Over at the Salon, a Dear Prudie column discusses the issue of miss-attributed parentage.
My husband was estranged from his parents for many years. He reached out to them when he was diagnosed with a terminal illness. They didn't have enough time to discuss and resolve their past, but they were at peace with each other when he died. Now my husband's parents wish to keep in touch with me and my toddler-age son, as he is the only link they have to their only child. The problem is that my son is not my husband's biological child. I had an affair
So the interesting part here is the growing trend for personal genetic services such as 23 and me. Among other things that you can do with 23 and me is to identify possible relatives in their database. As in, strangers who you don't know are related to you, perhaps as closely as a 2nd or even 1st cousin (yikes). You might also have known family who are also 23 and me users, of course.
See where we're going with this?
A simple search for nonpaternity on PubMed pulls up some interesting studies.
Wolf et al 2012 used human leukocyte antigen (HLA) typing to estimate a nonpaternity rate of 0.94% in a sample of 971 German children.
Voracek et al, 2008 conducted a meta-analysis of 32 published samples from 1932-1993 and identified a decline in nonpaternity over time. The mean and median rate was 3.1%, btw.
Li and Lao, 2008 and Lucast 2007 discuss the ethics of discovering nonpaternity during medical screening and research studies. These types of works touch most directly on the Dear Prudie issue, I reckon. The 23 and me era presents interesting new possibilities. There is no longer any professional "genetic counselor" intermediary or paternalistic physician deciding whether to "tell". People might find each other more or less automatically and contact each other through the 23 and me website. But, and here's the rub, the information propagates. Let's assume there is a mother who knows she had an affair that produced the kid or a father who impregnated someone unknown to his current family. Along comes the 23 and me contact to their child? Grandchild? Niece or nephew? Brother or sister? And some stranger asks them, gee, do you have a relative with these approximate racial characteristics, of approximately such and such age, who was in City or State circa 19blahdeblah? And then this person blast emails their family about it? or posts it on Facebook?
Williams 2005 shows how paternity can affect major societal knowledge such as the infamous Jefferson/Hemings case.
Cerda-Flores et al 1999 found a 8.1% interloper rate but discuss the issue of being able to exclude fathers based on blood group systems. The very earliest literature on PubMed clearly focused on these issues, i.e., how to use the available imprecise markers to come up with statistical probabilities. The new era of personal genetic identification presumably improves substantially on these issues.
There are more articles and you can scan them for yourself. But I'll end on two key notes. First, clearly this is a nightmare issue for people who are looking into rare genetic variants that are related to disease because the familial patterning is so important. It is critical to know who the father really was to keep the dataset valid and clean. Second, there seems to be a rumor of 10% nonpaternity that floats about in the literature, perhaps as a lasting straw argument or from some early flawed dataset. What the more recent studies seem to conclude is that this is an overestimate.
Or, you know, you could just read Dear Prudie's response which includes:
But I don't think your late husband's parents need to hear this. ... And I don't see any reason to deprive your child of a potential inheritance.