An Analysis of Fourth Cousins and Other Near Distant Relationships

· DNA
Authors

dnastudies

THE BACKGROUND

Since 2010, the Owston/Ouston One Name Study has conducted a two pronged DNA analysis by using Y-DNA and Autosomal DNA (atDNA). Culled from the Y chromosome, Y-DNA is passed from father to son and mutates very slowly. Autosomal DNA (atDNA) is derived from the 22 pairs of human autosomes and has been transmitted by all of our recent ancestors who received their DNA from their recent ancestors.

At this present writing, The Owston/Ouston DNA project has 40 participants – some have tested only Y-DNA, some only autosomal DNA, and some both. Of our 22 autosomal participants, 17 match multiple people in the study, two have non-paternal events (NPEs) in their ancestry, and three are related so distantly they share no significant (5cM+) matching atDNA; however, they are confirmed matches through Y-DNA.

All 17 matching participants share a single ancestral couple: William Owston (1778-1857) and Frances Wilson (1782-1852) – the founders of the Cobourg line of the Owston/Ouston clan. Not counting those with NPEs and those who haven’t tested, the relational makeup of this cohort is enumerated in the chart below.

relationships_matching_participantsThere are 136 relationships at all levels, but only 88 match others in the study and everyone matches multiple participants. For privacy, contributors are identified with the Cobourg name along with an alphanumeric designation. Men bearing the Owston surname have a numeric designation, such as Cobourg06 or C06 for short. For related females and males who do not bear the surname, a hyphen and an alphabetical alias is used. This designation appears as Cobourg-D, which is abbreviated as C-D.

PROJECT PARAMETERS

As a separate project until 2015, the autosomal study originally concentrated on the descendants of William Owston (1778-1857) and Frances Wilson (1782-1852). The couple produced eight known children – all of whom attained adulthood and produced issue. Of these eight, it appears that lineal descent from two of the daughters, Mary Ann Margaret Owston Smith (1819-1909) and Euphemia Owston Smith (1824-1892), has ceased. The issue that remains descends from all five sons and one daughter, Frances Janet Owston Williamson Sutherland (1815-1902).

Although Frances Sutherland had a considerable number of descendants, we have been remiss in securing any participants from this family at the present. As for the current matching autosomal participants, they represent all five of William and Frances Owston’s sons: Thomas (1804-1874); William, Jr. (1807-1892), James Wilson (1809-1858), Charles Paget Herbert (1817-1858), and John Gillon (1826-1901).

treeAs the chart above depicts, participants from John Gillon Owston’s progeny are overrepresented. Being the author’s line, a convenience sample produced more participants in this line of descent than from any of the others. The majority of close relative matches came from John Gillon Owston and include four out of five parental relationships, four out of five sibling relationships, all eight avuncular relationships, all five first cousin relationships, seven out of eight second cousin relationships, and eight out of nine second cousin, once removed relationships.

NEAR DISTANT COUSINS

I use the term “near distant cousins” to describe those in the range of third to fifth cousins – a range of relationships where autosomal DNA is typically shared by those more distant than close relatives. Third cousins share second great grandparents and on average match about 0.78% of our DNA. With closer relationships, we have a 99% or greater chance of matching our cousins; however, the probability of matching a documented third cousin drops to around 90% (“The Probability,” 2015; “What is the Probability,” 2015).

As we move up the consanguinity spectrum, the chances lessen for fourth cousins (45%-50%) and with fifth cousins (10%-15%) (“The Probability,” 2015; “What is the Probability,” 2015). Theoretically, we share on average nearly 0.20% of our DNA with a fourth cousin and nearly 0.05% with fifth cousins (“Autosomal DNA Statistics,” 2015). As the old automobile disclaimer use to state, “Your mileage may vary”; the same applies with sharing atDNA – as your matching will vary.

As an example of this variation, our study has a fourth cousin pair who shares more DNA (0.96%) than typical third cousins; however, there are also 14 fourth cousin pairs that share no measurable DNA. In another example, one person shares more with a fourth cousin, once removed (0.51%) than she does with a second cousin, once removed (0.34%). While in a perfect world, the numbers would be predictable; however, this is not always the case with atDNA; and after we leave the realm of close relatives, anything is reasonably possible.

Once we get beyond six generations, atDNA loses significance as a relationship predictor; however, there are cases where two (or more) distant cousins will share a modicum of DNA. This may be one of the many reasons individuals have no idea how they are related to the hundreds of cousins listed in autosomal DNA databases – as most people cannot verify every lineage past their second great-grandparents. Other factors that contribute to this include non-paternal events, adoptions, name changes, and the lack of records needed to confirm relationships.

