Can I use NGS to analyze DNA methylation in only 5 regions of interest?

[jdperez]

Dear Community,

Im seeking advice in whether there is a trustworthy and economical way to use NGS for highly targeted DNA methylation analysis (only 5 loci!), rather than doing RRBS.

We are interested on allele-specific methylation patterns within a cluster of imprinted genes in mouse. As far as I know, people usually purify DNA from hybrids derived from distantly related mouse strains, which thanks to their numerous polymorphisms, increase heterozygosity and allow distinction of the two alleles. DNA is bisulfite converted, then primers are used to target the region(s) of interest, the amplicon is cloned and multiple clones are sequenced by Sanger.

Our problem is that we want to analyze allele-specific methylation in five regions across multiple tissues of multiple animals. Due to the number of samples the experiment becomes very laborious and also very expensive. I am wondering if it would be possible to perform the same protocol up to the PCR but then take the amplified DNA and use them to prepare samples for DNA-seq where we can multiplex, can sequence rather shallow and maybe get all of our results in two lanes.

Because of the PCR there will be a massive number of PCR duplicates. In the classic protocol the PCR has 35 cycles and I’m not sure of how much lower I can go since PCR in BS-converted DNA is already pretty finicky. However, I wonder how much of a problem PCR duplicates will be, because in the end all I care about are the ratios of methylation between the two alleles, and the alleles should be equally represented. I will know whether this is true since all polymorphic bases should be 1:1.

I just have 2 questions: What problems am I overlooking with this approach? Are there other ways?

Sorry for the long explanation and thanks in advance for your help!!

[nucacidhunter]

As far as I know, people usually purify DNA from hybrids derived from distantly related mouse strains, which thanks to their numerous polymorphisms, increase heterozygosity and allow distinction of the two alleles. DNA is bisulfite converted, then primers are used to target the region(s) of interest, the amplicon is cloned and multiple clones are sequenced by Sanger.

You can follow the described workflow with some modification. After BS conversion use primers with Illumina Nextera adapter overhangs similar to amplicon primer design described in post#2 linked document: http://seqanswers.com/forums/showthr...16s#post154446

Some optimisation of first step would be required to adapt two step PCR method. Amplicon length (excluding adapters) should not be more than 300 bp to obtain good quality reads. Every read would be equal to one clone.

[jdperez]

Thanks so much!! Having the index in my PCR primers is a much better way to go. Thanks for clarifying my concern about PCR duplicates as well.

[nucacidhunter]

That just was the simplest way to clarify the workflow. Some techniques that are used to increase amplicon diversity also can be integrated into protocol.

[jdperez]

Some techniques that are used to increase amplicon diversity also can be integrated into protocol.

Could you tell me about some of those? Thanks!

[nucacidhunter]

Easiest approach is to add 0-3 bases to 5’ end of locus specific primers resulting in 4 set of primers for each loci (for paired end sequencing). The bases can be degenerate or designed to introduce diversity in cycles. Ideally, in each initial cycle each nucleotide should be present in 25% of clusters. For degenerate approach equimolar mix of each primer(s) can be used in PCR. For designed approach every set of primers can be used during PCR for different samples with the aim of multiplexing amplicons from each 4 set for sequencing. Following is an example from an Illumina presentation for two step PCR:

Append to 5’ end of forward PCR primer:
5’ TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG-[locus specific sequence]
5’ TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGN-[locus specific sequence]
5’ TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGNN-[locus specific sequence]
5’ TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGNNN-[locus specific sequence]

Append to 5’ end of reverse PCR primers:
5’ GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG-[locus specific sequence]
5’ GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGN-[locus specific sequence]
5’ GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGNN-[locus specific sequence]
5’ GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGNNN-[locus specific sequence]

If sequencing output will be more than requirement for downstream analysis, diversity issues can be solved by PhiX or a high diversity library spike in into sequencing run. This in addition to diversity would overcome base balance issues as well and is more cost effective.

[jdperez]

Thanks so much, this is very useful!!