Cutting corners with Illumina Sample Prep

[Rocketknight]

Hey all, just a quick question regarding Illumina's TruSeq DNA sample prep protocol. After adapter ligation, there's a cleanup step with Ampure XP. If you're doing the gel-cut size-selection step though, that step comes directly afterwards. The manual seems to suggest that you need to do the Ampure cleanup even if you're planning to do the gel-cut afterwards, but I feel like electrophoresis should remove any contaminants well enough on its own.

Has anyone tried skipping straight to electrophoresis without the Ampure cleanup? Is there any reason that shouldn't work?

[ECO]

I just skip the electrophoresis altogether, having sized the population with AMPure before the ligation.

The AMPure prior to the gel is likely to ensure that a majority of the ligase and free adapter (not to mention buffer/salt/DTT/ATP) are removed before sizing the sample. That much protein will almost certainly alter migration patterns...but by how much I can't say. It's an easy experiment to do though.

[ETHANol]

Agree with ECO you need to do the AMPure purification before you run it on a gel. You are isolating small DNA fragments and you can't run the gel for very long or your gel slice will be too big. The salt front will likely run up against were you want to cut. If you want to cut corners, skip the gel purification.

[Rocketknight]

Ah, good points, both of you. I'm afraid I'll have to keep the gel-cut size-selection step though. I'm using a Bioruptor rather than a Covaris and I'm doing exome enrichment followed by paired-end sequencing. I'm worried that if I allow big inserts through I'll end up getting a lot of reads that lie outside exons. Is that just paranoia?

[Heisman]

Ah, good points, both of you. I'm afraid I'll have to keep the gel-cut size-selection step though. I'm using a Bioruptor rather than a Covaris and I'm doing exome enrichment followed by paired-end sequencing. I'm worried that if I allow big inserts through I'll end up getting a lot of reads that lie outside exons. Is that just paranoia?

If you sonicate on high for long enough I think you can get the Bioruptor size distribution down to a suitable size. May be worth a try.

[Rocketknight]

I can definitely get a peak in the size range I want, my main problem is that I feel the distribution will be broader than with Covaris, and so I'll get a lot of undesirably large inserts, even if the peak is in the right place.

[ECO]

Look into double sided AMPure cuts to avoid the gel....we do it all the time. They are described here and in the literature... I can't find the links now but that should get you started.

[Rocketknight]

Aha, awesome. I didn't realize bigger fragments outcompeted smaller ones when the beads were limited, I thought they just magically didn't bind fragments below a certain size or something. Is there a standard table for what bead:dna ratio I should use for each size cutoff, or should I determine my own?

[pmiguel]

Aha, awesome. I didn't realize bigger fragments outcompeted smaller ones when the beads were limited, I thought they just magically didn't bind fragments below a certain size or something. Is there a standard table for what bead:dna ratio I should use for each size cutoff, or should I determine my own?

The beads themselves are not the issue, are they? I understood that AmpureXP is a PEG cut with a fancy way to recover the precipitant, right?

Except I guess the beads themselves do have a maximum binding capacity. Anyone know what that is?

Oh, never mind, that is all covered here.
--
Phillip

[Rocketknight]

Ahh, my bad again. It's the concentration of PEG in the Ampure XP buffer that governs what fragment size will precipitate. Sorry for the confusion.

[tbanks]

This paper has some information on salt and PEG concentrations to size select using AMPure beads. They used a Bioanalyzer too.

[Rocketknight]

Thanks all for the replies! I just started another batch of samples and got really nice size selection with Ampure. I sheared with a Bioruptor down to a peak size of about 260bp, then did the standard Illumina prep up until the first cleanup after end repair, then replaced their Ampure cleanup with the following:

1) Add 90ul Ampure XP to the 100ul end-repaired DNA.
2) Incubate 20 minutes
3) Separate beads and take supernatant
4) Add 20ul Ampure XP to the supernatant
5) Incubate 7 minutes
6) Separate the beads and clean them as normal (ethanol wash, dry, resuspend, etc)

This got me pretty much the perfect size range for 2x100bp exome sequencing. The only problem is I estimate I lost about 80-85% of my input DNA in the process. I have plenty of sample DNA to work with, but I'd prefer to only do the size-selection on DNA I've end-repaired. Does anyone know how well the Illumina end repair mix works if you use 4ug of sheared DNA at that step instead of 1ug?

Also, Bioanalyzer trace for the Ampure size-selected DNA:

[Hamid]

Hi Rocketnight,

Did you by any chance do a Qubit analysis after you shearing to see if you are losing most of your DNA as single stranded DNA as a result of high energy over-shearing?
Comparing the Qubit results of your unsheared sample to the sheared sample will let you know if that is the cause of high loss of available sample for library preparation.

hamid

[Rocketknight]

I checked that, yeah. I didn't lose much DNA in shearing - I think it just gets lost during double-Ampure selection. Do other people lose much less DNA with it?

[luc]

You might recover a bit more DNA if you reduce the first bead incubation time to 10 minutes. It is my impression that the size selection is anything but absolute - i.e. with longer incubation more DNA will keep on precipitating.