While our autosomal project has studied a variety of relationships, we have concentrated on fourth cousin matches. For the purpose of comparison, we will refer to those individuals as Gen 5 relatives as they are five generations from our common ancestor. While we have 13 in the project at that level, only 11 match others in the project. In addition, we also have one Gen 4 person and five Gen 6 individuals.

GENERATION FOUR

Cobourg-J, a great-granddaughter of William Owston, Jr., is our sole Gen 4 participant. The numbers of potential living Gen 4 descendants of William and Frances Owston, however, are limited and probably number about seven in total. Cobourg-J, our 40th project participant, was tested this summer and she matched all of the Gen 5 participants – one of whom is her daughter (Cobourg-A). Her matches of shared DNA with Gen 5 participants are listed below in centimorgans. With exception of her daughter, these relationships are all at the third cousin, once removed level.

Gen4-5-Matrix A typical third cousin, once removed will share in the neighborhood of 26.56cM (“Autosomal DNA Statistics,” 2015). With two relationships (Cobourg08 and Cobourg-H), Cobourg-J shared at or above triple the amount of DNA that one would have expected. The amount of matching atDNA is consistent with that of second cousins, once removed. Two relationships with Cobourg-J (Cobourg06 and Cobourg-E), shared a third of the predicted average for the relationship and are in the typical and predicted range of fourth cousins, once removed. The average amount of DNA that Cobourg-J shared with these ten participants is calculated at 43.2cM and is skewed due to her considerably higher than average shares with many of the participants.

Gen4-6-MatrixAs expected, Cobourg-J’s matches with Gen 6 participants are fewer in number and lesser in the amounts shared. While John G. Owston’s Gen 6 descendants are overrepresented in project, it is interesting when you compare Cobourg-J’s sharing. While Cobourg-J shares 94cM with Cobourg08, she doesn’t share any DNA with his second cousin, once removed – Cobourg-D. Additionally, she shares a fairly large amount with Cobourg02, but shares no atDNA with his son, Cobourg07. Ironically, while she shares slightly less than average (19cM) with Cobourg03, a portion of this DNA (14cM) transferred to his son, Cobourg09.

GENERATION FIVE

With this study, the Gen 5 relationships are better represented than any other group with a total of 13 participants. Discounting those with non-paternal events, we have 11 participants that match others in the project. In addition to having 43 fourth cousin relationship pairs, the number also includes four sibling pairs and nine second cousin pairs. The matching segments between participants are illustrated in the diagram below with blue lines indicating fourth cousin matches. The sibling matches between Gen 5 participants are in red and the second cousin relationships are in green. fourthcousinsWhile there is a 45-50% probability of a genetic match between fourth cousins, 67.4% of our fourth cousin pairs match with a minimum of 5cM (“The Probability,” 2015; “What is the Probability,” 2015). The average shares among fourth cousins is 13.28cM, the matrix below shows some interesting results among Gen 5 participants (“Autosomal DNA Statistics,” 2015). The results are very typical for several of our participants. For example Cobourg01, Cobourg03, and Cobourg06 share atDNA at 5cM and above with 50% of their tested fourth cousins with fairly typical sharing amounts. Of the three, only Cobourg01 has an atypical share at that is with Cobourg-A at 42cM.

Gen5-MatrixSome interesting patterns have developed with three participants. Cobourg-A, our fourth participant in the study and daughter of Cobourg-J, had three shares at or above triple the expected amount of shared DNA – she also shared with 70% of her fourth cousins who were tested. Cobourg-H, who unfortunately died this year, is unique is that she shared DNA with every single person (sans NPE participants) in the study who descend from William and Frances Owston. While some of her matches are higher than normal, they are not as high as others. It is interesting to note that the second highest share among our tested fourth cousins occurs between Cobourg-A and Cobourg-H at 58cM.

largest_matchesAnother fairly new participant, Cobourg08, matched 70% of his fourth cousins in the project – some with very high amounts. While two matches (Cobourg-A and Cobourg-E) were three times greater than the average fourth cousin share, his match with Cobourg-G at 65cM was five times greater than the fourth cousin average and at a level above the average for third cousins. Without the participation of Cobourg-A, Cobourg-H, and Cobourg08 in our projects, the results would be very typical at best.

