Why use DMEA
Even though many methods have been developed to predict beneficial drugs, most patients remain ineligible for targeted therapies and only about half of cancer patients deemed eligible respond to treatment. The power of DMEA lies in aggregating information across many drugs that share a common mechanism of action, rather than relying on a result from just a single drug. Simply put, if most of the drugs annotated with a mechanism of action are strong candidates, then we can be more confident the mechanism may truly be beneficial.
However, there are still limitations to DMEA. Perhaps the greatest limitation is that DMEA relies on the known mechanism of action annotations for each drug. Nevertheless, DMEA mitigates the risk of false positives by evaluating groups of drugs which share a mechanism of action rather than relying on a result for a single drug. Even if a drug is misannotated with a mechanism of action, that mechanism will not be significantly enriched unless most of the drugs in that mechanism set are strong candidates based on their rank in the input list. Also, even if DMEA works as intended to identify drug mechanisms for one phenotype versus another, not all samples represented by one phenotype may respond similarly to the same drug and other factors should be considered such as tissue type or tumor heterogeneity.
To learn more about DMEA, you can read our publication in BMC Bioinformatics here: https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-023-05343-8
To learn more about enrichment analysis in general, you can refer to this paper by Subramanian et al. and watch this YouTube video: