NON-PATERNITY EVENTS
Science July/August 2007
The unintended consequences of genetic screening for disease
by Steve Olson
Who's Your Daddy?
A few months ago, I sat down at my desk to open a letter that could tell me whether my father was really my father. In fact, that letter could tell me whether the men going back 10 generations on my paternal side were the biological fathers of their children.
I wasn't caught up in some bizarre multigenerational paternity suit. A scientific officer at a genetic testing company knew that I was interested in genealogy, and he had offered to run my DNA through a sequencer. A few weeks earlier, I'd swished mouthwash inside my cheeks, sealed the mouthwash in a tube, and mailed the tube to the company.
My doughty Scandinavian ancestors passed the test. My DNA revealed no obvious instances where the man named on a birth certificate differed from the man who was my biological ancestor. But I was lucky. Many efforts to trace male ancestry using DNA terminate at what geneticists delicately call a "non-paternity event." According to Bennett Greenspan, whose company, Family Tree DNA, sponsors projects that attempt to link different families to common ancestors, "Any project that has more than 20 or 30 people in it is likely to have an oops in it."
The law of unintended consequences is about to catch up with the genetic-testing industry. Geneticists and physicians would like us all to have our DNA sequenced. That way we'll know about our genetic flaws, and this knowledge could let us take steps to prevent future health problems. But genetic tests can also identify the individuals from whom we got our DNA. Widespread genetic testing could reveal many uncomfortable details about what went on in our parents' and grandparents' bedrooms.
The problem would not loom so large if non-paternity were rare. But it isn't. When geneticists do large-scale studies of populations, they sometimes can't help but learn about the paternity of the research subjects. They rarely publish their findings, but the numbers are common knowledge within the genetics community. In graduate school, genetics students typically are taught that 5 to 15 percent of the men on birth certificates are not the biological fathers of their children. In other words, as many as one of every seven men who proudly carry their newborn children out of a hospital could be a cuckold.
Non-paternity rates appear to be substantially lower in some populations. The Sorenson Molecular Genealogy Foundation, which is based in Salt Lake City, now has a genetic and genealogical database covering almost 100,000 volunteers, with an overrepresentation of people interested in genealogy. The non-paternity rate for a representative sample of its father-son pairs is less than 2 percent. But other reputed non-paternity rates are higher than the canonical numbers. One unpublished study of blood groups in a town in southeastern England indicated that 30 percent of the town's husbands could not have been the biological fathers of their children.
Even with a low non-paternity rate, the odds increase with each successive generation. Given an average non-paternity rate of 5 percent, the chance of such an event occurring over 10 generations exceeds 40 percent.
Most people can't look that far back on their family trees, but I can. Someone on the Olson side of my family once spent an inordinate amount of time tracing the family's male lineage. My relative's genealogical research indicated that my father's father's father's father's father's father's father's father's father's father migrated from Finland to Norway in the middle of the 17th century. If that is the case, I have a particular connection to that man.
Men pass most of their Y chromosomes down to their sons intact and unadulterated. I therefore have the same Y chromosome as my father, and his father, and so on. (In fact, all men living today have inherited the Y chromosome of a single man who lived about 50,000 years ago, probably in eastern Africa. But mutations have slowly changed the Y chromosome over many generations, which is why the Y chromosomes of Finns generally differ from those of Greeks. Nevertheless, over the course of 10 or even 100 generations, the changes typically are small and the heritage is clear.) The continuity of the Y chromosome is how we know that Thomas Jefferson almost certainly had children with his slave Sally Hemings: Her direct male descendants have the same Y chromosome as Jefferson's paternal uncle, who presumably had the same Y chromosome as Jefferson. (Similar tests can reveal whether sons and daughters are really descended from their mothers and grandmothers, though non-maternity is much rarer than non-paternity.)
My Y chromosome turned out to be as Finnish as sautéed reindeer-I al- most certainly inherited it from that 17th-century Finnish émigré. But even if my Y chromosome had turned out to be suspiciously un-Finnish, I probably could have come up with a story to protect my legitimacy. I could have said that my Finnish ancestor was the descendant of a Mongolian invader, or the son of a trader from Istanbul, or even a Spanish diplomat fallen on hard times (though in fact I know that he was a peasant farmer). I could have said that one of the men in my paternal lineage was adopted after his mother and father died. The imagination is a wonderful balm for bruised expectations.
