I am looking to compare RNASeq results from 3 different studies that have uploaded their data to repositories. However, study 1 uploaded the raw fastq files, study 2 uploads raw mapped counts or normalized counts from HISAT2 data for study 2, and a third study that uses data for study 3. I want to compare these 3 studies, but unsure if I can compare the results.
My primary questions are 1) Is there any way I can compare these different studies RNASeq samples from different mapping pipelines, and if so how?
2) Assuming I ran study 1 raw fastq file to match that from study 2 or 3 depening on the pipeline, am I able to compare the results, as well can I group the the samples into the same control and diseased groups, or would I have to strictly compare study 1 disease to study 1 control and study 2/3 disease to study 2/3 control? (might be confusing, not quite sure how to word it).
Thanks for any help
There are inevitable technical differences and biological differences between studies, I do not recommend to mix samples as it will likely not give meaningful results. But you could try to perform a meta-analysis of the three studies : perform differential analysis or retrieve differential analysis results for all three studies (logFC, p-value for each transcript between control and condition groups) and then use a meta-analysis package such as MetaVolcanoR for instance