Gen5-6-MatrixSince we only have five Gen 6 participants and several of these are closely related to Gen 5 participants, generalizations are difficult to make. Four out of the five participants are descended from John Gillon Owston. Of these, Cobourg07 is the son of Cobourg02, Cobourg09 is the son of Cobourg03, and Cobourg-B and Cobourg-C are daughters of Cobourg01. Their close connections are manifested with several avuncular and second cousin, once removed relationships among other Gen 5 participants.

Only Cobourg-D is from a different family group and, as a descendant of James Wilson Owston, she is a second cousin, once removed to Cobourg08. As expected, there are very few fourth cousin, once removed shares at a level greater than 5cM. The average fourth cousin, once removed share is calculated at 6.64cM (“Autosomal DNA Statistics,” 2015).

TRACKING GENERATION 5 SHARES AMONG CLOSE RELATIVES

Parental

In analyzing the amount shared with Gen 5 participants, a mother/daughter comparison is very similar, as expected. As atDNA theoretically halves in every generation, it was expected that Cobourg-A would have smaller shares and non-shares with other Gen 5 participants than her Gen 4 mother (Cobourg-J). Where her mother has lower sharing amounts, several of these matches are non-existent with Cobourg-A, as she shares no atDNA with Cobourg06, Cobourg-E, and Cobourg-I. The only exception is where she shares a slightly lower amount than her mother with Cobourg03.

mother-daughter-matchesSiblings

In looking at our first sibling group (Cobourg01, Cobourg02, and Cobourg03), we see very symmetrical results among these brothers. Where there are generally lower shares with one participant, the others follow suit with lower or non-existent shares. Although lower in number, Cobourg03’s match with Cobourg-A is his highest fourth cousin match – as it is with his two brothers.

Sibling01-4thcousinA second sibling group that comprises a brother and sister, however, shows radically different match patterns with fourth cousins. With the exception of a similar match to Cobourg02, there is no symmetry among the two siblings’ matches.

Sibling02-4thcousinSecond Cousins

In the chart below, brothers (Cobourg01, Cobourg02, and Cobourg03) are compared with their second cousins, Cobourg06 and Cobourg-H. Cobourg06 and Cobourg-H are second cousins to each other as well. While some symmetry occurs with Cobourg-H to the brothers, none is present with Cobourg06 to either Cobourg-H or the three brothers. This is interesting as two of the brothers share higher (391cM and 333cM) than the average amount (212.5cM) of DNA with Cobourg06 (“Autosomal DNA Statistics,” 2015). The only explanation might be that their higher than normal shares are through their shared non-Owston/Wilson ancestries.

2ndC01-4thcousin

Another second cousin pair, Cobourg-E and Cobourg-I are interesting as they share an extremely low amount of atDNA at 70cM, while the average is 212.5cM (“Autosomal DNA Statistics,” 2015). With exception of their nearly identical share with Cobourg-G and similar shares with Cobourg-H, their lack of symmetry, as with the second sibling group, emphasizes the randomness of DNA transmission that even occurs within the same family group.

2ndC02-4thcousin

TRIANGULATION

The idea of triangulation was derived from calculating distances and/or sizes of objects by comparing two points at a fixed baseline to a third point. Calculation occurs by using with the length of one line and two known angles to determine third angle of the triangle and the corresponding distances. In research, the term is used to compare results from two or more methods to determine the accuracy of the overall results.

When applied to genetic genealogy, Dr. Ann Turner (2015) stated that “Triangulation (identifying clusters of three or more people who all share the same DNA segment with each other) is a rigorous technique for generating pointers to a common ancestor. But the very requirements that make it so robust also restrict its scope” (para 1). In attempting to triangulate results, we analyzed shared segments of 12 project members.

While we concentrated on fourth cousins in the autosomal portion of the Owston/Ouston DNA project, we have expanded our reach to include one Gen 4 person (Cobourg-J) and eliminated her daughter (Coboug-A) so as to avoid natural overlap. We also added one Gen 6 participant, Cobourg-D, as she shares a significant amount of atDNA with Cobourg-E and Cobourg-H. Other Gen 6 participants were not included as their fathers were represented.

To provide purposeful triangulation, each one of the three individuals had to be descended from a unique Owston brother – therefore matching segments between participants who were second cousins, once removed and closer were eliminated. Therefore, this experiment included 66 relationships from third cousin, once removed to fourth cousin, once removed. There are a total of 60 segments at 5cM or greater that are shared by two or more these individuals in the project. With descendants from five brothers, there are ten possible tripartite combinations.

