PhiX spike-in issue

[edwardwong1070584]

Hi all,

I have come across issues with the Illumina PhiX v3 in my project. I need to sequence my low diverse library on NextSeq 500. To balance out the diversity, I tried to spike-in 50% of the Illumina PhiX v3 control along with my library. But, the run wasn't good and I only managed to see a 9% alignment in my first run. At first, I thought it was resulted from the degradation of my PhiX. Thereby, I quantified my PhiX using picrogreen and the reading seems fine after conversion and the bioanalyzer showed no sign of degradation. I repeated my run the 2nd time. This time the Illumina FAS came over and supervised my library denaturation method to ensure I followed exactly the instructions recommended in the protocol, ie. library denaturation->library dilution to 1.8pM followed by PhiX denaturation->dilution of denatured PhiX to 1.8pm and, eventually, mixed both library and PhiX at 50:50 ratio. Everything was fine but the results of the 2nd run was still the same as in we couldn't achieved near 50% of PhiX spike-in. In addition, we observed a high error rate too. Does anyone have any suggestion on this issue? Thank you in advance for the guidance.

[nucacidhunter]

Three issues to consider:
1- your library and PhiX
2- sequencing chemistry
3- sequencing hardware

For correct diagnosis you need to provide some information from runs SAV metrics such as cluster density, PF%, % aligned to PhiX, Q30 scores

[edwardwong1070584]

Hi @nucacidhunter,

Thank you for your reply. Please see below for the metrics. Thank you in advance for your guidance.

[nucacidhunter]

The issues that is apparent:

1- Low cluster density that could be due to errors in quantification, denaturation or loading resulting in under loading or overloading (over clustering where several cluster are identified as a single cluster). Over clustering possibility can be checked by looking at images and also selecting intensity in Flow Cell Chart from analysis page which should show higher values (yellowish) in upper tiles in comparison to lower ones. In standard clustering lower tiles has higher density than the upper ones.

2- PF% very low: indicates low diversity (as you have mentioned) that has not been improved by PhiX spike-in because PhiX% is not high enough (this can be checked by looking at %Base in Data By Cycle plot in analysis tab).

3- High error rate: could be result of over clustering or poor sequencing primer compatibility (if custom primer was used)

4- If you can try quantifying PhiX by qPCR

I hope this and looking to similarities in second run helps you to identify the issue. If you need more help please post images from those analysis page plots.

[edwardwong1070584]

Thank you very much for your reply, @nucacidhunter. The Illumina FAS is currently checking through the run metrics. He is also suspecting that the library sample itself might be overclustered, resulting in the template generation to perform poorly and also exlaining on why the alignment didn't make sense. We had actually run the 3rd time, in which we lowered the library sample loading concentration from 1.8 to 1.5pM and it did improve the cluster density. But, the error rate was still high.

We did quantify our PhiX and the concentration was fine. So I guess we could omit the possibility of the degraded PhiX being used. I have attached the plot of data by cycle for this run for your reference.

Thank you