Uncertainty in Biological Networks; Challenges and Opportunities
summary
Proposes representing uncertain interactions with probability-generating polynomials so coefficients give exact degree and joint-degree distributions (no enumeration); uses a two-variable trick for shared edges and shows better likelihood fits across five protein networks (Drosophila, C. elegans, S. cerevisiae, Helicobacter, E. coli). Shows network alignment over ~90 examples where incorporating probabilities improves functional coherence in ~90% of cases, sometimes >10x, and proves reachability is #P-hard, then compresses exponential cases with X and Y collapse variables* and sparsifies by sampling edges to scale.
AI-written summary · about
chapters (17)
- 0:00Intro
- 0:47Probabilistic biological networks
- 2:05From Probabilistic to Deterministic Networks
- 4:13Challenges under uncertainty
- 5:57Degree of a node?
- 7:32Joint degree distribution of nodes
- 10:47Alignment: Topological similarity
- 12:32Support matrix under uncertainty
- 13:40Statistical Coherence 100
- 15:05The reachability problem
- 16:57Preach: Our algorithm: first steps
- 18:10Preach: Computing reachability probability
- 21:03PReach: state of the art
- 21:53PReach with deterministic edges
- 23:40Reachability Probability of GO terms
- 25:08Reachability in cancer types
- 26:48Motif counting