SIFT and Polyphen prediction with SnpSift and dbnsfp version 5

Question

[Edmond]

Hi there,

I am having trouble with the annotation of SIFT, Polyphen, and other predictions, with the output of freebayes. I have downloaded dbnsfp version 5 and GRch37.75. The command line I make use of is the following.

java -Xmx8g -jar SnpSift.jar dbnsfp -db db/dbnsfp/dbNSFP4.1a.txt.gz myFile.vcf  > prueba.vcf

java -Xmx8g -jar snpEff.jar GRCh37.75 prueba.vcf >prueba2.vcf

dbNSFP4.1a.txt.gz is indexed with tabix. However, I read on snpsift docsthat for versions higher than 3, one should perform the following conversion:

On the other hand, snpeff version 5 does not have /scripts_build/dbNSFP_sort.pl script.

How can I fix that?

Set to your downloaded dbNSFP version
version="3.2a"

Replace coordinates by columns 7 and 8 (hg19 coordinates) and sort by those coordinates
cat dbNSFP${version}_variant.chr* \
    | $HOME/snpEff/scripts_build/dbNSFP_sort.pl 7 8 \
    > dbNSFP${version}_hg19.txt

# Compress and index
bgzip dbNSFP${version}_hg19.txt
tabix -s 1 -b 2 -e 2 dbNSFP${version}_hg19.txt.gz