Hi, let expect reads with forward/reverse sequence policy. If I find paired read with forward/forward policy, there could be inversion in right end, but also inversion in left end and rearrangement, etc. I could find a lot of similar examples. Can anyone help me how to find out right variations?
Thanks!
Your question, as written, is a bit difficult to parse, so I'll first write what I think you're asking:
"I have paired-end reads and am interested in finding structural variations, particularly inversions. Given that one would normally expect my reads to align to the reference in a forward/reverse (concordant) orientation pair, are there any tools available that can process my aligned data and use the forward/forward (or reverse/reverse, also termed discordant) orientation pairs to call inversions?"
A quick google would have turned up this review, which has a table listing a programs that would seem to be able to do what I assume you're asking. You can find more by searching pubmed for various combinations of "structural variation", "detection", and "NGS".
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Thanks, but I want to know how exactly decide between these variations, because I am implementing my own tool.
Ah, so you're trying to write a tool to make these structural variation calls, but don't know how to make the calls. In that case, read the source code and accompanying papers for some of the published tools (such as those mentioned in the review I linked to above). Usually one would think of a way to make the calls before deciding to write an implementation. Afterall, there might already be packages out there that works in whatever way you're thinking of doing it...in which case there's little point in writing yet another one.
I know how to make these calls, but I don't know if there is a way to decide between variations like I mentioned in my question. I have stored mapped reads in BAM file. To make variation calls I am using appropriate flags. But only with these flags I don't see any way to decide between these variations, which is right. Also I could have reads witch are aligned in forward/reverse orientation, which looks OK, but there could be rearrangement and both reads are inverted.
If you don't know how to decide between the various possible variations, then, by definition, you don't know how to make the structural variation calls. Obviously one can't make such determinations with single reads (in fact, with single reads you can't reliably determine if a discordant mapping is even interesting), so you'll have to take multiple reads surrounding a putative variation into account. With that, you could then properly call the variation (as well as use various filtering criteria to decide if the variations are likely to be real). As I said above, you should see how other programs do it if you can't picture the method yourself.
SVdetect manual was helpful. So I am doing it right, but I just wasn't sure about that. Anyway, thanks for your help.