Whole Genome Dna Amplification Before Applying Wgs

Question

[Thomas]

Dear all

We have some DNA samples that I would like to run whole genome sequencing on. Unfortunately these samples only contain very little DNA (too little for WGS). It has then been suggested to do whole genome amplification of the DNA before sequencing. This approach is commonly used before doing genotyping of a specific SNP.

Do you think that the pre-amplification will create a lot of bias in the seqeuncing (i.e generate many false postive genotype call, indell calls, CNV, etc.)?

Another approach could be to do single cell sequencing. What are the pros and cons when applying this?

Thanks a lot for any comments!

best wishes, Thomas

[Thomas]

Thanks to all the comments/concerns you have provided below. This is indeed very helpfull. I will talk to the lab about applying MALBAC to small amounts of DNA (not single Cell). If anyone has hands on experience with MALBAC I would very much like more comments. Thanks again. Thomas

[Mitch Bekritsky]

MALBAC is pretty recent...the paper was only published at the end of last year, and I first heard of it about two months ago. It may be hard to track down people with hands on experience outside of the lab that developed the technique.

Answer 1

[Christof Winter]

I guess the biggest in whole genome amplification is introduced by PCR: you will get an underrepresentation of both GC-rich and GC-low DNA fragments in your WGS results. Have a look at this paper: http://www.ncbi.nlm.nih.gov/pubmed/22323520 That means any variant calls in underrepresented regions will be affected (producing false negatives). Copy number might turn out fine if properly corrected for GC content.

Answer 2

[Chris Miller]

WGA introduces a ton of bias, most notably in the creation of regions with loss of heterozygosity, where one allele didn't amplify. I'm not saying it's never useful, but be cautious in your interpretation.

[Mitch Bekritsky]

That's why I recommended that Thomas consider MALBAC--at least according to the paper, it significantly reduces loss of heterozygosity, which would ease discovery.

Answer 3

[Mitch Bekritsky]

I'm no expert, but maybe try a linear amplification protocol? I've heard good things about MALBAC. It has reduced amplification bias compared to standard WGA, which will mean you can more accurately call CNVs, indels and SNPs, although according to the paper, there is still a ~1% allele dropout rate (which I think is still much better than WGA). The paper describes its use for single cell sequencing, but it could work in any experimental setup where you don't have a lot of input DNA. In this case, I don't know that there's a tradeoff between low input vs single cell, provided you have a good method to isolate single cells prior to amplification.

There's no mention in the paper of how the method performs on extreme GC biased regions though, so Christof's concerns are still valid.