It's important to understand in the ancestry.com case they didn't ask for information about a specific individual. They provided a sparse sample with a few mark

It's important to understand in the ancestry.com case they didn't ask for information about a specific individual. They provided a sparse sample with a few markers and asked ancestry.com, without a warrant, to do a fuzzy cold hit search of their entire database and provide them with information about any individuals that had partial matches. Ancestry found one, and related this existence, which was then speculated to be a possible relative. At that point police got an actual warrant to enable getting the name of this remote match, and ancestry.com provided it. The person was not even an ancestry.com customer. He was a person that participated in a scientific study years ago, the database results of which was later sold to by ancestry.com without his notified consent. There are massive statistical problems with this approach, without even having to get to the obvious privacy problems.