Weird Bioanalyzer Image - Overamplification

[ETHANol]

I gave the Illumina TruSeq RNA Prep Kit v2 an initial go with two samples. I followed the Illumina protocol exactly, starting with 1 μg of HeLa cell RNA purified from Qiagen RNeasy columns.

I ran 1 ul of the library after amplification on a Bioanalyzer DNA1000 chip got a really strange image (see attached image A). I really can't make sense of what is going on here. My only thought was that it was some weird result of over-amplification so I took 1 ul of the library, mixed it with 1 ul of a homemade (IDT) PCR Primer Cocktail, 8 ul of H2O and 10 ul of Kapa SYBR FAST 2X master mix. I then subjected the sample to 5 cycles of PCR, purified the reaction on a Qiagen Qiaquick PCR purification spin column and ran 1 ul of that on another DNA1000 chip (see attached image B).

In the re-amplification the library looks better. There are some higher molecular weight fragments even though you can still see that there are plenty of primers left in the reaction. But I would still guess that the larger material is DNA fragments annealing by their adapter sequences.

Anyway, my question is has anyone ever seen a Bioanalyzer image like this? Did you sequence the library? What is your guess to what is going on?

[Susanne]

Could you maybe share the electropherogram images from the chip? The gel images could be misleading sometimes, depending on their gray-scale. Thanks.

[ETHANol]

I agree the electorpherogram is a much better representation, but in this case it really looks just like the gel image. I'd do it, but, not to sound too stupid, I can't figure out how to export the electropherograms out of the program.

[Susanne]

...just right-click on the epg. image and save as :-)

[Heisman]

When you click to export you should have the option of highlighting a bubble that has electropherogram image.

[pmiguel]

Hi Ethan,
About the electropherogram image that you are unable to export. Not sure what you are looking at it with. The best is 2100 Expert software -- if you actually own the Bioanalyzer that might be what you are using. Then you can do the right-click and copy trick. But some windows programs seem not to be able to accept the images via a paste from the paste buffer. So on my windows boxes I open powerpoint slide and paste the image in a slide there. From powerpoint, I right-click just on the image and chose "save as png" from the context menu.

Also the 2100 Expert software has a fairly comprehensive "export" function that can be used to export any or all of the .png images. (But if you want a pdf, you must use "print" to create it.)

About the gel images you did post. I get the strong impression that we are seeing only part of the story because the DNA 1000 chip is cutting off ds cDNA above 1500 bases. That should not be an issue, but HeLa cultures may give abundant mRNA. In which case you may (as you suspect) simply have over-amplified and ended up with the ectopic-annealling primed multimers that TruSeq kits frequently produce.

We typically do 10 cycles of amplification instead of 15 with the TruSeq RNA kits.

That said, I wonder if your RNA fragmentation step was less than what the protocol recommends? We mainly do PE RNA sequencing so we dropped from the recommended 8 minutes of high temp to 4 minutes. If you did something like that then the "B" images are about what I would expect.

The "A" images suggest a level of daisy chaining beyond anything I remember seeing before. But we have used only v1 kits thus far. It could be the v2 kits are more efficient and thus "overshoot" more easily.

Or, when you have sample smearing over the bioanalyzer's spike in oligo (15 bp, in this case), the 2100 Expert software will usually give completely incorrect sizes -- often fixing on the wrong "peak" as being its oligo. Same for the 1500 bp spike-in standard. So, you would want to check that, if possible.

--
Phillip

[ETHANol]

Looking into my image exporting problem, it basically comes down to the computer is running Windows 2000, is not connected to the network and has no software on it except for some really old version of the Bioanalyzer software. I'll have to get by without exporting the electropherograms.

Against my better judgement I followed the TruSeq protocol exactly as written (except for the typo where they say add 1 ul of Super Script II to 79.6 ul of First Strand Master Mix). It really is just inane that Illumina says do 15 cycles of PCR regardless. I used 1 ul of the library before PCR and did a qPCR using KAPA SYBER FAST 2x and achieved 50% amplification after 10 cycles, thus I am nearly certain the library was over-amplified so I understand the larger material. If I were using the KAPA HF polymerase to amplify these libraries, I would only need 8 cycles. It is the smaller stuff that I find really strange. There I think you are probably correct about one of the markers being masked. I would guess that if I diluted the sample and ran it on another chip things would be more explainable.

On a side note, it is annoying that Illumina doesn't give a good method to determine the number of cycles needed, like what I do when using the Kapa polymerase as explained above. I guess the more we have to work on the methodology the more kit we use.

[pmiguel]

Major typo in TruSeq RNA v2 protocol?

Against my better judgement I followed the TruSeq protocol exactly as written (except for the typo where they say add 1 ul of Super Script II to 79.6 ul of First Strand Master Mix).

Hi Ethan,
I have just gotten around to looking at the v2 protocol. From reading you blog post it looks like you noticed the issue because the kit did not contain enough First Strand Master Mix to support that level of dilution.

How did you come to "9.6" ul being correct?

The v2 protocol (page 47) has:

5 Add 50 ul SuperScript II to the First Strand Master Mix tube (ratio: 1 ul SuperScript II for each 79.6 ul First Strand Master Mix).

Whereas the v1 protocol (page 45) has:

5 Add 50 ul SuperScript II to the First Strand Master Mix tube (ratio: 1 ul SuperScript II for each 7 ul First Strand Master Mix).


Not sure what would be correct. Either 7 or 9.6 would probably work. But since you chose one, maybe you had some reason for choosing it?

--
Phillip

[ETHANol]

If you divide the number of microliters of the 'First Strand Master Mix' provided by Illumina in the TruSeq v2 kit by the number of microliters in the tube of SuperScript II they say to buy you get 9.6 ul.

So now I understand how the 7 got in there.

I've been meaning to post this but been too busy, but I diluted my over amplified library 1:10 and ran it on the bioanalyzer it it looked more like a standard over amplified library, i.e. a smear where you expect it and then another one higher up. So you were right, too much DNA was masking the marker, which resulted in the crazy bioanalyzer result. Thanks.

I've made about 30 libraries now and about 10 cycles seems to be towards the higher end of what is good. 15 cycles for sure is way over done.