|Fergus(s)on Y-DNA Project
A non-paternity event is a term in genetic genealogy used to describe the case where the biological father of a child is someone other than who it is presumed to be. It could be the result of an illicit affair, an adoption or even a surname change if one is presumes their unknown ancestor has the same surname as their earliest known ancestor. It was not uncommon in the landed families lacking a male heir to leave the estate to a son-in-law with the provisio that he change his surname to that of his father in law. Likewise it was not unusual for a brother to adopt his sister's orphaned children and give them his name. One example of a name change is that of the MacGregors who in Scotland in the 1600s were forced to change their name or be executed. It was not until 1774 that the laws against the MacGregors were repealed but not all families changed their name back to MacGregor.
The non-paternity rate is the number of non-paternal events that occur per 100 families in one generation expressed as a percentage. Results of various studied show the rate varies from 0.8% to 30% according to the population studied. Since socioeconomic factors are involved one can expect that rate to vary according to the time period as customs 200 years differ from those of today. If one looks at results of studies in the domiciles of the Ferguson, i.e. the UK and the US, the rate is on the order of 1 to 2% and very uncertain [Ref: ISOGG Wiki, 2013]
A spreadsheet calculator by Robert Tarín is available by clicking here. The results given as a percentage of whether the true ancestry agrees or disagrees with the paper-trail. Results of computations for varying non-paternity rates and different numbers of generations back to a person's earliest known ancestor . The typical Ferguson is 5 generations removed from their earliest known ancestor and the deepest pedigrees get back about 15 generations. The table below illustrates disagreement with a paper trail due to a non-paternal event could be as high as 14% for a typical pedigree and as high as 37% for an extraordinarily deep pedigree.
|Non-Paternity Rate||Generations||Agree with paper-trail||Disagree with paper-trail|
In our project about 20% of the participants do not seem to match other Ferguson, known close relatives aside, these people will be referred to as singletons. The table above indicates that 5% to 14% are perhaps the result of an NPE. The remainder is attributed to insufficient sampling in the project, in other words some participants are not matching other Ferguson because their genetic cousins have not yet joined the project. To keep it simple say 10% of the Ferguson tested fail to match others because of NPE and another 10% fail to match because their genetic cousins have yet to test.
Consider a group of related people where each person has a different earliest known ancestor aka brick wall and the most recent common ancestor aka the founder of the group is unknown. The time to the founder is at least as far back as the time to most distant of the brick walls within the group.
In our project the Haplogroup/Clade structure used to group the various Ferguson is such that a Ferguson in one Haplogroup/Clade is unrelated to a Ferguson in a different Haplogroup/Clade. It is not uncommon for two unrelated persons to claim the same ancestor. Either the genealogy of one person is incorrect or an NPE is involved.
On the other hand everyone within a group may or may not be related to one another. Look at Haplogroup R1b-L21 Clade: MacWho for an example of a group wherein every person is related to every other person. This is revealed in the table of genetic distances where the dominant highlight color is green used to denote two persons are related according to the FTDNA guidelines.
Now look at Haplogroup: I2-Y4752 Clade Isles/Sc-15. In this case one can see within the table of genetic distances two distinct squares where the dominant highlight color is green, these are called Branches I and II. Everybody in a Branch can be assumed related to one another but without further analysis is unclear if the two branches are related to one another or represent an independent genesis of the surname. In the upper left is a small square of no dominant color and it is unclear if all therein are related to each other.
Here now follow some fictitious example of conflicting genealogies constructed for the purposes of illustration. First consider the case of a John Ferguson born 1800 claimed by two individuals as their earliest known ancestor. Further assume one individual is in R-ABC and the other is in I-DEF. Within R-ABC the earliest of the brick walls is 1673 and within I-DEF it is 1709. Since both 1673 and 1709 are before 1800 there is no NPE to consider and at least one of the two conflicting genealogies is in error.
If instead the conflicting genealogies involved a John Ferguson born 1680 there is a small chance an NPE is involved in the pedigree of the person within I-DEF. There is only one generation separating 1680 and 1709. Hence for a non-paternity rate of 2% the chances are 98% at least one one of the two genealogies is in error and 2% the conflict is due to an NPE.
If the conflicting genealogies involved a John Ferguson born 1600 the probability of an NPE in R-ABC would be about 6% and in I-DEF about 8%. The probability of a NPE in at least one of the genealogies is about 13.5%. Hence the most likely reason for the conflict is that at least one of the two genealogies is in error.
Assume instead of being in R-ABC consider a singleton participant. For conflicting genealogies involving a John Ferguson born 1800 the probability of an NPE for this singleton Ferguson based on 7 generations would be about 13%. The chances are 87% at least one of the two genealogies is in error and 13% the conflict is due to an NPE. Finally for two singletons the chances are 76% and 24% respectively.
The actual numbers are considerably uncertain but not particularly important to the bottom line when it comes to trying to resolve the conflicting genealogies. A critical examination of the conflicting genealogies is relatively easy and an error in there is the most likely reason for the conflict. The logical place to start is looking for an error in the genealogies looking especially for assumptions and undocumented assertions. It is important to recognize the difference between primary, secondary and tertiary sources of documentation; see the article What Every Genealogist Should Know ...by Natalie D. Cottrill, 2000. If the conflict survives a source audit then a strategy unique to the situation is needed to look for a NPE.
If a person is a singleton as defined earlier, it may be possible to rule out an NPE as the reason and the very least reduce the probability that an NPE is reason for the failure to match. The basic method is to have as distant a cousin test as can be found.
In the family tree above if Andrew and Phillip were to both test and match then they will have ruled out an NPE between them and their most common recent ancestor Ralph. That is not a particularly good strategy as the NPE is ruled out for just one generation back. If Andrew and Gary were to test and match then they would rule out an NPE three generations back to Harry. Andrew's most distant cousin is Walter and if he were to test and match then they would have ruled out an NPE four generations back to Thomas.
Whether or not ruling out an NPE back four generations is adequate is dependent upon the problem Andrew is addressing. If for example Andrew and Gary had tested but did not match they would have concluded that one of their genealogies was in error or an NPE had occurred in one of their lines. The testing of Walter and his match to Andrew rules out the NPE in his line. If the genealogy is correct then the NPE occurred with either the birth of William or Albert.
If Andrew is supposed to share a most recent common ancestor with Samuel (not shown) say two generations earlier than Thomas then the probability that a NPE is involved is reduced from a six generation computation or about 10% to a two generation computation or about 5%.
To seriously consider a match to another surname one should test to at least 67 markers. In some haplogroups, especially R1b, tests are needed to be sure the terminal SNP is matched as well.
If you are a singleton, have tested adequately and the match list generated by your myFTDNA page for the entire FTDNA database includes other names then those may be names involved in the NPE.
If you are not a singleton, have tested adequately and the match list generated by your myFTDNA page for the entire FTDNA database includes other names then there may have been an NPE which involving that name which predates the earliest brick wall of the Fergus(s)on in your match list. The appearance of a set of names is more significant than an isolated name. An isolated name might be a singleton in some other project and the result of an NPE with a Ferguson. When a set of names is matched, so far as those names are concerned an NPE might have occurred in their line which predates the earliest brick wall in that set.
The general idea singleton or not is to look for people by the names in your match list who were in the same locale at the same time as the earliest brickwall Fergus(s)on ancestor. Look especially at female Ferguson who married men by any of those names as her children may have been orphaned and then adopted by one of her brothers. Court records and wills often deal with the guardianship of orphaned children.
Copyright © 2006 Fergus(s)on Y-DNA Project