Extracting Columns From Vcf File Using Vcftools/Perl

Question

[bioinfo]

What's the easiest way to extract 2 or more columns simultaneously from my vcf SNP file and export them in CSV format from command line in linux? Any commands in vcftools or do I have to use perl?

[dfornika]

You may want to specify if you're trying to extract 2 or more samples or simply arbitrary columns. It looks like matted's answer applies to the former, and Dk and Rm's apply to the latter.

Answer 1

[Damian Kao]

This linux command will output column 1,2, and 3 of your file into csv. Change the 1,2,3 according to the column (1-based) you want to extract. You can extract as many columns as you want.

cut -f 1,2,3 fileName.vcf | sed 's/[\t]/,/g' > cols.csv

[bioinfo]

I tried with the command but not sure why i m getting it wrong..!!

cut -f 2,6,9 file.vcf | sed 's/[\t]/,/g' > out.csv
sed: -e expression #1, char 1: unknown command: `\ufffd'

sed: -e expression #1, char 1: unknown command: `\ufffd'

By the way what else can I do for INFO column (column 9) if I want to get just DP from a list of things in that column 9 (INFO column contains e,g. AN=2;DP=87;Dels=0.00..)

Answer 2

[matted]

For completeness, here's a vcftools solution:

vcf-subset -c name1,name3,name5 in.vcf.gz | tr "\t" "," > out.csv

[bioinfo]

vcf-subset -c POS,QUAL my.vcf.gz | tr "\t" "," > out.csv

Its giving me just the header line of the two columns in the VCF file POS and QUAL

#CHROM,POS,ID,REF,ALT,QUAL,FILTER,INFO,FORMAT,POS,QUAL

by the way I don't want any other info except the data in the columns with header

Answer 3

[Rm]

Using Awk :

awk 'BEGIN {OFS ="," ; FS = "\t"};{print $1, $2, $4}' input.vcf > output.csv

In perl:

perl -F"\t" -lane '$, = ","; print $F[0], $F[1] , $F[3]' input.vcf > output.csv