I'm manually reviewing Insertions and deletions. When counting separate reads: these don't add up to show the number, shown in the coverage-band...
when opening the file multiple times, the amount of reads is different each time (when counting manually). The coverage numbers stay the same.
Someone an idea?
You get different results every time? Just keep counting reads until you get a p-value < 0.05
Hehe, i joke. Chances are your reads are stacking up and disappearing off the bottom of the screen. Try right clicking and selecting 'repack reads' or something to that effect. Worst case scenario, use "samtools tview chr:0-1" in the terminal to see reads more "directly".
Finally, if you used GATK and what you really mean is that the number of reads dont match up to the number of deletions/insertions your seeing in a VCF file, thats because GATK will do local realignment around suspicious areas...
As much as IGV drives me crazy, every single time I thought it made a mistake, it was I (or my data) which was wrong.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
IGV does not always show you all the reads, there is some sort of cutoff there to control the resource usage. Showing all reads on a very high coverage could overload a system.
Hence you can't really count the reads unless you have very few a them (10 or so). For all other uses the coverage tracks will tell you exactly how many reads you have at a given base.
I really don't know what happens... It not that they are not visible: I sort them so, all the insertions are show on top of the row. The coverage band says I must have 34 insertions. But I get lower numbers. When I move the selected base and do a new sort, I get the correct amount. This scares me... Bioinformatics with IGV is not reproducible...
I need to check manually if the amount of insertions and deletions is correct. But I find a lot more deletions and insertions than given by the firm due to misalignment. But I really have to be careful to have a view, where everything is shown!
The sad state of genomic visualization is that most tools have substantial usability and/or other flaws. Unfortunately the state of the science is that it is much easier to get research support for collecting data than to visualize it. Of course this latter is essential for making sense of the former.