Control sample in Next-gen sequencing

[cub103]

Hello all,

I never done any sequencing but am planning to run some cancer DNA on Illumina GA. My question is: 1) Am I right that one flow cell can run 8 sample with one sample in each lane? 2) If 1) is yes, is it good to run only 1 control (or we need more) to compare the cases and control in order to get useful information?

Thanks a lot!

Cub

[cbrennan]

1) You can multiplex and run several samples in each lane -- we've actually had a client design their own barcodes and run over 3,000 samples on one flowcell. If you stick with the Illumina protocol, I believe their multiplex limit is 96 per flowcell.

As far as #cases/controls - from what we see (working at a CoreLab) that seems to be limited by how much sequencing the client can afford. I would suggest you first determine how much coverage you need in in your area of interest and then determine the best way to obtain the coverage.

For human genomic DNA, with a 120 PE we typically get 2X mapped, de-duped coverage per lane, for 80SR we see roughly .75X coverage. Of course those numbers can very a lot depending on the quality of your libraries.

[krobison]

What do you mean by control sample? What are you controlling for?

The standard for sequencing genomic DNA from cancer is to run a normal tissue sample for each patient; there are still far too many SNPs (and other variants) in the human population to rely on databases to distinguish mutations from germline variation. I think the U87MG paper is the only one which has deviated from this, simply because a normal sample for this cell line is not available.

If you mean the Illumina phiX control, then most labs seem to be sticking to burning one lane with it. Some labs spike it in, but people I talk to (I don't have my own instrument) seem to think this is still bleeding edge.

[cub103]

Load of thanks to cbrennan and krobison for the informative reply.

To Cbrenan- We are actually planning to run Exome sequencing using Agilent Whole Exon Kit (~3.3Mb).

To Krobison- I think I was talking about both controls. We may finally run a group of cancer Genomic DNA and compare them with controls from other institute, wondering if this is a good idea?

In addition, I can't not really understand the "read depth" and the "-X coverage, wondering if anybody could give a plain explanation?

Cub

[huguesparri]

The read depth is how many time you're going to sequence each base of your genome. A read depth of 3 means that you're able to read at least 3 times every bases of your genome. An otpimum read depth is depending on your experiment and on your sequencing read length (De novo sequencing is usually the most demanding applicaton).

The coverage is the sequenced percentage of the whole genome. A 0.75 coverage means that after aligning your sequence to your reference genome, there is still 25% of this reference genome that wasn't sequenced.

[cub103]

The read depth is how many time you're going to sequence each base of your genome. A read depth of 3 means that you're able to read at least 3 times every bases of your genome. An otpimum read depth is depending on your experiment and on your sequencing read length (De novo sequencing is usually the most demanding applicaton).

The coverage is the sequenced percentage of the whole genome. A 0.75 coverage means that after aligning your sequence to your reference genome, there is still 25% of this reference genome that wasn't sequenced.

Thanks for your reply. Do you have any suggestion about the minimal read depth to get meaningful results and minimal acceptable coverage? Thanks.