Network Data Integration Enables Precision Medicine
summary
Uses semi supervised non negative matrix tri factorization with graph Laplacian regularizers and multiplicative update rules to fuse networks of genes GO terms patients and drugs and to derive consensus clusters and predictions. Method reconstructs GO with 96% overlap to existing annotations, recovers gene annotations at AUC ≈ 0.87, predicts 132 GO term links of which 58 are validated, predicts 199 putative driver genes with ~40% independently validated, produces three patient strata with significantly different survival curves, and confirms ~37% of novel drug target links.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 2:16Network Data Types
- 3:34Methodology
- 13:11Results
- 14:54Exclude
- 15:11Evaluation
- 24:47Multiple Network Alignment
- 26:38Sequence Alignment
- 28:17Summary
related talks
30mLeveraging Molecular Data for Drug Discovery
33mUnderstanding Molecular Complexity for Precision Medicine
40mGraph neural networks for biomedical insights
32mMulti-Omic Integration for Understanding Disease
53mThe Algorithm for Precision Medicine
01h 19mAnalyzing Genetic Networks Using Neural Networks
17mLeveraging Knowledge Graph and Imaging to Predict Gene–Disease Links in Cardiovascular Diseases
35mFrom cells to drug responses - machine learning in cancer research