Hi all,

i am wondering what are possible causes of low Q30 scores, apart from overclustering?

I am working with amplicon libraries which were sequenced on MiSeq using v2 500 PE kit. I sequenced 12 libraries, all prepared simultaneously using the same method. PhiX comprised 21% of the run. The run metrics looked good with avg Q30 of 88%, and the output was very high with nearly 43 million PE reads (expected 25-30 mil). Cluster density was high, about 1,200 K/mm2 which is the upper end for this kit, but ~94% reads passed filter.
However when I ran FastQC on my samples I got very poor per base quality scores for some libraries (but not all!). For some libraries the scores were worse for Read 1, and for others for Read 2. This resulted in a lot of data being discarded.
It basically looks like Q30 data from BaseSpace (where median Q30 score is low only for late cycles for each read) does not match with quality scores given by FastQC.

I would be grateful for any insight as to why this would happen!

Many thanks,