Reduce PhiX

[Gaetan]

Dear all,

I was wondering if there was any tool to help me reduce the PhiX% used in our NextSeq 500 runs? We run similar samples continuously, with "low" diversity as we enrich specific regions by hybrid capture. Instead of empirical & successive testings to reduce the PhiX%, can I guestimate the "ideal" PhiX% to add before reaching the slippery slope?
Thank you for your help.
Gaetan

[GSviral]

Hi Gaetan,

For sequencing on the MiSeq at least I know there is a recommendation to use a 5% PhiX spike-in for low diversity libraries (dependent on which version of the MiSeq Control Software you are using, you'll have to look that up.)

When I sequenced on the NextSeq I also used a hybridisation/capture approach however this was exome capture and the diversity was fine so I used a 1% PhiX spike-in for QC purposes.

For your purposes, a 5% minimum PhiX spike-in would probably do the trick, but I may stand corrected if someone else chimes in.

Cheers!

[Gaetan]

Hello,

Thank you for your quick answer!

Unfortunately, I cannot test right down to 5% as my samples are way to precious to risk a failed run because of low diversity.

So there no tools/tests to gauge the diversity based on my collected data that could help me to reduce the "wasted" output in sequencing "too much" PhiX?

I could safely slowly reduce the PhiX% each run until I see a deterioration in quality metrics but this empirical testing sounds like a waste of time if I am far off...

Thank you in advance!

Cheers

[GSviral]

Hi Gaetan,

Yeah it's always a bother when working with precious samples and NGS! I took a quick look and found this from August 2018:

http://emea.support.illumina.com/bulletins/2017/02/how-much-phix-spike-in-is-recommended-when-sequencing-low-divers.html

The NextSeq guidelines for PhiX spike-ins for low diversity libraries is 10% - 50%. I would be relatively confident that a 10% spike-in would be a good starting point. If you are still hesitant, increase that to 15% - 20%.


For your other question - unfortunately I am not aware of any tool to deduce diversity I'm afraid. Sorry!

Cheers.

[Gaetan]

Thank you again for your even quicker response!

I got those numbers too, but I guess it depends a lot on your target panel size, dups, or depth required for your analysis...

I will keep digging and if I found something, I'll post it!

Thanks again!