But genetic tests don't lie, which means that our imaginations may be in for a workout. For example, groups of people in many parts of the world trace their lineage to particularly prominent male ancestors. In some cases, genetic tests reveal a kernel of truth behind these stories. Genghis Khan's Y chromosome really is widely distributed in Asia, for instance. Still, many of these stories have social rather than genealogical roots. "Many times we romanticize about the different groups that we have ancestry with," says Rick Kittles, a geneticist at the University of Chicago who founded the company African Ancestry. When Kittles has told clients that their genetic tests don't coincide with what they believe, a few, he says, have been shattered.
Frankly, I hadn't thought much about these issues before sitting down to open that letter from the genetic testing company. If I had, I doubt I would have agreed to the test. If my Y chromosome was not what I expected, would I tell other family members about it-including my teenaged son? Would I have been tempted to encourage my brother, then my male cousins through my father's brothers, then my male second cousins through my grandfather's brothers, and so on to be tested so that I could determine where the non-paternity occurred? I think we'd all have been better off assuming the best and shunning the test.
But the pressure to undergo genetic testing is about to increase. New technologies are reducing the cost of sequencing DNA. Researchers are now establishing extensive databases of DNA sequences combined with health information so they can link specific genes to diseases. And once the contributions of our genes to common diseases are discovered, everyone could benefit from DNA testing. Already, the Personal Genome Project at Harvard University is seeking volunteers who are willing to have their DNA sequences and medical information posted on the Web for biomedical purposes, even though the project warns that a person's DNA could be used to "infer paternity or other features of the volunteer's genealogy."
Two of the men most responsible for the sequencing of the human genome-James Watson and Craig Venter-are making most of their genomes available on the Web. But if their sons ever decide to have their DNA tested, they could face the same situation I did in opening that letter. Watson has kept part of his genome private because he doesn't want his sons and the public to know whether he has a genetic variant predisposing him to Alzheimer's disease; he seems unconcerned about what the rest might reveal.
Genetic counselors have been struggling with the issue of non-paternity for years. When a child is born with a genetic disorder, the parents may go to a counselor to learn whether they should try to have more children. If tests reveal that the presumed father of the child is not the biological father, most counselors will tell only the mother. But a vocal minority insists that paternity should be known to all.
So far, the expense of these tests has limited their use to cases like the one above, where a serious genetic disorder is already apparent. But what will happen when people begin sequencing large parts of their DNA routinely, to see whether they are vulnerable to specific diseases? If you discovered a predisposition to heart attacks or prostate cancer, and medications could reduce your vulnerability, wouldn't you want to tell your siblings and cousins? And shouldn't they be tested, too? Yet in the absence of stringent and possibly unattainable privacy protections, widespread testing will lead to many unpleasant surprises.
Geneticists have only begun to think about how to protect people from knowing themselves too well. But they probably should have seen this problem coming a long time ago. An oft-quoted definition of their field is: "Genetics explains why you look like your father-and if you don't, why you should."
The URL for this page is http://www.theatlantic.com/doc/200707/paternity
Non-Paternal Event (NPE)
This page was original titled False Paternal Event
15 December 2006
Non-paternal event, non-paternity event, false paternal event, false paternity, misattributed paternity: all these terms refer to a break in the Y chromosome line due to a formal or informal adoption, name change, "extramarital event" (infidelity), child known by other surname (mother's maiden name, stepfather's name), etc. There's no agreement about the rate at which these occur (the 5%-10% frequently reported as being used by geneticists may be based on folklore). Using a conservative estimate of 2%, in a group of 50 participants, we might expect that at least one would not match at all.
There is always a possibility that you could get disappointing test results. Samples that vary by three or more markers from the main group may do so for a number of reasons. One possibility is that they represent distinct lines either older or younger than the currently observed most frequent line. Another is that there has been a non-paternal event at an unknown time in the past. This means the male tested may be carrying the surname but his Y chromosome does not appear to be associated with that surname.
Types of non-paternal events include but are not limited to:
Pregnancy outside of a marriage
Pregnant female married man who was not father of child
Adoption (formal or informal)
Man married pregnant widow
Children known by step-father's name
Man took wife's name and/or children given the wife's surname
Man changed name - various reasons
Aliases
Illegitimacy - child given mother's surname
Clerical error in recording administrative data such as assigning a name to the wrong person
It should be stressed that adoptions were quite common in every age: parents died by disease or war and a relative took in the children and raised them with their name, daughters had children out of wedlock and the grandparents (or other relatives) raised the children as their own. A teen-age girl who gets pregnant by one boy and marries another - for whatever reason - might be a more frequent occurrence than maternal infidelity in earlier generations. Taking into consideration the strong pressure against "unwed mothers" until the last generation or so, one might expect such cases to account for some of the paternal irregularity indicated by Y-chromosome testing. Very young mothers of first-child sons in the line could be indicators for a higher probability of this phenomenon.
