Miseq read 2 bad quality

[MisUser]

Hi all,

We did a Miseq run (read 1-48 cycles, read 2-50 cycles) on a custom made library. The read quality for read 1 is OK but the read 2 quality is very low. Since it is an amplicon library, the client added some stuffer nucleotides for color balancing.
I attached some SAV screen captures since I don't have the FastQC data yet. Should we suspect an issue with the library construction or could it be faulty reagents?
Thank you very much for your help.

[GenoMax]

I doubt it would be reagent related. Can you post cluster density and PF?

In any case having Illumina tech support do a remote login to look at the run should rule out any hardware/software/reagent related issues in advance.

[MisUser]

Thank you for your quick reply,

We had 1200k/mm2 but only 30% of them PF. After talking to the client, it may be possible that there is no A in the first 5 cycles of the read 2 but the library is well color balanced though. Could it be the reason why the quality drops dramatically during the 2nd read squencing?

[nucacidhunter]

One reason could be high cluster density. Actual density could have been higher than reported 1200 which is at high end. From SAV analysis Tab, %Base under Data By Cycle pane would show if the colour balance is an issue. Looking at density tab under flowcell Chart would indicate overclustering if cluster density does not decrease from bottom to flowcell top. Also looking at images will indicate overclustering issue.

[nucacidhunter]

Your library description also points to a custom prep. If library was not prepared by ligating commercial adapters then some adapters may not have full complement of sequencing primers and this also will contribute to low %PF and low Q scores.

[MisUser]

Your library description also points to a custom prep. If library was not prepared by ligating commercial adapters then some adapters may not have full complement of sequencing primers and this also will contribute to low %PF and low Q scores.

Hi nucacidhunter,

Thank you for your input. The client did not use a ligation process to prepare his libraries (instead he built them by PCR) but he used illumina's adapter sequences. We quantified them with qPCR (commercial kit with Illumina's primers) as usual. Since we got signal, I guess the adapter sequences are present.
I joined the SAV screencapt (sorry for the bad quality). Indeed there seems to be a big issue with the color balancing...

[GenoMax]

Was phiX (or some other genomic DNA) spiked-in to help normalize the nucleotide distribution somewhat? If not that would be one thing to try along with under-loading of the library. You can probably read the amplicon sequence from the posted plot

[nucacidhunter]

I think there are at least two contributing factors:

1- PCR based addition of adapter is done in one or two step PCR protocols. The quality of oligos will affect both %PF and Q scores because any substitution or indel in oligos will affect sequencing primer binding and phasing/prephasing.

2- The client staggering techniques does not seem very successful. As GenoMax has mentioned this type of library need to be clustered under 800 K/mm2 and %15 or more PhiX to be added to obtain high Q scores. For longer reads cluster density need to be reduced further.

[thermophile]

something strange is happening to read1 too. I sequence mostly low diversity libraries and agree with the lower density and phiX spike in. How long are these amplicons?