Next-gen sequencing of targeted genes with many DNA templates

[gmoxley]

I wish to use next-gen sequencing for rare variants in about 800 disease and 200 control samples. The limitation is that total template available for the project is 500 ng total per individual sample. Targeted regions are promoter regions (2 kb upstream) plus short exons 100-300 bp long with adjacent exon-intron boundaries plus 3' UTR 500 bp, totaling maybe 30 kb, distributed over 400+ kb. The exons are pretty far apart for long-range PCR, I think.

Short questions for the gurus:
Does one of the genomic target enrichment methods lend itself to quite limited template like this?
Does the strategy of pooling the samples, then doing individual PCRs for each short fragment, and then sequencing to detect common and rare variants make more sense?
Which template normalization method would work best for shorter fragments, less than 500 bp? For long-range PCR products?

Thank you. George Moxley

[krobison]

The amounts of DNA you have are definitely a challenge for any of the target enrichment schemes out there -- they tend to recommend closer to 3-5ug of input DNA. It might be worth looking in detail how many long-range PCRs would be required to cover your targets -- that would probably be your best shot for working with such small sample amounts -- but that is probably also a lot of LR-PCR to get working.

You might try using a Whole Genome Amplification technique prior to library construction.

If you are looking for common variants, then pooling could be a reasonable strategy -- you just will need to save template so you can deconvolve the mutations later.

If you did a simple 20-fold pooling, then you would have 40 disease pools & 10 control pools. -- might be a bit small for RainDance (tends to work best I understand in even batches of 96) but that or SureSelect would be reasonable options.

[der_eiskern]

I wish to use next-gen sequencing for rare variants in about 800 disease and 200 control samples. The limitation is that total template available for the project is 500 ng total per individual sample. Targeted regions are promoter regions (2 kb upstream) plus short exons 100-300 bp long with adjacent exon-intron boundaries plus 3' UTR 500 bp, totaling maybe 30 kb, distributed over 400+ kb. The exons are pretty far apart for long-range PCR, I think.

Short questions for the gurus:
Does one of the genomic target enrichment methods lend itself to quite limited template like this?
Does the strategy of pooling the samples, then doing individual PCRs for each short fragment, and then sequencing to detect common and rare variants make more sense?
Which template normalization method would work best for shorter fragments, less than 500 bp? For long-range PCR products?

Thank you. George Moxley

many of the NGS companies have protocols for such small quantities of isolated nucleic acid. i'd call your favorites and ask tech support to ask R&D.

one strategy you might want to consider is to shear them all using the covaris ultrasonicator. attach a universal adaptor to the ends that has a restriction site you can use to cleave of the adaptor and do rolling circle amplification. then digest. use your favorite exome capture method. you could pool all your samples for the exome capture if you're on a budget but i would think capturing each would be better because you'd be able to determine if the lack of sequences could be traced back to not having enough starting genomic dna that could be captured. 500 ng of worm or fly disease dna would give you a lot more pulled down than 500 ng of human dna.

i should warn everyone that I haven't tried this approach myself but assuming you can get good ligation efficiency in your hands, i think it'll work.

-Todd

[gmoxley]

Thank you!

[dottomarco]

MYcroarray.com are starting to provide liquid enrichment altough they target few Mbs, and they require few ngs of libraries.
No idea about how their probes work for actual enrichment. I am testing them myself on a 454