Difference between dataset analysis

Question

[BioQueen]

Hi! I'm going to merge several datasets from different studies from GEO, and I was wondering how I can check how different the datasets are compared to each other?

Do I merge the datasets to one big set first, or do I do the analysis separately on each dataset?

Thanks!

Answer 1

[ATpoint]

This is super open-ended and therefore cannot be answered in detail. In general OMICS datasets can only be compared directly when processed identically in wet- and drylab. That is not the case for public data when using multiple datasets. This is where meta/rank-based comparisons come into play, e.g. RobustRankAggregation where you compare the rank distribution of e.g. genes between studies, or something like Gene Set Enrichment Analysis where one study serves to build a hypothesis (a gene set, e.g. DEGs) and then the others are checked against it, whether the overall distribution (aka as a "trend") is the same.