Hi all

I am trying to understand appropriate methods for quantile normalization. As is mentioned in multiple papers (e.g. DOI: 10.7717/peerj.1621), quantile normalization seems like a suitable method for making the datasets generated across different studies/platforms or even the data generated using different approaches (i.e. microarray and RNA-seq) comparable.

However, I could not find recommendations or guidelines regarding the required preprocessing steps in order to prepare the data for quantile normalization. In specific, I am wondering if the data from each study (be it microarray or RNA-seq) should be filtered based on intensities/counts, log2 transformed, or subjected to any type of prior normalizations before pooling the samples from different studies together and opting for quantile normalization. Starting with raw counts/intensities, I would greatly appreciate it if anyone could help me in any of the following scenarios:

1. Preprocessing of the data prior to pooling samples from two microarray studies
2. Preprocessing of the data prior to pooling samples from microarray experiment with those of an RNA-seq experiment
3. Preprocessing of the data prior to pooling samples from two RNA-seq experiments

Thanks in advance for your time and recommendations