Astronomical quotation

[mfors]

Hi everyone,

I hope you can help me, I have requested a quotation from a sequencing service company and the price is not what we expected.

We want to sequence a single gene from a large number of patients to look for mutations. The gene is less than 30kb and we are initially planning for hundreds of patients, maybe up to a thousand. The quotation was for the following set-up: sequence capture to isolate the gene of interest, followed by extensive barcoding and pooling before sequencing. Reducing the sequencing to our short gene and increasing the number of samples in each lane were considered clever ideas to minimise the price. However, the most costly part turns out to be the library prep which is done on each sample before sequence capture and any kind of pooling. Hence the hefty price.

I would be grateful if you could give me any tips on how to redesign my experiment!

My thoughts:
- Pool the samples prior to sending them in. Does anyone have experience with pooling samples before capture and sequencing? What do I have to think about when choosing how many samples to pool? The gene is only 30kb so I suppose we could pool quite a lot of samples, right?
- How many barcoded samples wold you recommend in a single sequencing lane? I have heard that 150 is feasible.
- We will buy the basic/standard bioinformatic analysis that will include SNP calling. Do you think that will be enough or will we need a skilled bioinformatician to help us out with the analysis? We have basic knowledge of bioinformatics.
- Long PCR instead of capture? But that would mean a lot of PCR runs. And I haven’t found any PCR kits that promises long enough amplicons to cover 30kb. And it doesn’t eliminate the LibPrep.

If anyone has an idea that does not involve high throughput sequencing but would give the results we want, I would be very grateful for that too.

Thank you!

[kmcarr]

If you want to track each patient individually then each sample needs to be uniquely barcoded, that's what is driving your cost.

This project sounds like it might be tailor made for the Illumina TruSeq Custom Amplicon application. Basically Illumina will design and synthesize a custom set of overlapping primers for your region (each pair spanning ~250 bp). Barcodes are added by nesting the target PCR reaction with an additional PCR to add on the barcode/library adapters. The products of the highly multiplexed PCR reaction are ready to go directly onto a flow cell. Currently they they only support the application on the MiSeq platform (it requires 2x150nt reads to sequence the 250bp products). I have no experience with this myself beyond viewing Illumina's webinar on the application but you might want to check it out.

[tbanks]

Can you consider multi-dimensional pool strategies with tagging? This article demonstrates using 454 to find EMS mutations in a plant population of 3000 M1 families. It would be a bit trickier using Illumina reads but not too difficult. You'd still need to make a bunch of libraries but should be less than 50 for 1000 individuals. If you went with this method you likely need to collaborate with a bioinformatician to deconvolute the results for you