library prep batch effects

[chrisbala]

Hi Everyone,

Do any of you have a sense of how often (or how badly) you see batch effects of library prep and/or sequencing in RNA-seq data (Illumina Tru-Seq kits) ?

Say I've done an experiment, but I'm not totally convinced I want to run all my samples at once. How bad would it be to prep and sequence 1/2 the libraries now, and then 1/2 later once I am more certain the experiment is worth pursuing further?

In my case, I'm outsourcing both the library prep and the sequencing.

Thanks!

Chris

[kwaraska]

As long as it is the same service provider, I would not be too concerned. That is one of the advantages of NGS-you can add data run at different times and still be able to compare them. This is the major downfall of arrays, in my opinion. You can't run samples at different times.

I would just confirm with the provider that they will be using the same technology to prep both sets and run both sets. For example, they are currently making libraries on one robot and will be switching to a different one soon.

[chrisbala]

Thanks, thats what I'm thinking. I've had problems in the past, but that was when a number of things varied on the provider side. Barring that, perhaps its ok..let's see if anyone else has thoughts.

[weigrc]

We prepare libraries by ourselves. With the same one kits (TruSeq DNA + Exome kits), we did see the duplicate rate increased between 1st and 2nd batch of library preparation. The cDNA libraries prepared by TruSeq RNA kit may have the same issue since both DNA/RNA are using the same enzymes.

Wei