Index Read not mapped

[koiologist]

Hello,

We are recently experiencing demultiplexing issues on HiSeq 4000, NextSeq 500 and MiSeq, where we have good cluster formations, high read PFs, and greater than 90% >Q30 scores but only a very small fraction <1% of the sequenced indexes passed PF and were mapped. Majority of our read output (>60%) were in the undetermined fastq file but each sample in our pooled library do get a fastq file containing very little reads. We spiked in about 5% PhiX as it is a low diversity library (predetermined viral peptide sequences).

We are using custom 7-nucleotide IS7 indexes which can be mapped (98%) if only one sample (one index) was ran. Library was gel and pcr purified and bioanalyzer showed a single sharp peak at the expected fragment size.

Any thoughts?

[GenoMax]

Over-clustered FC's may have good data for main reads but will fail on the index reads (do you have a large number of N's or are the sequences something you do not expect to get in the index).

You should deliberately under load by 15-20% and/or increase the % of phiX to 10-20% and see if that helps.

[koiologist]

Thanks for the reply. We had one run on the MiSeq where only a single indexed sample (input library) was sequenced. The run was very over clustered and with very low main read PFs (less than half of read output) but most of the indexes (98%) were mapped. I think in one of our NEXTSeq run we did see a bunch of N's in our index sequences. I do not know if the same things (the N's) are also seen in our MiSeq and HiSeq runs.

Thanks.