MiSeq 500 Cycle Kit - Low Read 2 Qualities

[tmm343]

Very specific question but has anyone encountered low read quality in only read 2 on the MiSeq using a 500 cycle v2 kit with run configuration of 2x250?

I have does 3 consecutive runs now with two different labs and all 3 are giving me the same phenotype with drastic read 2 quality drop off. The runs do use custom primers but all have been used by collaborators with success and Tm's have been checked. I have done similar runs with the 300 cycle kit with no issue.

[nucacidhunter]

I wonder if the custom primer in all those runs are from the same batch. If primers quality is low, for example, some are short by one base that would affect phasing and hence quality. There could be other reasons as well.

[tmm343]

The custom primers are different between the runs and from two separate labs. One set is straight from the Caparoso paper for 16S sequencing. The other set has been ordered by the lab but same sequences used by a collaborator and have been successful.

I originally thought it was a primer issue but it doesn't add up since they are not the same.

[GenoMax]

In this case it may be best to work with Illumina tech support who can remote in to look at your original data (or in basespace if you are using that) and help diagnose the problem.

We can ask you to post FastQC plots etc but that may not get us very far towards an answer (you could post them if you have those available).

[nucacidhunter]

I meant the primer synthesis or storage issues. Custom sequencing primers without 3' phosphotioate linkage and gel purification would most likely comprise oligo population of varying length. If shorter and full length oligos hybridise to cluster strands and prime sequencing then phasing could be an issue. This can be rulled out by comparison of phasing/prephasing values with successful runs or with read1 of the same run.

[tmm343]

GenoMax - I have been working with tech support and they believe it is a primer or library issue most likely but aren't 100% sure. I find both a little hard to believe since it is the same issue in two different labs with two different samples and two different primer sets and Read 1 qualities are spot on.

I tried to upload images but seems they are too large. Trying to get it to work to show some reference!

nucacidhunter - The primers for read 1 and read 2 were synthesized the same. I cannot speak for storage conditions since the labs are responsible for this. The phasing from Read 1 and Read 2 are much different. See below for the two runs.

Read 1: 0.0002/0.015
Read 2: 0.836/0.272

Read 1: 0.343/0.307
Read 2: 1.065/0.555

[nucacidhunter]

I assume those figures are from SAV and represent Read 1 and 2 figures not the index read (in SAV index1 indicated as read2 and index 2 if present as read 3 and read 2 as read 3 or 4). Here are some figures from my runs with Illumina primers:
R1 0.070/0.008 corresponding R2 0.041/0.011
R1 0.175/0.043 R2 0.166/0.037
R1 0.146/0.071 R2 0.188/0.104
R1 0.099/0.122 R2 0.029/0.065

I note those figures are higher in your runs R2.

[GW_OK]

Maybe cluster densities compounding low diversity sequences?

I've run some 16S amplicon libraries (similar to your Caporaso) had had beautiful read 1's that absolutely tanked in read 2. This was after the "fix" to allow less PhiX spike-ins, so we didn't have but maybe 2% PhiX spiked in. We ended up clustering to a lower density (~700k/mm2 instead of ~1100k/mm2) and it worked much better.

[GenoMax]

Illumina tech support must have looked at that possibility (and perhaps eliminated that as a cause since tmm343 does not mention it).

@tmm343: Can you post the densities for these runs and reagent version?