Mining cosmicDB

cosmic.py

#!/usr/bin/env python

# Mining "COSMIC: Catalogue Of Somatic Mutations In Cancer" database

import os
from subprocess import call
import pandas as pd

# Cosmic DB has requires login to download data, so neither BioMart or wget actually work.
call(["wget", "http://cancer.sanger.ac.uk/files/cosmic/current_release/CosmicCompleteExport.tsv.gz"])
call(["gzip", "-d", "CosmicCompleteExport.tsv.gz"])

# read CosmicDB data
cosmic = pd.read_csv('CosmicCompleteExport.tsv', delimiter="\t")

# read in epigenetic protein info
epi = pd.read_csv('epigenetic_proteins.ids.csv')

# intersect and write csvs
epi = epi[['class', 'Associated Gene Name', 'Ensembl Gene ID']]
merged = pd.merge(epi, cosmic, left_on = "Associated Gene Name", right_on = "Gene name")
merged.to_csv("epigenetic_proteins.cosmicDBMutations.csv", index = False)

tmp = merged.groupby('Gene name')['Primary site']
df = pd.DataFrame([
	tmp.apply(len).index,			# names
	tmp.apply(len),				# n. tumours
	tmp.apply(unique).apply(len),		# n. unique tissues
	tmp.apply(unique).apply("|".join)	# unique tissues
	],
).T
df.columns = ["Associated Gene Name", "Entries Count", "Tissue Count", "Tissues"]

df.to_csv("epigenetic_proteins.cosmicDBMutations.tissues.csv", index = False)