Interaction-based Methods for Uncovering Genes Functionally Important in Cancers
summary
Pipeline maps tumor missense variants onto homologous structures requiring 60% sequence identity and models about 22% of human proteins, finding binding residues are depleted in healthy polymorphisms but enriched in cancer missense mutations (example: Dicer and magnesium‑binding acidic residues mutated in endometrial and glioblastoma). A second method greedily grows connected subnetworks to trade off coverage versus size using an alpha parameter (chosen ≈0.5), validated by holdouts on 416 kidney cancer patients using an 8,000‑protein network and improving recovery of cancer census genes.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:23Protein networks
- 1:18Two complementary problems
- 3:32Cancer Genomics
- 5:05Mutations and Protein Interactions
- 6:13Structural Approach
- 16:11Summary
- 17:15Joint Work
- 26:25TCGA Data
- 27:01Picking Alpha
related talks
30mUnlocking the Genome: Exploring the Potential of CRISPR/Cas9
17mExploring the Source Code of the Pfizer/BioNTech Vaccine
58mStrong Generalization from Small Brains and No Training Data
01h 01mThe Gospel According to TCS
26mTed Willke, Senior Principal Engineer, Intel Labs
17mSimulation of Transient Effects in Accelerator Magnets
01h 02mFrom here to there and back again
01h 10mDistribution Estimation and Applications I