Resource Summary Report  New Search Previous Search Results

URL: https://dx.doi.org/DOI:10.17504/protocols.io.bpdvmi66

Authors: Lillian Thistlethwaite, xiqi.li2 , Varduhipetrosyan

Group: Metabolomics Protocols & Workflows

Summary: We consider the following algorithmic problem that arises in transcriptomics, metabolomics and other fields: given a weighted graph and a subset of its nodes, find subsets of perturbed variables that show significant connectedness within the graph. While solutions to this problem may be discovered by devising a variety of scoring functions, statistically rigorous discovery is obstructed by the high computational cost of permutation testing, required to establish p-values. In this work, we develop CTD, a novel information-theoretic method that evaluates the connectedness of perturbation signatures in a network context and assigns an upper bounds on their p-values without use of permutation testing. We apply CTD to interpret multi-metabolite perturbations due to inborn errors of metabolism and multi-gene perturbations associated with breast cancer in the context of disease-specific Gaussian Markov Random Field networks learned directly from respective molecular profiling data.

Associated Publications: Thistlethwaite LR, Petrosyan V, Li X, Miller MJ, Elsea SH, Milosavljevic A (2021) CTD: An information-theoretic algorithm to interpret sets of metabolomic and transcriptomic perturbations in the context of graphical models. PLoS Comput Biol 17(1): e1008550. doi: 10.1371/journal.pcbi.1008550

Affiliations: Baylor College of Medicine, Baylor College of Medicine, Baylor College of Medicine

External URL: https://doi.org/10.1371/journal.pcbi.1008550

Version: 4

Publication Date: 2020

Expand All Collapse