biolab/red

Inference of Epistatic Gene Networks

Red

Epistasis analysis is a major tool from classical genetics to infer the order of function of genes in a common pathway. Commonly, it considers single and double mutant phenotypes and for a pair of genes observes if change in one masks the effects of the other one. Despite recent emergence of biotechnology techniques that can provide gene interaction data on a large, possibly genomic scale, very few methods are available for quantitative epistasis analysis.

Red is a conceptually new probabilistic approach to gene network inference from quantitative interaction data. The advantage of Red is the global treatment of the phenotype data through a factorized model and probabilistic scoring of pairwise gene relationships from latent gene presentations. Resulting gene network is assembled from scored relationships.

This repository contains supplementary material for Gene network inference by probabilistic scoring of relationships from a factorized model of interactions, presented at ISMB 2014 by Marinka Zitnik and Blaz Zupan.

Usage

Inferring preferential order-of-action factorized model with default parameters:

>>> from red import Red
>>> from data import loader
>>> G, S, H, genes = loader.load_jonikas_data("data/080930a_DM_data.mat")
>>> gene_red = Red(G, S, H, genes)
>>> gene_red.order()

For more examples see examples.py or run:

$ python examples.py

Input:

• matrix of double mutant phenotypes, G,
• matrix of expected no-interaction double mutant phenotypes, H,
• a vector of single mutant phenotypes, S,
• latent dimension, rank,
• regularization of gene latent representation, lambda_u, lambda_v,
• learning rate of gene latent profiles, alpha,
• learning rate of logistic map, beta.

Inferred factorized model gene_red includes:

• preferential order-of-action scores,
• completed matrix of double mutant phenotypes,
• gene-dependent logistic function model,
• inferred gene network for a given gene set of interest,
• gene latent representation,
• quality (Fro. error and NRMSE) of matrix completion.

Citing

@article{Zitnik2014,
  title     = {Gene network inference by probabilistic scoring of relationships from a factorized model of interactions},
  author    = {{\v{Z}}itnik, Marinka and Zupan, Bla{\v{z}}},
  journal   = {Bioinformatics},
  volume    = {30},
  number    = {12},
  pages     = {i246--i254},
  year      = {2014},
  publisher = {Oxford University Press}
}