How to clean bad tiles from Illumina reads?

Question

[blazer9191]

Hello all,

I have a few datasets which seem to have some sort of error with the flow cel. There are entire lanes which do not have good data according to my fastQC report. attached is a sample image of one of the per tile quality plots.

How can I fix this? This is single end Illumina RNA-Seq data. 100bp read length. I'm not sure what machine it was done on.

http://i.imgur.com/zLi8O3l.png

Thanks.

Answer 1

[harold.smith.tarheel]

You can filter by tile with the BBMap command of that name (see this thread), although quality filtering should also remove the affected data.

[blazer9191]

Yes, actually I was dumb and didn't look at my cleaned FastQC report until after I posted this. I ran trimmomatic and I ended up with the following: http://i.imgur.com/V3dPYmR.png

It seems pretty good, not sure if I should clean it up more than that.