While there are numerous shared segments between two individuals, there was great difficulty in finding segments that were shared by three or more individuals at 5cM. Of the 60 segments, only six at 5cM or greater were shared among a group of three individuals. This represented triangulated relationships among six tripartite combinations. Unlike what we’ve seen in other studies with triangulated segments of the exact same length for all three individuals, none of the matching segments fit this mold. We’ve seen several combinations that show partial shares of a large segment with one individual and an abbreviated share with another. abbreviatedIn addition, the start and stop points were not the same for the two segments and many showed an overlap that indicated that two of the individuals received portions of what was originally a longer segment. overlapAlthough not considered as one of the triangulated segments, the following illustration of adjacent segments demonstrates some of what was observed when comparing matches. Again, this is probably an example of where a larger segment was split in the recombination processes for two of the matching individuals. AdjacentAlthough the topic of the veracity of identical by descent (IBD) segments less than 5cM has been debated ad nauseam, we’ve decided to drop the triangulation threshold to 3cM. In each case, the subjects matched each other at a minimum of 5cM – showing a legitimate relationship. At 3cM, five additional triangulated segments were identified. These segments represented three of the tripartite combinations that were found in the 5cM matches and a previously unrepresented combination. The following diagram shows the triangles of matches between participants. In one case, four individuals matched – two of whom are brothers and their match to each other was not considered (see the yellow match). triangulation Three of the tripartite combinations were not represented in the triangulation process. They are the descendants of the following brothers:

  1. Thomas Owston; William Owston, Jr.; and Charles P.H. Owston
  2. Thomas Owston; James W. Owston; and Charles P.H. Owston
  3. James W. Owston; Charles P.H. Owston; and John G. Owston

This non-representation is probably due to the limited numbers of participants in these families. Adding others at the near distant level may provide triangulated segments for these relationships.

An examination of the above diagram indicates that three individuals were significant in the triangulation process of the 12 individuals: Cobourg08, Cobourg-H, and Cobourg-J, the mother of Cobourg-A. While the triangulation process when extended to 3cM provided 11 segments shared by at least three individuals from unique progenitors, it does not seem to be solely effective in determining relationships.

For example, Cobourg06 has no triangulated segments. Because he matches Cobourg-E, Cobourg-G, Cobourg-I, and Cobourg-J, as well as having four second cousins who have triangulated segments, it can be argued that he has a solid genetic connection to others in this grouping. Since the majority of 29 matching fourth cousins only share one segment of atDNA, triangulation for some participants may be an exercise in futility and other methods might need to be employed.

In his recent blog post regarding “Genetic Networks,” Blaine Bettinger (2015) stated the following: “Although Triangulation is the gold standard, I’m not convinced that triangulation alone should be utilized for identifying IBD segments, or that triangulation alone should be utilized to assign segments to an ancestor or ancestral couple” [his emphasis] (para 10). While some may consider this radical, I agree. Dr. Bettinger’s comments are food for thought and may spawn further discussion in the genetic genealogy community.

MEIOTIC EVENTS

Finally, I wanted to see if meiotic events had an influence on the sharing among the 43 fourth cousin pairs. Chowdhury, Bois, Feingold, Sherman, and Cheung (2009) confirmed that females are more likely produce meiotic recombination than males. According to Chowdhury et al. (2009), female to male recombination events occur at a ratio of 1.6 to 1. While recombination could produce smaller segments to be passed on to descendants, fewer recombination events in males may be represented by several all or nothing transmissions of DNA that was derived from a specific ancestral couple.

While the number (43) of our project’s fourth cousin relationships is relatively small, it is impossible to make accurate generalizations from the data, but it is worth considering. To begin this process, we’ve counted the number of female meiotic events that occurred with the 43 paired relationships. Since we had only one share with four female meiotic events and only three with three female meiotic events, we only compared those with 0, 1, and 2 female meiotic events – a total of 39 relationships.

Although outliers occurred with 0 and 1 female meiotic events, the clustering of the majority of relationships can be seen in the following chart. Although the largest fourth cousin share (65cM) occurred with no female meiosis events in the direct lineage from the common ancestral couple, the next highest share was only 11cM. Likewise, one female meiotic event produced a share at 46cM; however, the next highest share was 26cM.

matches-female-transitionsWhen more female meiotic events are represented, the spread of results (sans outliers) expands. Zero female meiotic events produced the lowest cluster and the range of the clusters expanded as more female meioses were introduced in the lineage. In addition, the percent of non-matching relationships decreased with the number of female meiotic events.

