Tweet
New NextSeq500. A client brought us their own multiplexed library of 30 samples to sequence, which we did using the MID, 2x150 kit. Libraries were NEB Ultra2, indices were BIOO HT 384 12bp, Sequencing kit was Illumina MID 2 x 150.
The run looked good (89% >q=30, 155M clusters PF, 45B bp). Running stand-alone, we demultipexed and got the expected 30 x 4 fastq files and the eight "Undetermined" fastq files. The overall cluster density was slightly high (233K per sq.mm).
The actual read files look odd using fastqc. Read 2 has lots (10's of thousands per sample) of reads which are 35 Ns (correct index), a few thousand PhiX per sample, and a large number (10's of thousands) of reads comprised of long strings of G with a dozen or so As in the middle.
Reads 1 looked more realistic with a few thousand PhiX (bleeds) and a few thousand N-35s.
Could we have set up the plate wrong? Single index means same index on both ends right? The single index BIOO indices were 12bp long, and Illumina told us to reduce the read count to 145 to accommodate the extra index reads. But no one acted like this combo of kits would cause a big problem.
We know about over representation of Gs, and tile-edge artifacts. Advice appreciated as we are new to the NextSeq.
-p
The run looked good (89% >q=30, 155M clusters PF, 45B bp). Running stand-alone, we demultipexed and got the expected 30 x 4 fastq files and the eight "Undetermined" fastq files. The overall cluster density was slightly high (233K per sq.mm).
The actual read files look odd using fastqc. Read 2 has lots (10's of thousands per sample) of reads which are 35 Ns (correct index), a few thousand PhiX per sample, and a large number (10's of thousands) of reads comprised of long strings of G with a dozen or so As in the middle.
Reads 1 looked more realistic with a few thousand PhiX (bleeds) and a few thousand N-35s.
Could we have set up the plate wrong? Single index means same index on both ends right? The single index BIOO indices were 12bp long, and Illumina told us to reduce the read count to 145 to accommodate the extra index reads. But no one acted like this combo of kits would cause a big problem.
We know about over representation of Gs, and tile-edge artifacts. Advice appreciated as we are new to the NextSeq.
-p
Comment