MiSeq v2 cluster density 1/2?

[pmiguel]

Did a run last week on a v1 hardware MiSeq using a v2 chemistry 500 cycle kit on qPCR titrated libraries. Clustered at 10 pM. Instead of the expected 800-1000 clusters/mm^2 we got 425.

Again these were with KAPA qPCR titrated libraries using the same protocol that usually allows us to nail cluster density. These TruSeq RNAseq libraries were a little odd in that the insert sizes were a mean of 700-800 bp to get the full value of the 250 base reads. But we have gotten dead-on accurate clustering results for libraries with much larger mean inserts sizes than these in the past. (Obviously this does require compensation for the larger amplicon sizes compared to the KAPA standards.)

This is the first time we have used v2 chemistry, however. But Illumina Tech support says there should be no major difference in cluster densities with v2. I have another run to do with another set of similar libraries, but I don't want to waste the reagents getting 1/2 the amount of data I should. On the other hand clustering at 20 pM seems too aggressive.

Anyone seeing much lower than expected cluster densities running v2 MiSeq chemistry? Especially with fairly long amplicons?

--
Phillip

[Bucky]

We had the same problem here. We used to cluster at 8pM and that resulted in about 700k/mm2. Switching to v2, we ended up with 200k/mm2. Loaded the latest run at 15pm and that seemed to do the trick (1000k/mm2). Will probably set up the next run at 13 or 14pM.

[pmiguel]

Thanks Bucky,
I have been stripping my gears trying to figure out what was going on. That makes it seem like v2 performs more like a HiSeq with respect to raw cluster density[1].

Are you running v1 hardware, or has your instrument been upgraded to v2?

Anyone else see this or not see this?




[1]Of course as a matter of efficiency the cBot blows the MiSeq away -- 120 ul/lane for the HiSeq gets you the same density that 600 ul gets you for the MiSeq.

--
Phillip

[Bucky]

Hey Philip,

We are using v1 hardware. I should also mention that I pooled my sample at 4pm and then diluted it out more than usual in order to get a lower NaOH final concentration. Maybe that helped as well..

[GenoMax]

Philip,

Disclaimer: I do not work in the lab. I am merely passing the following info along.

We are not seeing this problem with v.2. kits but the inserts in question are not long.

[pmiguel]

Hey Philip,

We are using v1 hardware. I should also mention that I pooled my sample at 4pm and then diluted it out more than usual in order to get a lower NaOH final concentration. Maybe that helped as well..

Not sure I follow you. Normal protocol is to mix 10 ul of 0.2 M NaOH with 10 ul of 2 nM (total concentration of all libraries), incubate, then neutralize with 980 ul (50xdilution) of HT1. That gives you 1 ml of 20 pM denatured libraries, that you can dilute further, as needed.

Could you elaborate, please?

--
Phillip

[pmiguel]

Philip,

Disclaimer: I do not work in the lab. I am merely passing the following info along.

We are not seeing this problem with v.2. kits but the inserts in question are not long.

Is this v2 kits on v2 hardware, or v1 hardware?

--
Phillip

[GenoMax]

Is this v2 kits on v2 hardware, or v1 hardware?

--
Phillip

This is v.2. kits on v.2 hardware.

[pmiguel]

This is v.2. kits on v.2 hardware.

Okay, this may be a v1 hardware/v2 reagent thing then. I heard from another core that they saw lower cluster densities ("like 1/2") using v2 chemistry on a v1 instrument. But that apparently went away when they were upgraded to v2 hardware.

--
Phillip

[Bucky]

Not sure I follow you. Normal protocol is to mix 10 ul of 0.2 M NaOH with 10 ul of 2 nM (total concentration of all libraries), incubate, then neutralize with 980 ul (50xdilution) of HT1. That gives you 1 ml of 20 pM denatured libraries, that you can dilute further, as needed.

Could you elaborate, please?

--
Phillip

Right. Instead of starting with 10ul of 2nM, I start with 10ul of 4nM. So in order to get to my desired final concentration, I need to dilute out more, thus lower NaOH concentration. Does that make sense?

[pmiguel]

Right. Instead of starting with 10ul of 2nM, I start with 10ul of 4nM. So in order to get to my desired final concentration, I need to dilute out more, thus lower NaOH concentration. Does that make sense?

Ah, okay, got it.

--
Phillip

[GW_OK]

Illumina put out a new methodology for denaturation with the v2 chemistry. It's in the new Rev E version of the Miseq users guide.

[pmiguel]

Illumina put out a new methodology for denaturation with the v2 chemistry. It's in the new Rev E version of the Miseq users guide.

I downloaded it to take a look. Not seeing the difference in the denaturation method though. Looks pretty much the same to me.

It is different from the HiSeq -- there the concentration of NaOH used is 1/2 that than for the MiSeq denaturation. Is that new with Rev E of the MiSeq UG?

Anyway, we were using the 0.2 N NaOH, just as called for in Rev E.

--
Phillip

[pmiguel]

After complaining the Illumina about this issue they told me they did not equipment in-house to test it out. (Meaning, I presume, all their in-house MiSeq have v2 hardware.) So they offered to send a couple of MiSeq kits out to test this hypothesis. Just started the v1 kit this morning.

Clustered at 10 pM and got 758 Kclusters/mm^2. That seems a little lower than normal, I think. Actually, I ran the same library on v2 chemistry 2 days ago at 15 pM and got a cluster density less than 500 Kclusters/mm^2. But Illumina wants the experiment to be done by the same person, same day, etc.

I'll let it run until cycle 26 and then abort the run, so I can get the second run started today.

--
Phillip

[Bucky]

Cool, let us know what you find out!