Can data be normalised before exploring it with DiscoRhythm?

[JarrigeD]

Hello! I am currently analysing a circadian RNAseq time series dataset for a training project and I am interested in using your package in R to identify gene expression rhythmicity. 

This might be a very naive question, but I was wondering whether performing a normalisation on the raw gene counts data (for instance with the regularised-logarithm transformation from the DESeq2 package) prior to analysis with DiscoRhythm would be appropriate. Or should only raw data be provided to DiscoRhythm? I am sorry if this is written somewhere in the doc, but I could not find the answer so far.

Thanks in advance!

[matthewcarlucci]

Hi @JarrigeD,

DiscoRhythm does not perform normalization (maybe this should be more explicitly stated in the documentation) and expects that data has been appropriately preprocessed prior to use. So, for raw RNAseq counts, I would recommend some kind of transformation though cannot give any strong recommendations. This RNAseq study that applied DiscoRhythm simply chose reads per million. Reading about each oscillation detection method may help you in deciding whether your preprocessing is appropriate.

You might get more general comments by posting support and usage questions to the support site: https://support.bioconductor.org/.

[JarrigeD]

Thank you very much for your answer @matthewcarlucci!
I will search further to find the most appropriate normalisation method.