overclustered run?

[vanessamata]

Hi all,

I sent a custom made library pool consisting of 3 amplicons (nextera like adaptors with 7bp indexes) - 80% of the pool - and genomic library of birds (trueseq single index adaptors) - 20% of the pool. We also added 1% phix just for quality control and run this in a miseq nano kit 500 cycles at a final loading concentration of 8pM at our institute. The results were fairly good for a low diversity amplicon library:

cluster density: 870
cluster PF (%): 87.4
%>=Q30: 90.10
reads PF: 2,249,010

We then sent the exact same library at 4nM for a sequencing center to be sequenced in a HiSeq Rapid Run 500 cycles (2 lanes). We asked for the library to be loaded also at 8pM, for the clustering to be done on board, and to do the heat denaturation step before loading in order to better denature the amplicons. We have asked for this in the past and it worked very well, with results more consistent with what is obtained in MiSeq (as that is what we do with MiSeq as well).

The results were a little more disappointing with:

cluster density: 1002 (L1), 1008 (L2)
cluster PF (%): 60.8 (L1), 52.2 (L2)
%>=Q30: 82.6 (L1), 81.6 (L2)
reads PF: 223,630,120 (L1); 192,735,128 (L2)

Looking at the density PF I believe I see some signs of overcluster:



However, how could the library be overclustered at 8pM? It was ok at our MiSeq. Could the guys at the sequencing center messed up in the dilutions? Is there anything I can do for my run to be repeated? Another strange thing is that although the same library was shared among Lane 1 and Lane 2, in lane 1 75.5% of the reads PF were assigned to the amplicons, while in lane 2 only 36.1% of the reads were assigned to amplicons. Lane 2 was also the one with higher signs of overclustering.

Should I include a higher percentage of genomic DNA in the future? My amplicon library consisted in 70% of one amplicon (several hundred different indexes), and ~15% + 15% of two other amplicons (also many indexes).

Thanks

Vanessa