Issue with fetching whole chromosome

[ckeown]

The nugget of code below works beautifully:

import vcf
vcf_reader = vcf.Reader(open('../SNPs/snps_spretus_chr19.recode.vcf.gz', 'r'))
snps_vcf = vcf_reader.fetch('19', 0, 200000000)                                                                                                                                                         snps = {}
alt_snps = {}
for snp in snps_vcf:
    snps[int(snp.POS)] = snp.REF.upper()
    alt_snps[int(snp.POS)] = [str(alt_snp).upper() for alt_snp in snp.ALT]

I want to change it, however, so that I can read the whole chromosome in without having to specify a range on the positions (because ultimately I want to run this for all chromosomes). I have tried snps_vcf = vcf_reader.fetch('19', 0). However, I get an error:

TypeError: 'NoneType' object is not iterable

I have similarly tried leaving the starting position out as well, but that gives me this error:

TypeError: fetch() takes at least 3 arguments (2 given)

Any suggestions? The documentation states that I should be able to run both of the commands above. I would include the version number, but I can't seem to find it because vcf.__version__ doesn't exist. It was installed within the last week, however, using pip, so it's probably the most recent.

Thanks in advance for your help! chris

[martijnvermaat]

I'm guessing fetch is not the problem here, but snp.ALT is None.

[martijnvermaat]

By the way, if your VCF file is grouped by chromosome, you can also get records one chromosome at a time using itertools.groupby instead of fetch. This would be a bit more lightweight since fetch only works on tabix-indexed files and requires pysam.

For example:

import itertools
import operator
import vcf

reader = vcf.reader(filename='bla.vcf')

for chrom, records in itertools.groupby(reader, operator.attrgetter('CHROM')):
    print 'Chromosome:', chrom
    for record in records:
        print record

[ckeown]

Thanks for the response, Martijn. I'm pretty sure it's not snp.ALT because the below code gives me the same error:

import vcf
vcf_reader = vcf.Reader(open('../SNPs/snps_spretus_chr19.recode.vcf.gz', 'r'))
snps_vcf = vcf_reader.fetch('19', 0)
for snp in snps_vcf:
     print "Hello."

I will try the other method, but it's not ideal, because I can't guarantee the file is sorted. Thanks!

[martijnvermaat]

Could you paste the complete stack trace for the error?

[shashidhar22]

Hi I am trying to read the cosmic mutations vcf file one chromosome at a time, and I am having the same issue as stated above, is there any solution to this? Below is my code:

cosmic_vcf = vcf.Reader(open('CosmicMutantExport.vcf.gz','r'))
for lines in cosmic_vcf.fetch(chrom):
    print(records.CHROM,records.POS, records.INFO['GENE'],records.REF,records.ALT)

where chrom is an iterator over the chromosomes.

[martijnvermaat]

The first argument to fetch should be a string, not an iterator.

Could you paste some minimal but complete code that fails? Perhaps as a gist with a small VCF file included?

What versions of Python/PyVCF/pysam are you using?

[shashidhar22]

I am using python 2.7.6, pysam version 0.8.0 and pyvcf version 0.6.7

I tried the following pieces of code import pysam import vcf cosmic_vcf = vcf.Reader(open('CosmicMutantExport.vcf.gz','r')) for lines in cosmic_vcf.fetch('1'): print(records.CHROM,records.POS, records.INFO['GENE'],records.REF,records.ALT)

and got the following error

Traceback (most recent call last): File "test", line 4, in for lines in cosmic_vcf.fetch('1'): TypeError: fetch() takes at least 3 arguments (2 given)

Here is a sample from my vcf file

fileformat=VCFv4.1

source=COSMICv70

reference=GRCh37

fileDate=20140805

comment="Missing nucleotide details indicate ambiguity during curation process"

comment="URL stub for COSM ID field (use numeric portion of ID)='http://cancer.sanger.ac.uk/cosmic/mutation/overview?id='"

comment="REF and ALT sequences are both forward strand

INFO=

INFO=

INFO=

INFO=

INFO=

CHROM POS ID REF ALT QUAL FILTER INFO

1 69345 COSM911918 C A . . GENE=OR4F5;STRAND=+;CDS=c.255C>A;AA=p.I85I;CNT=1 1 69523 COSM426644 G T . . GENE=OR4F5;STRAND=+;CDS=c.433G>T;AA=p.G145C;CNT=1 1 69538 COSM75742 G A . . GENE=OR4F5;STRAND=+;CDS=c.448G>A;AA=p.V150M;CNT=1 1 69539 COSM1343690 T C . . GENE=OR4F5;STRAND=+;CDS=c.449T>C;AA=p.V150A;CNT=1 1 69540 COSM1560546 G T . . GENE=OR4F5;STRAND=+;CDS=c.450G>T;AA=p.V150V;CNT=1 1 69569 COSM1599955 T C . . GENE=OR4F5;STRAND=+;CDS=c.479T>C;AA=p.L160P;CNT=2 1 69591 COSM3419425 C T . . GENE=OR4F5;STRAND=+;CDS=c.501C>T;AA=p.V167V;CNT=1 1 861390 COSM460103 G C . . GENE=SAMD11;STRAND=+;CDS=c.69G>C;AA=p.P23P;CNT=1 1 865609 COSM336143 C T . . GENE=SAMD11;STRAND=+;CDS=c.147C>T;AA=p.P49P;CNT=1 1 865617 COSM3790304 C G . . GENE=SAMD11;STRAND=+;CDS=c.155C>G;AA=p.S52C;CNT=1

[martijnvermaat]

Ah indeed, omitting the start position is only possible with the current master branch, not with the released 0.6.7. (The commit is 4fba62c.)

You can try the current master branch with:

pip install -e git+https://github.com/jamescasbon/PyVCF.git#egg=PyVCF

As an alternative, you could use the code I suggested in an earlier comment.

[martijnvermaat]

I think this should also answer @ckeown's question (omitting the end position is possible in 0.6.7, but only returns one call at that exact position, or None if there is none).

[shashidhar22]

Thank you that resolved the issue