Cost of RRBS and miSeq for ~50-100 samples?

[sulicon]

Hi,

We are planning to measure the methylation and miRNA expression profiles for a set of samples (50-100). For methylation measurement, we just need to cover the promoter regions, so RRBS is fine. Any platform is OK for us.

Illumina's website says that its RRBS platform costs $300-360/sample. I am not sure whether the estimated price of RRBS has everything included or not. Also, I don't know the actual cost of microRNA sequencing.

Could anyone share me some information about this? Thanks!

[dpryan]

What sort of organism? Keep in mind that you're not going to get a huge amount of coverage, since a MiSeq only outputs ~50million paired-end reads (i.e., sequence for ~25 million fragments) per run. If you actually used all 48 barcodes in a run then you'd only get ~500,000 fragments per sample...which would be far too few for many organisms (e.g., humans or mice).

[sulicon]

Hi Devon,

We are dealing with human samples. I am not familiar with miSeq. Per my understanding, there are only ~1100 human miRNAs annotated in miRBase, so ~500K fragments per sample means ~500 fragments per miRNA on average. Given the short length of miRNAs (~22nt), 500 framgents/miRNA == 500X coverage, which should be sufficient for expression estimation. Am I missing anything there?

[dpryan]

You're not going to only get miRNAs. What you end up enriching for is just small RNAs.

But anyway, my main reply was in regards to your desire to do RRBS, rather than small RNAseq, for which the miSeq makes more sense.

[sulicon]

Thank you Devon! I thought I could get the reads from only miRNAs. And I misunderstood miSeq as an abbr of "miRNA Sequening"...

[dpryan]

Ah, a miSeq is a specific machine and has nothing to do with miRNA. RRBS won't allow you to only target miRNAs (in short, you digest the genome and then size-select a range that's enriched in promoter regions in CpG islands). You want to do targeted bisulfite sequencing. The human pre-made ones would be a bit too general for you (they're mostly promoter regions, I believe), but I would assume that one could get a custom one made.

[sulicon]

Ah, a miSeq is a specific machine and has nothing to do with miRNA. RRBS won't allow you to only target miRNAs (in short, you digest the genome and then size-select a range that's enriched in promoter regions in CpG islands). You want to do targeted bisulfite sequencing. The human pre-made ones would be a bit too general for you (they're mostly promoter regions, I believe), but I would assume that one could get a custom one made.

Thank you again. Sorry for possible confusion here: We are interested in both general promoter DNA methylation patterns and miRNA expression profiles in human samples, as many as possible. The major concern for us is the cost. I am not sure how many samples we can process under a given budget (say 20K).

[dpryan]

You might post your needs over on Genohub and get some vendor quotes there. Even if you plan on doing some of the stuff yourself, you'll at least get a ballpark estimate that way.

[JeremyDay]

50-100 samples of miRNA and RRBS, will cost big bucks if you try to do it on a Miseq! Library prep is a little more complicated to find costs, but... there is an update coming soon for Rapid Runs on a Hiseq 2500. Considering we offer a SE 100-cycle Rapid Run for approx ~$500 more than the v3 600-cycle on Miseq, you'd get 10X more data. Of course that highly depends on the new software dealing with low diversity. When the Miseq got that update, it was a huge increase in performance. Anyhow, I guess I am saying the cost would go down if you can use a Rapid Run (update should be very soon). I would forget Miseq.

[sulicon]

You might post your needs over on Genohub and get some vendor quotes there. Even if you plan on doing some of the stuff yourself, you'll at least get a ballpark estimate that way.

thanks for the info!

[sulicon]

50-100 samples of miRNA and RRBS, will cost big bucks if you try to do it on a Miseq! Library prep is a little more complicated to find costs, but... there is an update coming soon for Rapid Runs on a Hiseq 2500. Considering we offer a SE 100-cycle Rapid Run for approx ~$500 more than the v3 600-cycle on Miseq, you'd get 10X more data. Of course that highly depends on the new software dealing with low diversity. When the Miseq got that update, it was a huge increase in performance. Anyhow, I guess I am saying the cost would go down if you can use a Rapid Run (update should be very soon). I would forget Miseq.

Thanks for the suggestion. Will have a look at Rapid Run