flâneur — a map of the web's best reading

CS294 - Final Project - Google Slides

docs.google.com · saved by 1 readers

CS294 - Final Project Saved to Drive   Slideshow     Share FileEditViewInsertFormatSlideArrangeToolsExtensionsHelp          Arial           1 Final Project 2 Unsupervised Extraction from Perturb-seq So for any CRISPRi knockout, you know the target itself (that RNA). If we have some priors on how genes interact with each other, we can connect all the genes in the counts as X number of hops away from the knocked out gene itself If a gene that is knocked out is many hops away in our priors from a gene in the counts that gets differentially expressed, then by definition, there’s probably a link between those genes and should be reflected in the priors, so we are deconvoluting the data to find “new biology” 5 AARS2 ↔ ATP5C1: physical association (PSI-MI:MI:0915) via BioID; PMID 34079125 (Go 2020); BioGRID:2977523. ATR ↔ MCM2: direct interaction (MI:0407) via enzymatic study; PMID 15210935 (Cortez 2004); BioGRID:251212. (There’s also a Xenopus entry; human shown here.) AURKB ↔ XRCC5: physica

Explore this link on the map →

saved by