Making a custom run recipe on nextseq 500

[provirus]

Is it possible to design a complete custom run recipe by yourself on a Nextseq500. For instance I am considering a PE sequence run were some specific cycles are chemistry-only (with no pictures taken) due to some sequence patches in my library with no diversity.
It seems Illumina have no software to support this. They ask one to contact their tech support in order to discuss such possibilities.
While this is of cause the best way to proceed if one like to make a custom run, I am just curious as to whether one could actually make such recipe yourself ? What files are involved in instructing the apparatus and what are the format (etc). Any experience in this from anyone greatly appreciated ?
Thanks.

[MisUser]

Hi Provirus,

Ask your local FSA if he/she can provide you a custom recipe file with the spec you want. Illumina doesn't advertise it but they gave us the files necessary to run dark cycles (cycles without imaging).
Good luck

[provirus]

Thanks. We will indeed approach Illumina with this. I wanted to know if some users had experience with this, and also to learn about the "inner workings" of the programs that control the runs - it is a black box to me.

[austinso]

Been a while since I've done this on the MiSeq, and I know they've consistently updated these files to make it harder to modify...

All the protocols have traditionally been in XML files. So if you go into the Illumina folders on the c: drive, and into the control software folder (not sure what it is called in the NextSeq), under the recipes folder, you should have a subfolders corresponding to various recipes corresponding to preprogrammed methods that pop up in the menu. It's an interesting exercise to understand how they are organized. Frankly, they are messy IMHO...it helps to map out how they are connected.

My sense it that playing with the xml file in the exposure folder would be enough.

But...do you really need to do that? As long as you have diversity in the first 8 (?) bases it should be okay...or, spiking in PhiX at higher concentrations would also be an option.

I may be missing something, though...

Best, Austin

[provirus]

Hi Austin,
interestingly. I will have a look in the folders, and see if it makes any sense. You are right that PhiX spike would help. We are testing this at 35-45% at the expense of library.of-interest - it would be nice if a dark cycles-run would give us more sequence-per-dollar. It's annoying it has to be such a pest to customize runs. (By the way, we have low diversity in the very first cycles - 8 plus some).

Thanks.

[austinso]

Ultimately, the caveat is that the first few bases are used as the QC step to identify successful cluster generation. So having dark cycles may not be in your best interest. But again depends on your application.

I've used 1% phiX for lower diversity runs on the NextSeq (15K primer extension sets), which were fine.

What I've also done in the past is design into the adaptor a mix of pads to offset the reads for extremely low diversity libraries. YMMV of course.

[provirus]

Hi Austin. Is it correctly understood that you had the low diversity regions/patches offset from one another by using 2+ different adapter/amplicon ends (effectively 'moving' the patches relatively to each other) ?
Regards.

[austinso]

Essentially, yes. Either a fixed number of random bases (e.g. 5'-end of primer), or adding 0,+1, +2 bases of known composition.

Hi Austin. Is it correctly understood that you had the low diversity regions/patches offset from one another by using 2+ different adapter/amplicon ends (effectively 'moving' the patches relatively to each other) ?
Regards.

[JakobHedegaard]

Hi Provirus and Austinso,
What a coincidence. We have just been in contact with Illuminas Techsupport to have a custom program made for the NextSeq 500 to do a PE run with 5 dark cycles after the first 12 cycles : read 1: 151 (1-12 + 13-17 (chemistry-only) + 18-151) cycles read 2 (Index): 6 cycles read 3 (reverse): 151 (1-12 + 13-17 (chemistry-only) + 18-151) cycles.
Would be easier if we could modify the programs, but since we are certified it would be a pain.
/Jakob

[JakobHedegaard]

Hi Provirus,
Please send me a privat message.
/Jakob

[JakobHedegaard]

A update on the chemistry-only recipe....
We got a custom recipe from Illumina (without any guarantee) but it didn't work. It was supposed to do chemistry only in cycle 13-17 - but it didn’t.
/Jakob

[austinso]

Doing this for a clinical lab is always a caveat...

Anyhow, I think all you would need to do is remove/comment out the lines from the Sony.xml file that is in the Exposures folder corresponding to the reads that you want to be dark.

Note that this has to be modified in the main drive, as all the files are copied over to the run drive, so will overwrite any mods present there (could that have been a reason why the modified file didn't work...?)

[GenoMax]

Doing this for a clinical lab is always a caveat...

True. There is a reason not to be easily able to mess with a sequencer certified for clinical sequencing.

That said, Nextseq 550Dx (newly certified NextSeq for Clinical applications) supposedly has a "research" mode (distinct from diagnostic mode) that probably allows for thing like this while keeping the certification intact.