Cece Moore

So say matches one, three, and five share DNA with each other.

If I can build all their family trees, I should be able to identify their common ancestor.

The only reason two people would share these identical segments of DNA is if they inherited them from somebody in the past.

They have to have common ancestry.

And so if I can identify where that DNA comes from, which of the great, great grandparents are further back,

then that gives me one piece of my unknown person's tree.

And so I create what's called genetic networks.

I'll group the matches into networks of people that are sharing DNA with each other or clusters.

And each of those clusters will represent one branch of the unknown person's family tree.

So we start piecing it back together that way.

Maybe I'll have one set of great, great grandparents, one set of great, great, great grandparents.

Maybe if I'm lucky, I can identify great grandparents.

And then I have to find that one person or set of siblings that is related to all of those matches and descended from those sets of ancestors.

So it's like reverse engineering someone's family tree and eventually their identity based on their ancestors.

It depends how much data you have.

If you have enough matches that you can connect to someone's mother's side and their father's side, maybe three or even four of their grandparents' lines, you can narrow it down to just one immediate family.

But because these databases are so small, we often don't have that.

So say we could only identify one set of his great grandparents or great, great grandparents.

In that case, we would have to then do what's called reverse genealogy, identify all of their descendants and look for their descendants who are the right gender, age range, maybe live in the right area.

drive a white car.