amplicon sequencing via miseq troubleshooting

[greenmachine]

hi- i'm trying to sequence an amplicon library (7 yeast genome loci) using our miseq. my most recent run, i used a v2 300 cycle nano kit. i loaded a 10pm library with a 15% phix spike in. the dominant reads in each library is a string of ~40 Ns (ranges from 50-80% of my reads). my amplicons are ~400 bases. any ideas how to troubleshoot or what is going on?

[Ben3]

greenmachine is this a custom amplicon protocol? And have you run this before with different results?

Also, what do the metrics from the run look like? When I ran similar amplicon sequencing protocols, I had to use a lot more PhiX to keep the diversity high enough for a clean run.

[greenmachine]

how much phix do you normally spike in?

yes - this is a custom amplicon protocol. it has worked in the past and i'm not sure what's different / why it isn't working now.

i had ~70% clusters passing filter (obviously not good)

[Ben3]

greenmachine if I remember correctly it was very high, maybe around 30%. I also sometimes spiked in other libraries to add diversity but still get useful data instead. But I also wonder if there was a problem during the PCR step and if it was the libraries themselves.

How do you QC the libraries before pooling and loading? Did you notice anything different this time?

[greenmachine]

Ben3 previously (when this build was working), i would just do the final pcr (with nextera adapters), pcr purify (qiagen min elute kit), and then qubit and pool the final libraries together in an equal molar ratio.

since this library build hasn't been working (and there has been this mysterious smaller product), i have started running the final pcr on a gel, extract the 400bp band, and then qubit + pool.

could there be degraded primers? or low quality starting material? or something else causing this issue? or is this just overloading?

[Ben3]

greenmachine, I suspect it's the primers as you mentioned. I would either order new primers or if you store the stock primers separately, I would take a new aliquot to test. If you didn't change the protocol, then that would be the first place I would start. I've had primers I used a lot in the past go bad and I just ordered new ones and everything worked great.

If that's not the case, I would try and see if you changed anything. I feel like amplicon sequencing is normally pretty robust and I've done it successfully on low-quality samples. It's normally only a big issue if they're really degraded. What do you do to QC your samples before the library prep?

[greenmachine]

Thanks Ben3 ! I've ordered new primers. Would that explain that the dominant sequence in my fastq is Ns?

After extracting the gDNA, I just PCR then qubit then PCR then qubit, no real qc. What do you normally do?

[Ben3]

greenmachine I suspect that it would explain why there are so many Ns, but then again it still seems a bit odd. I think it will at least help get rid of the smaller band you're seeing in your gel. My two other ideas are: 1) the flow cell could be a little overloaded. I usually loaded less on amplicon sequencing runs since the diversity was bad. 2) I added much higher amounts of PhiX to increase the diversity. It may take some playing around to get the exact amount you need and the best loading concentration.


Additionally, I would check the gDNA with the qubit to ensure I had a good extraction, and often times I would run it on the TapeStation using a gDNA screentape. This way I could make sure I didn't have too much degradation before going into PCR. You can run it on a gel, QIAxcel, or similar electrophoresis device just to check the quality. It may not always be necessary but I'm thinking it might be good to check before proceeding with the PCR on your next prep to be sure.