View on GitHub

iPaD

Efficient drug-pathway association analysis via integrative penalized matrix decomposition

Download this project as a .zip file Download this project as a tar.gz file

iPaD

Authors:

Cong Li, cong.li@yale.edu Can Yang, eeyang@hkbu.edu.hk Greg Hather, ghather@gmail.com Ray Liu, ray.liu@takeda.com Hongyu Zhao, hongyu.zhao@yale.edu

'iPaD' is a package written in Matlab and the name stands for 'integraive Penalized matrix Decomposition (for drug-pathway association analysis)'. It performs drug-pathay association analysis on paired transcription/drug sensitivity profile data using an efficient bi-convex optimization algorithm. For instructions on how to use this package, please refer to 'CCEL_analysis.m' and 'NCI60_analysis.m' as vignettes.

It requires the following four matrices as input: 1) An N by G transcription profile matrix Y1; 2) An N by D drug sensitivity matrix Y2; 3) A P by G pathway-gene relationship indicator matrix L1; 4) A P by D drug-pathway association indicator matrix L2_prior. N is the number of cell lines (samples), G is the number of genes, D is the number of drugs and P is the number of pathways. Of course, the samples, genes, drugs and pathways have to be in the same order across these matrices.

This repository contains the following files:

iPaD.m This is the main function of the iPaD package. iPaD_cv.m Perform cross-validation to choose an appropriate penalty parameter. iPaD_permu.m Perform permutaton test to calculate the p-values for the drug-pathway associations.

(initialize_X.m LassoSolver.m update_B1.m update_B2.m update_X.m) These files are all sub-routines of the iPaD main function.

quality_control.m Perform the two following quality control steps for the input data set: 1) remove genes or drugs that have less than three unique values; 2) merge pathways that have identical member genes.

CCLE_analysis.m Code for analyzing an example real data set - the Cancer Cell Line Encyclopedia (CCLE) data set. Can be used as a vignette for the usage of the iPaD package.

(CCLE_L1.txt CCLE_L2_prior.txt CCLE_L2_validate.txt CCLE_Y1.txt CCLE_Y2.txt CCLE_genes.txt CCLE_drugs.txt CCLE_pathways.txt) These files are the CCEL data set.

NCI60_analysis.m Code for analyzing another example real data set - the NCI-60 data set. Also can be used as a vignette for the usage of the iPaD package.

(NCI60_L1.txt NCI60_L2_prior.txt NCI60_L2_validate.txt NCI60_Y1.txt NCI60_Y2.txt NCI60_genes.txt NCI60_drugs.txt NCI60_pathways.txt) These files are the NCI-60 data set.