Fastqc report

[melNGS]

Hi,

I'm having a problem with one of my sample!
I analysed it with fastqc and it shown the Quality score between 30 and 40 (I attached the pic), but when i looked the kmer (attacehd pics) it looked awful! I don't understand what it is and why it shows a good quality if my sequences are not good at all!! The analysis is terribile: I'm not able to identify SNP/indel and the coverage is low!! Do you know what it's going on with this sample?? And do you know if there is a way to clean or to restore it??

Thanks!!

M

[melNGS]

The run was performed with Illumina Hi-Seq 2000, with a multiplexed assay: I run twelve samples in one lane. The data from the other samples are fine, the problem is only for this one! I forgot to mention that it's paired end and the pics attached are only for the read number 1, but the read number 2 looks the same. How can I trim it? Do you think in galaxy can I find a tool? DO you think is it contamination or what?

Thanks so much!!

[westerman]

Did you rule out that possibility that your sample is not from the organism that you think it should be from? You ask, "DO you think is it contamination" ... that is the most plausible explanation. Take about 100 reads and blast them against 'nr' to see you you get.

[simonandrews]

This looks like the random hexamer problem which comes up a lot on RNA-Seq data. Basically the random primers used to create RNA-Seq libraries aren't actually random and give you a Kmer bias at the front end of your reads. We see this in most RNA-Seq libraries to a greater or lesser extent. Yours is maybe a little worse than normal but if it maps OK then it's probably alright to keep using it. If it is this then there's no point trying to remove this bias, since the sequence you've got is probably correct but from a slightly biased subset of positions within your data.

More details can be found in this paper.