First, we know that genetics begins with the premise that we are all either XX if female or XY if male. When I studied genetics in high school, the standard idea was that we are exactly 50 per cent of each of our parents. Well, no. It turns out that we are more our mothers than our fathers (more egg than sperm), and that certain aspects of a person come from a particular parent, such as intellect.
So, if we think of tracing our matrilineal line, i.e. our mother's mother's mother, we track the X, and wind up with what Bryan Sykes of Oxford calls the Daughters of Eve. And, there are only a handful of women from whom we all descend.
So, tracing the Y yields our father's father's father's line, the patrilineal line. So, the idea would then be that the Y doesn't change...but it does. In the work done to confirm the identity of Czar Nicholas II, it became apparent that not only does the Y mutate over time (roughly every 150 years), but that some people actually show the mutation (as did the Czar) in their genes. So, assuming that while a Jefferson fathered Sally Hemings' child, DNA evidence does not show which Jefferson. Family Tree DNA now offers a Jefferson-matching DNA service, noting that
"While DNA cannot prove that Sally Hemmings' child was Thomas' son, Y-DNA did prove that the child's father was 'a Jefferson.' Family Tree DNA has the Y-DNA signature of the Jefferson line:
If you want to have the test, the link is http://www.familytreedna.com/landing/matching-jefferson.aspx
For books by Professor Bryan Sykes, see http://books.google.com/books?q=+inauthor:%22Bryan+Sykes%22&source=gbs_authrefine_t