NCI60 transcriptomics, proteomics, mutations, cnvs contain genes not in `genes.csv`

[ymahlich]

combining and sorting (i.e. generating the union of) entrez_ids from nci60 transcriptomics, proteomics, mutations & cnvs generates a set of gene ids which is not contained in genes.tsv.

specifically:

• unique entrez_ids in the union of transcriptomics, proteomics, mutations & cnvs: 19,387
• unique entrez_ids in genes.csv.gz: 19,811
• intersection of the two: 19,145

steps to reproduce:

$ zcat < nci60_transcriptomics.csv.gz | cut -d, -f1 | tail -n+2 > nci60_genes.tsv

$ zcat < nci60_proteomics.csv.gz | cut -d, -f1 | tail -n+2 >> nci60_genes.tsv

$ zcat < nci60_mutations.csv.gz | cut -d, -f1 | tail -n+2 >> nci60_genes.tsv

$ zcat < nci60_copy_number.csv.gz | cut -d, -f1 | tail -n+2 >> nci60_genes.tsv

$ zcat < genes.csv.gz | cut -d, -f1 | tail -n+2 | tr -d '"' | sort -u > genes.tsv

$ wc -l genes.tsv
     19811 genes.tsv

$ sort -u nci60_genes.tsv > nci60_genes_sorted.tsv

$ wc -l nci60_genes_sorted.tsv
     19387 nci60_genes_sorted.tsv

$ grep -Fxf nci60_genes_sorted.tsv genes.tsv | wc -l 
     19145

[sgosline]

We need to double check the NCI60 build to see if we are excluding genes when we build the dataset, then decide if we want to exclude.