There are lots of streams of information you can use to help identify a set of candidates. Overlapping gene expression data, similarity ensemble approach, ChEMBL, and you can try docking your targets using something like AF3 or Autodock, just to name a few ideas.

However, your question doesn't suggest what class of molecules either list of compounds are, nor what their targets might be.

I don't work in human research, so maybe my view is biased from non-model research, but in my experience this can be a very difficult and time-consuming task with an extremely high error rate. My experience is in plant-pathogen immune protein interactions and the error rates are sky high. Only mentioning it as food for thought on proceeding with a purely in silico approach.