>>16694149 (OP)There are already a million people doing every stupid analysis possible of public datasets that aren't contributing significantly to much, even if they're really smart people. The best ways to contribute are to:
a.) Get in the fucking lab test things for effects on animals
b.) Get good at math and coding and talk to people who do get in the fucking lab and learn about what kinds of analyses would actually be useful
c.) Get rich and fund people who are in the fucking lab
Generally, if your idea does not involve you or someone you know closely getting in the fucking lab, then it's likely not useful.