Kmer sequences in 5' RNAseq reads

Question

[silas008]

Should I to trim the first bases of RNAseq reads that contain 7_k-mer?

[GouthamAtla]

First why do you want to trim first 7_k-mers ? Do u think they have very poor quality or they are adapter contaminations ? What your pipeline ? Expression or Assembly or something else ?

[silas008]

I don't know how to put images here to exemplify my question. :(

[h.mon]

upload them to some image hosting site and post the link here.

[venu]

try uploading to google+, and paste the link here.

Answer 1

[h.mon]

No, at least, not blindly. Even though FastQC will tell you the "per base sequence content" is problematic, and in general one can easily spot the first 5-10 bases as being non-randomly distributed, this does not necessarily means they are wrong. If you are really worried, you could map (or assemble and map, if you do not have a reference at hand) and check the alignment, if these bases do not align correctly, then you can trim.