During our curation, we notice that signatures from many studies are very heterogenous even for the same disease conditions. These reflect more contextual differences rather than the disease names. It is better to keep the specific study contexts rather than merging them. Users should explore the original studies to find out more specifics.
For instance, for sets tagged BugSigDB are one set per study, not one set per condition. BugSigDB curates published signatures individually, and studies of the same disease differ in cohort, country, sequencing method and statistics, so we kept them separate rather than merging them into a single “Obesity” set.
In another word: if your taxa enrich for five different obesity studies, that is stronger evidence than a hit on one merged set would be, and you can click View on each row to see which taxa overlap and follow the PubMed link. The older sets (e.g. Obesity (Increase) without a suffix) come from GutMDisorder and are still there; the Evidence field in the detail view tells you the source of every set.