Reindex A Vcf File / How To Handle Cosmic Vcf Files

Question

[secretjess]

So I have managed to successfully query 1000 Genome VCF files before by following tutorials, I thought I could apply this to the COSMIC database. I can't.

I'm trying to make use of some WGS data from COSMIC. As I don't know the file structure I ran the following in an attempt to view the headers:

tabix -H ftp://ngs.sanger.ac.uk/production/cosmic/wgs/CosmicCodingMuts_v64_26032013_noLimit_wgs.vcf.gz
[get_local_version] downloading the index file...
[kftp_connect_file] 550 No such file.
[download_from_remote] fail to open remote file.
[tabix] failed to load the index file.

...so then I downloaded the file:

tabix -H CosmicCodingMuts_v64_26032013_noLimit_wgs.vcf.gz 
[tabix] the index file either does not exist or is older than the vcf file. Please reindex.

How do I reindex a vcf file? I tried this (I don't know what I'm doing at all, sorry):

tabix -f CosmicCodingMuts_v64_26032013_noLimit_wgs.vcf.gz 
[tabix] was bgzip used to compress this file? CosmicCodingMuts_v64_26032013_noLimit_wgs.vcf.gz

....and so in a last ditch attempt I tried just running vcftools (I thought this would filter out any mutations not on X):

vcftools --gzvcf CosmicCodingMuts_v64_26032013_noLimit_wgs.vcf.gz --chr X --out test

VCF index is older than VCF file. Will regenerate.
Building new index file.

Reading Index file.
File contains 641910 entries and 0 individuals.
Filtering by chromosome.
(list of chromosomes)
Skipping Remainder.
Keeping 28481 entries on specified chromosomes.
Applying Required Filters.
After filtering, kept 0 out of 0 Individuals
After filtering, kept 28481 out of a possible 28481 Sites

So didn't work as expected either. I then tried running tabix again but it still gives me the error that an index file doesn't exist.

Are there any read me files or guides that I've just not been able to find? I literally have no idea what I'm doing.

[lh3]

You may have a look at the example in tabix man page. You need to compress with bgzip. The error message has mentioned that. Gzip won't work. Vcftools uses a different index format and it cannot really achieve random access in a gzip'd vcf.

[Sheila]

I am having a similar problem, and I'm trying to subset based on regions in the genome that I've predefined in a bed file.

I have a SNP.vcf file provided by Illumina so as per the instructions above, I bgziped it:

bgzip SNP.vcf

and got a SNP.vcf.gz

Then when I tried this command:

tabix SNPs.vcf.gz -B testbed.bed

I got this error:

[tabix] the index file either does not exist or is older than the vcf file. Please reindex.

Any advice?

[Sean Davis]

You have to create the tabix index first.

[bjlemmer]

I have an extension of this problem. I have a file

chr38.vcf.gz

ran:

tabix chr38.vfc.gz chr38:1-100000

I then ran these lines and get error.

rm chr38.vcf.gz.tbi
tabix -p vcf chr38.vcf.gz
tabix chr38.vfc.gz chr38:1-100000

[ti_index_load] wrong magic number 
[ti_index_load] fail to load the index: chr38.vcf.gz.tbi
[tabix] failed to load the index file

Answer 1

[Sean Davis]

Roughly, you'll need to do these steps (untested):

gunzip CosmicCodingMuts_v64_26032013_noLimit_wgs.vcf.gz
bgzip CosmicCodingMuts_v64_26032013_noLimit_wgs.vcf
tabix -p vcf CosmicCodingMuts_v64_26032013_noLimit_wgs.vcf.gz

[Sheila]

Hi Sean, It seems you know about tabix. Do you mind helping me with the tabix issue I'm having below as well. It's a similar problem...