At zero female meiotic events, non-matching relationships occurred at a frequency of 42.86%. With one female meiotic event, the frequency of non-matching atDNA dropped to 30%. With two female meiotic events, the percentage of non-matches further dropped to 20%. This leads us to believe that, at least in our study, more ancestral female representation produces more opportunities of matching atDNA segments among fourth cousins. In addition, those matches tended to have larger segments. The mean cM lengths were larger as the number of female meioses increased and are recorded as 0 = 8.67cM; 1 = 14.70cM; and 2 = 15.06cM. By removing the two outliers, the mean results are more dramatic as follows: 0 = 4.23cM; 1 = 11.22cM; and 2 = 15.06cM.

CONCLUSION

An analysis of near distant relatives in a surname study can add value to the overall project. We have already answered two problems that were not possible through traditional methods. Although in both cases, the genetic results confirmed existing hypotheses; these conclusions would have not been possible without the use of autosomal and Y-DNA.

The scope of fourth cousin matches in the Owston/Ouston project is a testimony of great variability one finds when studying near distant cousins. In the future, we are anticipating further testing of additional fourth cousins and hope this will add to the body of knowledge about our family and genetic genealogy in general.

REFERENCES

Autosomal DNA statistics (2015). International Society of Genetic Genealogy (ISOGG) Wiki. Retrieved from http://www.isogg.org/wiki/Autosomal_DNA_statistics

Bettinger, B. (2015, July 25). Creating DNA circles – Exploring the use of “genetic networks” in genetic genealogy. The Genetic Genealogist: Adding DNA to a Genealogist’s Toolbox. Retrieved from http://www.thegeneticgenealogist.com/2015/07/25/creating-dna-circles-exploring-the-use-of-genetic-networks-in-genetic-genealogy/

Chowdhury, R., Bois, P. R. J., Feingold, E, Sherman, S. L., & Cheung, V. G. (2009). Genetic analysis of variation in human meiotic recombination. PLOS Genetics, 5(9), 1-9. doi:10.1371/journal.pgen.1000648

The probability of detecting different types of cousins. (2015). Mountain View, CA: 23andMe. Retrieved from https://customercare.23andme.com/hc/en-us/articles/202907230-The-probability-of-detecting-different-types-of-cousins

Turner, A. (2015). The trouble with triangulation: Preliminary notes. Retrieved from https://drive.google.com/file/d/0B-wpDkkP5x0odkFRdDFaNHloOFk/view

What is the probability that my relative and I share enough DNA for Family Finder to detect? (2015). Houston, TX: Family Tree DNA. Retrieved from https://www.familytreedna.com/learn/autosomal-ancestry/universal-dna-matching/probability-relative-share-enough-dna-family-finder-detect/

3 Comments

Comments RSS
  1. Blaine T. Bettinger

    This is wonderful, what a terrific in-depth analysis! Out of curiosity, with 22 autosomal descendants (fewer, perhaps, if you ignore the children of a tested parent), how much of the ancestral couple’s genomes do you estimate you were able to recreate? Of course, the answer will depend on whether you’re including just triangulation or also including all shared segments (the Genetic Network theory). Related to this, are you aware of overlap in other lines of the test-takers’ family trees? If not, do you think you have enough information about everyone’s “non-Owston” lines to confidently use non-triangulated segments?

    Feel free not to take too much of your valuable time to answer, I just wanted to share some of the thoughts running through my head after reading this. It might also be helpful to others who might someday consider large-scale projects to recreate an ancestral couple’s DNA!

    Will you be posting this to the ISOGG Facebook group? I do hope you share it.

    • jowston

      Wow, Blaine. You got to this rather quickly. I would venture to say that we probably could capture about 200 cM of the ancestral couple – and I would be following Genetic Network theory to calculate this. I don’t know of any overlap, but that is always a possibility. These families were pretty mobile, though at one time four brothers lived in Pittsburgh – descendants lived all over. They really never communicated with each other until I began working on our history in the 1970s.

      I am confident that the descendants of Charles P.H. Owston have no overlap as there were two generations of marriages into German immigrants. One of those individuals matches both my half-cousin and my sister-in-law who are not genetically related to the Owstons. That’s a curiosity that I’d love to figure out as the shares are considerable. I am also confident that the descendants of William Owston, Jr. would have no overlap as they were the only ones that remained in Canada when everyone else moved to the US.

      Yes, I was fine tuning it as blog had several problems and I think I have it all fixed. I’ll be sharing it soon. Thanks so much for your comments.

      Jim

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: