Always worth looking into. To my knowledge, previous studies have produced things like "these 600 genes are regulated in a particular way in patients who were

Always worth looking into. To my knowledge, previous studies have produced things like "these 600 genes are regulated in a particular way in patients who were diagnosed with Alzheimers", which is much less persuasive of an argument; the contribution power per gene is quite low in that case. If you can conclusively say "these n<5 genes are exactly the things that make people 50% more likely to develop a disease", you're cooking with gas.