Vcf Statistics Given A Gff3 Annotation

Question

[Zev.Kronenberg]

Greetings,

Anyone know of a tool that summarizes where variants fall within a GFF3 file?

an ideal output would be something like:

number of variants within genes

number of variants within utrs

number of variants in CDS

...

Otherwise I can munge some code together. Thanks.

Answer 1

[Cyriac Kandoth]

snpEff is a variant effect annotator, which can build its gene transcript database from a GFF3 file using java -jar snpEff.jar build -gff3. See the section named "Building a database from GFF files" at this page. But make sure your GFF3 isn't already one of the pre-built databases listed by java -jar snpEff.jar databases. Run snpEff on the VCF using java -jar snpEff.jar eff to annotate each variant to all possible transcripts in your GFF3. Since you're not interested in variants flanking gene UTRs/CDS/introns, I recommend using the options -no-downstream -no-upstream. But look over the documentation using java -jar snpEff.jar -h to see all your options.

Once you have an annotated VCF, it should be easier to write a wrapper that counts the variants in UTRs, CDS, introns, etc. If you have variants than map to more than one gene/isoform in your GFF3.