Capture or amplicon?

[Rocketknight]

Hey all, I'm looking to sequence human exons across 189 genes totalling about 700kb. Is there anyone who's tried both target capture and amplicon sequencing who could tell me which would be better for this?

If I do go for amplicon sequencing, is there anything weird I should expect in the results? Does the non-physical shearing used by Haloplex etc. give you weird fragment sizes or loads of false PCR duplicates, or is that mostly not a problem?

[zxyeo]

I am also interested in learning the differences between HaloPlex, SeqCap EZ and AmpliSeq for sequencing of ~100 human genes.

More specifically,
1. Which one will give a better on target percentage of coverage? Any known on target bias?
2. Any additional intrumentation required?
3. Which platform gives better cost effectiveness for sequencing ~100 human genes?
4. Any platform specific technical difficulty (experimental preparation) from different platforms?
5. Which one is likely to take longer preparation time?

Thanks!

[kwaraska]

I have done both HaloPLex and Exome capture from Agilent.

While HaloPLex is faster, as it makes the libraries and captures them all at once, the downside for our users has been that since it is an enzymatic shearing, it is far more difficult to pick up PCR duplicates as they all will start at similar locations. Random shearing is better to determine PCR duplicates.

Sure Select takes longer but is very robust.

Both procedures can be done with standard Mol Bio equipment like PCR machines and mag plates/stands.

I believe that Haloplex may only be for Illumina now, but I could be wrong on that. SureSelect works the same for multiple platforms.

I only have anecdotal evidence, but I believe Sure Select gives better on target capture.

I would suggest capture over amplicon sequencing as you will not know if you are missing something until it is too late with amplicon, whereas the algorithms used to make SureSelect and Haloplex probes will tell you ahead of time if you are missing a region.

I don't pretend to be an expert on this subject....just giving you my experiences

[zxyeo]

Thanks a lot kwaraska! It is interesting to learn about the difficulty in distinguishing the PCR duplicates using HaloPlex.

I wonder if it is cost effective to use Agilent Exome capture for ~100 genes. Although there is a SureDesign application for smaller scale capture, I heard that the requirement of DNA amount is similar to whole exome capture. Whereas Haloplex required little amount of DNA (Correct me if I were wrong). We hope to use as little DNA as possible since we are dealing with rare tumor samples.

[kwaraska]

Haloplex does require a smaller amount of DNA. If you can generate a Haloplex design, I say go for it.

[Genomemirabilis]

If coverage and uniformity is important, AmpliSeq is probably the best one (mainly for uniformity they get over 93% of bases with over 20% of the mean coverage). Just compare their specifications and I would try to do the design at AmpliSeq website. It´s free and quick to do it. There is also, 2 posts at Core Genomics that could help.

http://core-genomics.blogspot.com.br/2012/05/ampliseq-2.html

http://core-genomics.blogspot.co.uk/2012/03/ampliseq-design-tool-demo.html

Regarding the number of genes, it depends on the size of the total region to be sequenced. If they too large, the total panel could get next to 1GB. In this case, you can get some unespecific amplification. Not a big deal, but could require further sequencing. Anyway, you would have similar issues with capture.

[fashar1]

We just ran Haloplex on 48 samples and were not very impressed with the results. Our target region was 2.5Mb and we get >95% reads aligning to our target region, but the coverage was extremely uneven (60-80% of target covered at 20x).
As mentioned before, the enzymatic fragmentation makes identifying PCR dups almost impossible, and really it is a PCR-based amplification method, so you don't really want to be removing PCR dups.
However, the one thing that was different about our design than is usual for most people, is that we were targeted 4 chunks of the genome, introns, intergenic regions, the whole shebang, and that apparently makes probe-design more challenging.

[MWN]

How much DNA input is needed for Haloplex? I have seen the latest 200 ng input suggestion and even 50 ng input in Agilent slides. Hower at 50 ng, the allele frequences start to show larger variations.
Has anybody tested FFPE DNA and is willing to share your experience? I assume higher DNA input with PCR QC.
With 100 bp amplicon design, does Agilent recommend a different set of enzymes (several in one of 8 digestion pool)?