A result indicating a non-paternal event would certainly be a disappointment to most participants. But your name is legally your name and a small sample size could be misleading. A DNA sequence suggesting a non-paternal event could be that of the original bloodline - e.g., 20 people are tested, 19 are very similar and yours is clearly different. It could be that the 19 descend from the same person 300 years ago who was adopted while your line links to the original blood line going back 800 years.
FAULTY RESEARCH
Faulty research can result in a non-match or a match to a different line than expected. For example, suppose a researcher traces an ancestor named John R. Kinney to an 1850 census in MO. The census indicates he was born about 1819 in MA. The researcher finds a publication about an MA Kinney line and looks up all the Johns in the index. He finds a John Robert Kinney, Jr., born in 1819, listed as a son of John Robert Kinney, Sr., who was born in MA. There is no additional information about JRK, Jr. The publication traces the family back to 1635 in England. The inexperienced researcher thinks this must be his John and links the line to the family in the publication. Later, a DNA test of the researcher does not match two other participants from the 1635 line. This could be due to a non-paternal event but it also could be due to faulty research. If the DNA project has many participants, our researcher may learn that he links to a different Kinney line (error was due to faulty research). If it is a large project and he has no matches, he will have to consider the possibility of a non-paternal event.
_____________
Some of the above was adapted (with permission) from the Blair surname project web site. The site also includes helpful DNA 101 and FAQ sections http://blairgenealogy.com/dna/
See also:
Subject: Non-paternity rate [was Re: [DNA] Father-son Study by Univ of AZ] - December 07, 2004
http://archiver.rootsweb....GY-DNA/2004-12/1102428402
The Abstract Factory
http://abstractfactory.bl...-is-rate-of-mistaken.html
The above refers to "how many bastards are there, anyway?" at:
http://unauthorised.org/a...ogy/august-1996/0125.html
Who's Your Daddy?
Short report on: "Who's Your Daddy?" "Measuring Paternal Discrepancy and Its Public Health Consequences." Mark A. Bellis et al. in Journal of Epidemiology and Community Health, Vol. 59, No. 9, pages 749-754; September 2005.
http://www.discover.com/i...ov-05/rd/whos-your-daddy/
December 19, 2005 World Science
Aug. 12, 2005 Courtesy BMJ Specialty Journals and World Science staff
http://www.world-science....ernews/050812_dadsfrm.htm
< . . . . "misattributed paternity." That's the phrase I employ from time to time now, as it covers both NPE and glitches in the paper trail. >
Ann Turner, November 13, 2006, post to ISOGG at yahoogroups.com, Subject:Re: [ISOGG] FW: Ysearch at the Conference.
http://freepages.genealog...opp/DNA/falsepaternal.htm
One in 25 dads could unknowingly be raising another man's child, researchers find
Aug. 12, 2005
Courtesy BMJ Specialty Journals
and World Science staff
Around one in 25 dads could unknowingly be raising another man's child, new research suggests. The study is published in the September issue of the Journal of Epidemiology and Community Health.
http://www.world-science....ernews/050812_dadsfrm.htm
At this point in the narrative we need to introduce the factor to which geneticists politely refer as 'non-paternity' - the term used when a child's father, the name on the birth certificate, is not the biological father. When a son bears the surname of his father but does not carry his genes there are only a few explanations available. The most straightforward, and innocent, is that the son has been adopted and taken the surname of his adoptive father. Of course, the same happens to adopted girls, but they will most likely not transmit this name to their children and they will certainly not pass on a Y-chromosome either. Y-chromosomes are only ever passed between father and son. Women just don't have them. The second explanation is that the entire family adopts a new surname. This was not a common practice in medieval England but it certainly was in Scotland, where a man often took the name of the clan chief on whose lands he lived or in whose army he fought without being related to him. That leaves us with the third and final explanation for the discordance between surname and Y-chromosome - infidelity by, or possibly rape of, the woman. Biologists have a rather more brutal name for it - extra-pair copulation. If a woman has a child with a man other than her husband and if that child is brought up within the family and is given the family name, the link between name and genes is broken. If the child is a boy, he will inherit his father's surname but not his Y-chromosome. That will have come from his mother's lover, or from her assailant, and not from her husband. When he has sons of his own, it will be this man's Y-chromosome that is passed on. Even if there are no non-paternity events in later generations, the link between the Y-chromosome and the original surname cannot be rescued. It is severed for good.
Bryan Sykes
http://www.wnyc.org/books/29529
Extensive list of studies through the years
http://www.childsupportan...sattributed_paternity.htm
