Hi

I know that certain genes are more influential in progression of/ as causes of disaeases. Neglecting this fact, I want the answer to this question:

Considering that they are certain currently available/prescribed drugs for a certain disease. And that the gene expression profile of diseased vs normal tissue/cases is present (GEO). And the gene expression profile of drug-treated diseased tissue/cells is at hand.

What method should I use to see which drug best reverses gene expression profile of the disease? I have also seen ExpressionBlast website.

I think the method should be something like cosine similarity used in connectivity map or other re-positioning methods, see for example [1]. Or something as simple as clustering[2]/correlation/PCA -after correction for platform (as batch)-with do?

Platforms may be different: RNA-seq vs microarrays of different build.

References

1 A Large-Scale Gene Expression Intensity-Based Similarity Metric for Drug Repositioning https://www.sciencedirect.com/science/article/pii/S2589004218301299

2 https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-8-29