read depth analysis on WGS data stored as .sra from ncbi

[Oscar.K]

greeting all,

We are currently interested in preforming read depth analysis in order to investigate a possible CNV in a loci on human chr 10.
The data we want to analyze is from WGS of CHM1_1.1 (Hiseq 2000, paired reads), one issue is that the raw data consists of approximately 400gb stored as 35 different .sra files.
what is the procedure we need to follow in order to get to the point where we can preform the analysis?
from what I understand, we need to download the entire repository (since we don't know where the reads covering the region is, or is it possible to find in which of the .sra the reads we're interested in are located and only download those?) and then use the SRA-toolkit to convert the reads to fastq before we align it to a reference.
And then we can use a software of our choice to analyze the aligned data.
Am I correct in my thinking so far, or is there any unnecessary or plain wrong steps somewhere?

very grateful for any advice.
sincerely
//Oscar

[dpryan]

The reads are likely randomly distributed throughout all of the sra files, so you'll have to download all of them. Yes, you'll have to use the sra toolkit (specifically fastq-dump) to then convert to fastq and an aligner of your choice for the alignments (I hope you have access to a cluster). Once you have the sorted BAMs you can proceed with whatever method you prefer to look at CNVs.

Note, you might download just one of the SRA files and see if it contains alignments as well as reads. It's not very common, but you might get lucky.