Bam QC, GC content

Question

[Martina]

Hello, I would like to ask you for advice with the assessment of BAM quality, especially QC content. This is the result of quality check by qualimap, exome sequencing with NEB and TWIST, only on-target regions are assessed. Have you got some idea why GC content is so shifted? It is a cancer sample, but still, I don't like it. Do you also have some advice about where to find some dataset so I can get some experiences assessing of what is normal and what is not? Thank you very much, Martina

[Carlo Yague]

Have you got some idea why GC content is so shifted?

Here are some suggestions on where to dig to investigate the question: What are you sequencing exactly (species, tissue, ...) ? Do you expect any kind of contamination from different species (virus, bacteria, ...) ? Have you mapped to the transcriptome or genome ? What is your mapping rate ? Do you have any over-represented reads ? Have you tried blasting them ? What does this GC-content plot looks like for the unmapped reads ?

Do you also have some advice about where to find some dataset so I can get some experiences assessing of what is normal and what is not?

Before wasting too much time on other datasets, a simple google search can give you a good idea on what to expect.

[finswimmer]

Hello Martina ,

AFAIK the GC-distribution given in red is the expected distribution of over the whole hg19 reference. You've said you were doing exom sequencing, which is a kind of targeted sequencing. As the several bases are not distributed equally over the genome some kind of difference is expected. Furthermore the library prep can prefer regions with a specific GC content, so that these region have more reads and the end.

If you not have done adapter trimming (which is finde depending on the goal of your analysis) it could also be that you often run into adapters which probably also introduce a bias in the gc content distribution.

fin swimmer