DNA test for familial relationship

Say you have tissue samples from two different people, and you want to determine if and how they’re related. Can modern forensic CSI-style DNA tests determine the exact relationship… parent, offspring, sibling, niece/nephew, aunt/uncle, cousin, grandparent, etc? For the purpose of this question, assume the investigator has only the samples, and doesn’t know anything else about the people they came from.

*** Ponder

If I understand correctly, you could only get as far as “these two people share a female relative” or “these two share a male relative.” And that’s only if you are using labeled probes for the sex chromosomes. If you’re looking at the rest of the genome, you rely on odds. “only 1 in 100,000 people have this pattern of labeling on this chromosome” kind of thing.

Though maybe there’s a new technique I haven’t heard about.

Also, you get a “##% match” result, which is then translated as “they’re related” or “they’re barely from the same country”. Spontaneous mutations and chromosomal transfer can lead to, for example, a father and son having slightly different Y chromosomes, even though they are father and son.

Usually, mtDNA is used for the female line and the Y chromosome is used for the male line. A girl will also get an X chromosome from the father.

Geneticists can look at DNA and determine paternal and sibling relationships to a high degree of accuracy. The accuracy diminishes the further in distance the relative is, although I’m not sure how quickly that accuracy falls off.

Most of us have heard of the instance where one of Sally Hemmings’ descendants was traced to Thomas Jefferson (or one of his male relatives). But this involved a fluke on Jefferson’s part in that he (through his father) inherited a Y chromosome that is very rare in European men *. Had he been just another run of the mill European, we couldn’t have been so clear about tracing someone that far back since there would have been many men living at that time with the same Y chromosome who could have father Sally Hemmings’ child (or children).

*Somewhat more common in Spain and Portugal, but extremely rare in Britain. It’s actually more common in the Middle East, leading some to think that an ancestral Jefferson male was Jewish (that being the most likely M.E. ethnic group to have made it to England).

Thanks for the responses, but I haven’t quite gotten what I was looking for.

Is there any kind of lookup table, that lets you take that numeric result, and say X% means they’re parent and child, and Y% means they’re siblings, and Z% means they’re half-siblings, and W% means uh oh, that’s your uncle grandpa! (incest)

*** Ponder

No, you can’t tell what exact relationship exists, unless you have further information.

For instance, if you have a baby, and an adult male, you can determine whether that adult male is the father of the baby or not, because you know the baby can’t be the father of the adult. But you’re only testing whether the baby and the putative father are related.

But if you just have the data that the two individuals share certain genetic markers, you don’t know anything further. You could with further analysis determine if they were 50% related, which would mean parent-child or sibling relationship, but you couldn’t tell which. If they are 25% related, you would know they were grandparent-grandchild, half-sibling, aunt/uncle-niece/nephew, but no further.

Although standard paternity tests don’t do such analysis. They just look to see if the baby and the putative father share a few genetic markers, and if they do then it is extremely unlikely that the male is not the father. But this is simplified because there are typically only a few candidates for paternity, those that the mother had sex with around the time of conception. It would be a harder test if the pool of potential fathers included several related males, check your local listings for the latest Jerry Springer schedules.

Geneticist here, and no… you can’t. As Lemur866 points out, paternity testing looks at a few (highly variable) markers to see if they are shared between potential fathers and the offspring. In most cases, all these markers would be shared between closely related people, or too variable to be of any use. In either case, you cannot really point to anything that specific.

What did you mean by “further analysis”? I guess I didn’t state what I meant clearly: I was asking what the full gamut of tests available to modern science would be able to determine.

Hee! There were two different Spingeresque situations that originally prompted my question. :slight_smile: I know full well that people often lie or are misinformed, so I deliberately framed my question so that the researcher had no knowlege about the subjects… i.e. they can’t trust what the subjects are saying, and have to go by what the SCIENCE says. :slight_smile:

One of the situations was prompted by a Law and Order episode in which a girl discovered the person she thought was her sister turned out to be her mother. Her true mother was underage at the time and the family didn’t want the social stigma, so they buried the truth and lied to everyone. The grandmother just said, “OK, I’ll continue raising you both, and let’s just say she’s my daughter and you’re her sister.”

You’re stating above that DNA tests comparing the mother and daughter wouldn’t be able to conclusively prove or disprove a later assertion that her “sister” was really her mother? I guess further comparisons with the grandmother and grandfather would help, but IIRC, they weren’t very cooperative with the investigators, and the father was completely unavailable. Would testing them as well help pin down the truth, or just potentially show that grandmom was unfaithful with granddad?

The other situation I was considering was prompted by a discussion on another board about what would a typical CSI investigator conclude if the match was greater than 50%. How much over 50% would the match have to be before they would say “definitely incest somewhere in the family tree”? I guess this depends on how many markers tested for, and the likelyhood that normal full siblings might inherit more than half of the markers through sheer chance?

*** Ponder

Ok caveat a lawyer, not a geneticist, but have dealt with situations like that you have descibed, not exactly but a bit. From what I have been able to pick out

Typically more information is required for a legally conclusive ruling. There was a case here a few years ago, that the both purpoted fathers shared the a similar Y chromosone and thus the test was unconclusive. Thus tests were ordered on each mens maternal grandfathers (both of whom were alive at the time) and it was proved that one was papa and the other a mere “uncle”. No idea what the tests they used were.

Those guys must have been half-brothers, sharing the same father, right? Else, how could one be the father and the other be the brother, yet they have different maternal grandfathers?

They were unrelated. “uncle” was in inverted commas like you may call your fathers collegue when small. They did share the same Y Chromosone, which apparently is not at all unheard of.

Oh, OK. Yes, as I noted above, it is not at all uncommon for people whose ancestors come from the same general area to share the same Y chromosome or the same mtDNA (from the maternal line).