whole blood RNA-seq without globin depletion

Question

[Sebastian Boegel]

I was asked to analyze whole blood RNA-Seq data and realized huge amounts of HBA1,2, HBB transcripts. So obviously the wet lab team has not done globin depletion. I would love to advice them for the next experiments, to do globing depletion, however I realized that all the public available SRA data I want to include in this study (to compare the samples with) also haven't done globin depletion. I am now wondering, if it is better to leave the protocol as it is (i.e. without globing depletion) to allow a better comparison of the own samples vs the public RNA-Seq data that also don't have globing depletion.

Any thoughts on this?

Thanks a lot,

Sebastian

[padwalmk]

Hi, For whole blood, you can remove the RBCs by using RBCs lysis buffer and it's much easy you can easily prepare in the lab and it takes an only half-hour. After that, you can store cells in the lysis buffer of any kit that you are using. I am already using this approach and carried out the RNA-Seq, working perfectly without any major effect. Even I can detect some transcripts that are coming only from my tumor samples [Because it does not express et al in blood].

[ATpoint]

I think OP was more referring to the impact of this batch effect when comparing to published non-depleted samples. I agree that ACK lysis is trivial in the lab, but this is not the issue here.

[padwalmk]

I knew he wanted some view on comparing globin depleted and non-depleted sample, but he has also asked for suggestions for his next experiment.

[Kevin Blighe]

If you are going to be working exclusively on datasets that have not done globin depletion, then neither perform the technique (globin depletion) on your own samples. This will, therefore, eliminate one extra source of bias. You will still have to account for batch, though.

You could probably still include non-globin-depleted datasets and include, as an extra covariate, GlobinDepleted ([TRUE/FA:LSE]).