I have PE300 illuminated reads, which only overlap over 100 bp. However, I still want to use most of the non-overlapping region for my analysis. As I understand it, the command “make.contigs” uses the q-scores for the overlapping region, but what about the non-overlapping region? Can I somehow use the q-scores for the non-overlapping region of the sequence as well? One way I can imagine to do this is to first run the “fastq.info” command, then use the q-scores as suggested in the 454-SOP, and lastly run “make.contig” on the now quality processed reads. For me this seems very tedious and may also produce some problems later I haven’t thought of.
I hope somebody can help solve this problem for me.
You’d be on your own with this one. The data from the non-overlapping region is generally poor and not to be trusted. The problems with the 300PE reads amplify this problem. This is why we encourage people to only sequence regions where they can generate fully overlapping reads (http://blog.mothur.org/2014/09/11/Why-such-a-large-distance-matrix/)
Thanks for the reply. I’m aware of the problems with PE300. That is actually also why GATC-Biotech (the sequencing company) in corporation with Illumina has worked out an in-house method to do PE300 on Hiseq machines. I were a bit skeptical at first, but after following the Miseq SOP on the data with the exception that I used “diff=4” (mean length of sequences =480) in the precluster command and deleting reads that were not found at least 3 times across all the samples. Going from 21.000.000 reads (I actually had 70.000.000 at first, but had to downgrade so our server could handle the process) I now roughly have 19.000 unique reads and a total of 1236 OTUs (at 97% similarity) ranging from 47 to 388 OTUs per sample (human oral samples). I think these numbers are pretty good, but I still want to do better, and thats why I asked about what to do with the non-overlapping regions.