100+ years of collective accumulated research by a global community of geneticists who freely share their research with each other coupled with modern genetic simulation technologies

But, a lot of it is deductive reasoning, by sequencing genomes of literally millions of individual organisms, they can pay attention to shared traits and find individual genetic sequences that are shared by multiple individuals with those similar physiological traits; for example, by sequencing the genomes of multiple down syndrome patients they'll eventually notice all of them carry an extra chromosome and the same one at that,

before the days of modern database technology this process was exhaustive and required hundreds of people working together just to sequence one genome, now it can be done quicker but it's still quite exhaustive

In short: clinical research and statistics