I am looking for gene sets that are associated with schizophrenia specifically, but also related neurological functioning. What is the best way to do this apart from delving into the literature? As far as I can tell, going through GSEA's mSigDB is one way. Any other recommendations? Thanks in advance!
•
written
by
Ellen O •
Maybe grab SNPs associated with the phenotype (example) and locally query them for their genomic positions, which gives a BED file.
Once you have positions in a BED file, you could use BEDOPS bedmap with a set of BED-formatted gene annotations to map SNPs to genes.
For example, to get gene annotations:
$ wget -qO- ftp://ftp.sanger.ac.uk/pub/gencode/Gencode_human/release_27/gencode.v27.basic.annotation.gff3.gz \
| gunzip --stdout - \
| awk '$3 == "gene"' \
| convert2bed -i gff - \
> genes.bed
Then:
$ bedmap --echo --echo-map-id-uniq --delim '\t' snps.bed genes.bed > answer.bed
Once you have gene IDs, you could do GO analysis or other hypergeometic-based analyses etc. that compare your set of genes to those in background.
Sounds like a good approach, but I would like to add that the nearest gene isn't necessarily the "functionally involved gene" for associated SNPs. In some cases, the functional gene is 100's of kb further.
Log in to answer this question.
