Whole-genome amplification prior to Illumina sequencing?

[Crawford505]

Hello,

Does anyone have any experience sequencing samples on the Illumina platform that have been whole-genome amplified? Has anyone explicitly tested whether it introduces allelic/amplification bias that could be a problem for SNP calling?

We've heard rumors and impressions that say it's a problem and should be avoided. But we have valuable samples that we would like to sequence and WGA seems to be the only way.

Thanks in advance.

[henry.wood]

How tiny are your samples. If you are gentle you can make a perfectly good library with a lot less than 1ug.

[Crawford505]

Many of them are in the 400-500ng region. Have you had success with samples in this range. Do you modify the standard protocol to accommodate the lower grammage?
Thanks.

[henry.wood]

We don't bother tweaking anything until we get to less than 100ng, so hopefully 400-500ng should be fine using normal protocols. If you are feeling twitchy you could practice using 400ng of a non-precious sample. You can tell if a library has worked by looking at the agilent/picogreen results without having to sequence it.
Once we get to low concentration we start to fiddle around with adaptor concentrations and PCR cycles.

[Crawford505]

Thanks very much that is helpful.

However, we are considering outsourcing our sequencing (the price is right) and the facility that would do it for us is requiring what i'm assuming they are considering 'safe' amounts of starting material.

So whole-genome amplification is still something we may want to look into. Anyone have any experience with this?

[jasonwood]

I used to use Sigma WGA to amplify ChIP samples for microarrays. Haven't needed to do this since I routinely make illumina libraries from 20ng or so starting material, and have been getting good sequencing results. One thing to keep in mind is that many WGA systems are PCR based, so if you then make libraries from them, you'll just end up sequencing the WGA adapter oligo instead of the sample. Maybe another WGA technique like MDA might work, but I don't have any experience with that.
But since the illumina library prep already has an amplification built in, i don't see any need to pre-amplify. you can use very small starting amounts. Find a place that will sequence pre-made libraries if necessary, and make them yourself.

[csquared]

We haven't had any trouble making libraries down to 10ng input in our service lab and lower in the R&D lab. You need to appropriately alter the ratios in the library prep but it's nothing difficult.

[krobison]

A number of vendors are introducing kits for smaller amounts of DNA; my notes say that NuGEN's requests 200ng and I know somewhere another company claims to have kits with a 100ng requirement (but those notes aren't making themselves found).

I also have data from WGA & at a crude level it looks fine. One thing to be careful of is any procedure using PCR -- you do risk having the PCR primer(s) occupying the ends of many of your reads.

[Crawford505]

Just so we're clear, we are interested in library preps of genomic DNA for DNA-seq runs. Are the numbers referred to here applicable to gDNA preps?

[jasonwood]

Yes, you should be fine with just a few ng of fragmented gDNA. Most of my samples are chip-seq (which is also fragmented genomic), and i start with 20ng if available, but routinely go down below 10ng without many problems if the IP is low efficiency. Certainly anything in the 20-100 range is quite robust (with diluted adapter primers).

[kainsteven]

The Ovation WGA FFPE System from NuGEN can be used for DNA from FFPE or Fresh Frozen (FF) material with input amounts as low as 100 ng.

Steve

[rthornton4]

Hi,
We just attempted to sequence a WGA ChIP-seq library on our GA-II and it failed miserably. The adapter that Sigma uses in their WGA kit confused the RTA software and what the sequencer did call was junk. From talking with our FAS, the problem is that the new software needs to identify all four bases to call any single cluster and this adapter does not accommodate that. Currently we are being told that the only work-around is to design a sequencing primer that will hybridize to the adapter sequence or run these samples on the older software.

[frozenlyse]

Alternatively you could concatamerise your WGA'd material and refragment, that way the WGA primer sequences will not be on the end of your fragments - of course your downstream analysis will be made that much harder... You are probably better off optimising Illumina library prep for small amounts of DNA (which many labs have done)

[rthornton4]

Thanks frozenlyse for the advice! Unfortunately I didn't prepare the WGA sample. It was prepped by an outside user and I made the library. I do agree with optimizing library preps with smaller amounts of input but I don't have all of the information on why the user WGA'd their material. Another user had done this to compare WGA material vs. non-WGA material to see what, if any, the differences were in the sequencing. They had found the WGA material was highly comparable to the non-WGA material. I don't know if that was this users intent. All I do know is that they are not happy with Illumina currently.

[elaney_k]

Thanks frozenlyse for the advice! Unfortunately I didn't prepare the WGA sample. It was prepped by an outside user and I made the library. I do agree with optimizing library preps with smaller amounts of input but I don't have all of the information on why the user WGA'd their material. Another user had done this to compare WGA material vs. non-WGA material to see what, if any, the differences were in the sequencing. They had found the WGA material was highly comparable to the non-WGA material. I don't know if that was this users intent. All I do know is that they are not happy with Illumina currently.

Hi,

I've managed to sequence samples for a collaborator on the Illumina GA that were target-enriched using the nimblegen/454 prep. The input DNA had the 454 sequencing adapters on either end of the DNA (22 bases long on the 5' end). I thought about concatamerising the DNA before doing an Illumina prep but figured the downstream bioinformatics might be a bit of a nightmare so prepped them without it. I made sure not to overload the flowcell and recalled all the data from the saved images. I got great usable data from it. I have to say in your case Illumina can't really be blamed, if you understand how the cluster calling algorithm/chemistry works then it's much easier to predict these kinds of problems.

Elaine