illumina Interval List: 0- or 1-based?

[user 31888]

Hi,

A bit confusing...
I am using a human reference genome (human_g1k_v37.fasta) from GATK, a Ensemble annotation file and the TruSeq Exome Targeted Regions Manifest v1.2 (http://support.illumina.com/downloads.html).


1) To fit with the 2 other files I formatted the 'chr' (removing it), but is the TruSeq .bed file 0- or 1-based? Where does the file come from?
If it is 0-based, I think I have to make it 1-based to fit the annotation file, right? How could I do that?

2) Picard complains about the TruSeq interval list not having a header. Is it ok if I append the .dict file (generated from the reference genome with GATK) to it?
What is the format of the interval list file?

Thanks !

[Jessica_L]

I think I have a partial answer for you, at least:

The bed file numbering is most likely 1-based because the first base of a chromosome in most references is usually numbered as one, not zero. Based on the references you've downloaded, you probably won't need to do anything to the bed file.

As for where the file comes from, it's generated by Illumina-- it represents the targeted chromosome regions based on the probes they designed. In a low tech operation with few targets, bed files can be generated manually. At the larger scale of targeted exomes, manual design of probes is pretty impractical, so I assume that Illumina is using a design software to generate all the probes and that the .bed files are automatically generated by the same software.

I'm not really certain how to answer the second set of questions-- Is Picard giving you an error and stopping or giving you a warning and continuing? There's additional information on interval list formatting here: https://www.broadinstitute.org/gatk/...rticle?id=1204 and look at item number three-- you have a couple options for how that file should be formatted. It sounds like you can put the contents of the .dict file into the list file, but that you will still need the intervals, too.

I hope that helps!