Gene expression script

First version of CosmicDB adapter for retrieval of gene expression data and parallel calculation of average Z-scores. It uses mpi4py for parallel execution and pandas for indexing CSV tables.

Gene argument (-g --gene) can be provided as single name (Uniprot notation) or as path to file containing gene Uniprot IDs, every gene ID in a new line.

Tissue argument is always provided as a single name and all gene expressions are calculated for a single tissue sample.

• requirements: Python 2.7, mpi4py, pandas, requests
• setup COSMICDB_USER and COSMICDB_PASS environment variables
• download token has to be manually obtained from https://cancer.sanger.ac.uk/cosmic/download - it is valid for unknown period of time (48h - 14 days)
• default WORKING_DIR is ./genes/expressions/

Usage:

export COSMICDB_USER=<Cosmic email>

export COSMICDB_PASS=<Cosmic password>

Get expression data from CosmicDB:

python get_gene_expression_parallel.py -g <gene Uniprot ID or file path> -t <tissue name> -n <number of cores(for splitting CSV)>

Calculate average Z-score expression:

mpiexec -n <number of cores> python calculate_gene_expression.py -g <gene name or file path> -t <tissue_name> -o <path-to-output-file>

After execution of script get_gene_expression_parallel.py files generated in ./genes/expressions/ are not removed.

Batch download of expression data from a list in file:

python get_gene_expression_parallel.py -g <path-to--gene-list-file> -t <tissue name> -n <number of cores>

Example of batch download from CosmicDB:

python get_gene_expression_parallel.py -g ./genes/uniprot_entries_small.txt -t breast -n 12

Batch calculation of Z-scores for list of genes:

python batch_calculate_gene_expressions.py -g <path-to-gene-list-file> -t breast -o <path-to-csv-output> -n <number of cores>

Example:

python batch_calculate_gene_expressions.py -g ./genes/uniprot_entries_small.txt -t breast -o ./genes/expression_scores.csv -n 12