Snpsift Regular Expression Assist

Question

[rob234king]

I've successfully annotated a VCF file using snpEff but would like to filter the resulting VCF file using snpSift.

The VCF file is in the following format:

#CHROM    POS    ID    REF    ALT    QUAL    FILTER    INFO                            FORMAT                                                    Sample1                                        Sample2                                            Sample3
A       8725    .    C    T    .        PASS    ADP=99;WT=2;HET=1;HOM=0;NC=0    GT:GQ:SDP:DP:RD:AD:FREQ:PVAL:RBQ:ABQ:RDF:RDR:ADF:ADR    0/0:145:77:77:77:0:0%:1E0:37:0:41:36:0:0    0/1:0:58:58:55:3:5.17%:9.8E-1:36:33:25:30:3:0    0/0:285:163:162:161:1:0.62%:9.8E-1:36:39:78:83:1:0

I can filter the main columns e.g. INFO column but I don't know how to filter using the "Sample1 Sample2 Sample3" columns by, for instance, only taking vcf entries where one sample has a minimum FREQ of 5% and the other two samples less than 2%.

So far I have this but would like to add the other expressions I detailed above, any ideas how I can do this?

cat A.vcf | java -jar SnpSift.jar filter " ((NC = 0) & ( REF = 'C' ) & ( ALT = 'T')) " > A_filtered.vcf

Answer 1

[Pavel Senin]

Will that work for something like this? (editing brackets, can't really test it)

cat input.vcf | java -jar SnpSift.jar filter \
 "((NC = 0) & ( REF = 'C' ) & ( ALT = 'T')) & ( \
  ( (FREQ[1]>5) & ((FREQ[2]<2) & (FREQ[3]<2 )) ) | \
  ( (FREQ[2]>5) & ((FREQ[1]<2) & (FREQ[3]<2 )) ) | \ 
  ( (FREQ[3]>5) & ((FREQ[1]<2) & (FREQ[2]<2 )) ) )" > output.vcf

[rob234king]

Thanks but doesn't work, doesn't recognise it as a field or Sample1. I tried adding the Sample1 to the vcf header so I could at least search for a string but still does not recognise it. I thought the program would look at the header to determine what fields are available. May have to just stick with a perl script but was hoping I could do this with existing tools and make it simplier for others.

[Pavel Senin]

What if instead of FREQ[1] you will try GEN[0].FREQ>5 and so on? For me it works when I use it for DP in INFO field of my file, I do not have a FREQ field unfortunately. I tried with your VCF example, but it throws an exception that Genotype numer '0' does not exists.