro-F1: 66.13 vs. 63.74Improved
build PPI graph from trainset only (GCT) instead of all data (GCA) (ablation)GNN-PPI · STRING BFSmicro-F1: 78.37 -> 73.23Hurt
build PPI graph from trainset only (GCT) instead of all data (GCA) (ablation)GNN-PPI · SHS27k DFSmicro-F1: 74.72 -> 73.42Hurt
remove PGE graph module (PIE only) (ablation)GNN-PPI · STRING BFSmicro-F1: 78.37 -> 50.03Hurt
remove PIE sequence encoder (PGE only, handcrafted features) (ablation)GNN-PPI · STRING BFSmicro-F1: 78.37 -> 73.81Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy