You can try using the 'cut' unix command:

cut yourFile.vcf -f 12,13

That would return column 12 and 13. 'f' parameter specifies columns you want to return, 'd' specifies delimiter to cut.

I recently wrote a brief entry about using unix commands to work with tab delimited files: http://blog.nextgenetics.net/?e=28

If you want to use a script solution, you can try this in python:

import sys
for line in open(sys.argv[1],'r'):
  columns = line.strip().split('\t')
  print columns[1], columns[2]

Save as yourName.py and use by: python yourName.py yourFile.vcf

PyVCF (code, documentation) has already done all the hard work of figuring out how to parse a *.vcf file, and presents it to you in a way similar to the csv module in the Python stdlib. The code below should allow you to get started:

import vcf

vcf_reader = vcf.Reader(open(r'path\to\zebrafish.vcf', 'rb'))
for record in vcf_reader:
    for call in record.samples:
        if call.sample=='TUB' or call.sample=='TA':
            if call.gt_nums != '0/0':
                print call.gt_nums

I just get all lines that don't contain #

sed -n '/#/!p' g1.vcf

and then pipe into the cut command:

sed -n '/#/!p' g1.vcf | cut -f 1,8,11

hope that helps.

A command-line one liner might be cooler, but this should work and it's not mysterious. Replace idx_first and idx_second with the index of your genotypes of interest, remembering that python arrays are 0-based.

python
f = open('foo.vcf')
fo = open('foo_stripped.vcf', 'w')
idx_first = 11
idx_first = 12
for line in f:
    if line[0]=='#':
        fo.write(line)
        continue
    a = line.rstrip('\r\n').split('\t')
    out = a[0:9]
    if not "0/0" in a[idx_first] and not "0/0" in a[idx_second]:
        out.append( a[idx_first] )  # first column with genotype
        out.append( a[idx_second] ) # second column with genotype
        fo.write( '\t'.join(out) + '\n')

fo.close()
f.close()