Filtering no-calls from VCF file

Question

[SOHAIL]

Hi Everyone, i want to remove all sites with more than 5% of no-call (./.) among all the samples in my VCF file, could you please guide me how i can do that variant filtration??

Thanks!

Answer 1

[Pierre Lindenbaum]

using vcffilterjs https://github.com/lindenb/jvarkit/wiki/VCFFilterJS

 java -jar dist/vcffilterjs.jar -e 'var n=0.0;for(var i=0;i< variant.getNSamples();i++ ) { if(variant.getGenotype(i).isNoCall()) n++;} n/ variant.getNSamples() < 0.05'  input.vcf

[SOHAIL]

Thanks @Pierre.. That worked perfectly fine for me..

However i am curious and interested to use VCFFILTERJS more, Could you please indicate any guideline for writing java scripts expressions particularly for VCFFILTER -e. I have seen some examples but i wanna know about basic functions like: variant.getNSamples() or variant.getGenotype().isNoCall() , etc.

[Pierre Lindenbaum]

this tools uses a javascript engine (nashorn) that can read the underlying java classes of the htsjdk library like https://samtools.github.io/htsjdk/javadoc/htsjdk/htsjdk/variant/variantcontext/VariantContext.html . It uses a old and well-known java technology called reflection: https://docs.oracle.com/javase/tutorial/reflect/ .