Can you read VCF.idx?

[kgulukota]

Is there a way to use VFC.idx file (other than for viewing)?

I have a large (42 Gig) VCF and associated VCF.idx - both generated by GATK. I have some simple questions about this VCF like how many variants are in it? How many of them are on CHR1, 2, ... etc? Is there a particular variant (KRAS G13D) in this VCF?

Is there some way to answer these questions without having to open it in IGV? Clearly this data exists within the idx file. Right?

[dpryan]

I don't know why you think this information would exist in the index file. The index is just used to be able to calculate where in a file to move to extract some desired data. If you want to get various metrics about a VCF file, then just write a program to calculate it (unless one already exists that does exactly what you want).

[kgulukota]

Right. That's what I was afraid of. And I did write some scripts for things I need.

But, when the file is so large everything takes time and I was hoping that some info in an index file can be used to speed it up.

The question remains though: What is the structure of the idx file? What information does it have?

[dpryan]

You might still be able to use the index if you have a question that needs random file access. If, for example, you wanted to know about a particular stretch of the genome, you could use the index to randomly extract just that part of the VCF file. I believe that the index is a tabix index, so you can have a look at tabix for a C interface, if that's what you want. Alternatively, you could probably use tabix in a script from the command line, though I've never personally tried that.