README meta_analysis_code.R (makashir2014)

The meta_analysis_code.R program takes m gene expression datasets from m different studies as an input and performs meta-analysis of differential gene co-expression on pairwise combinations of genes.

All the input data should be located in one single folder; the path of this folder should be specified as data.loc in the code. Similarly, the output path for results should be specified as res.loc in the code.

Datasets must be in tab separated matrix format, each row containing expression of that gene and each column indicating that sample.

There must be the only two unique column names in a dataset (each referring to one condition e.g. cancer, healthy) and these should be the same across all datasets. The first column must contain gene names.

There must be at least 4 samples in each study.

The following output files will be produced:

1. gene list

   contains a list of all genes used for meta-analysis

2. gene-pairs

   contains the list of all gene pairs on which differential co-expression analysis is performed

3. counts

   matrix indicating the number of studies / datasets involved in differential co-expression meta-analysis of that gene pairs

4. q_scores

   matrix indicating the q scores (meta-analysis standard normal scores) for that gene pair

5. p_values

   matrix indicating the p values of differential co-expression meta-analysis for that gene pair

Note that this program does not adjust the p-values for multiple hypothesis testing, this needs to be done separately.