Question

Joint Preprocessing of RNA-Seq and Microarray datasets

3

Entering edit mode

Jo ▴ 30

@jo-8608

Last seen 8.9 years ago

Hi,

I would like to jointly pre-process RNA-Seq datasets and Microarray datasets together. Are there any pipelines or ways to do this? Please advise.

Thanks

preprocessing rnaseq microarray • 2.9k views

ADD COMMENT • link updated 9.0 years ago by Gordon Smyth 51k • written 9.0 years ago by Jo ▴ 30

0

Entering edit mode

That's going to depend very heavily on what you want to do with this data.

ADD REPLY • link 9.0 years ago Ryan C. Thompson ★ 7.9k

0

Entering edit mode

I want to normalize, remove batch effects (ComBat), and run limma on the datasets jointly.

ADD REPLY • link 9.0 years ago Jo ▴ 30

score 2 · Answer 1 · 2015-11-29

2

Entering edit mode

Gordon Smyth 51k

@gordon-smyth

Last seen 19 minutes ago

WEHI, Melbourne, Australia

No, you can't preprocess RNA-seq and microarrays together, nor would I include them in the same linear model in limma. The technologies are just too different.

Edit: The way that we usually handle this is to analyze the RNA-seq and microarray studies separately, then relate one experiment to the other using gene set tests (roast, barcodeplot etc) or a scatterplot of genewise log-fold-changes (e.g., microarray on the x-axis vs RNA-seq on the y-axis).

None of this requires batch correction or meta analysis.

ADD COMMENT • link 9.0 years ago Gordon Smyth 51k

0

Entering edit mode

Thanks for the reply. Is it possible to do a meta-analyses combining RNA-Seq datasets and Microarray datasets? If so how? Thanks in advance.

ADD REPLY • link 9.0 years ago Jo ▴ 30

0

Entering edit mode

Assuming that the RNA-Seq and microarray datasets are based on the same types of samples, the basic idea would be to process each separately, and then use something like the GeneMeta package to analyze.

ADD REPLY • link 9.0 years ago James W. MacDonald 67k