A DNA variation is a position in the DNA where people's letters differ — one person might have an A there, another a G. To link such a position to a disease, researchers compare two large groups: people who have the disease and people who do not. They read the same positions in everyone's DNA and count how often each letter appears in each group. If one letter turns up clearly more often in the disease group, that position is called associated with the disease.
The word "associated" is doing real work here. It describes a tendency across a whole population, not a prediction about any single person. A variation can raise risk and still leave most people who carry it healthy, and most people with the disease may not carry it at all. The size of the effect matters too: many genuine associations shift risk only slightly, which is why researchers need very large groups to see them above the noise of ordinary human variation.
There is also a location problem. Many associated positions sit in stretches of DNA that do not directly spell out a protein, so the link points to a region rather than to an obvious culprit. Turning that regional hint into a specific gene or protein is the next step.