This chart from the Shared cM Project shows ranges of total shared cMs observed for known relationship levels at each of the major DNA testing companies. An interactive version is available through DNA Painter’s Shared cM Project Tool. Courtesy of Blaine T. Bettinger, thegeneticgenealogist.com, CC 4.0 Attribution License.
In previous articles, I have introduced the Shared cM Project and tools for evaluating shared DNA at DNA Painter. This column explores the resources in more detail and describes how to navigate genetically equivalent genealogical relationships. I recommend use of these tools to evaluate “known” relationships for potential half relationships and possible relationship levels with unknown genetic cousins.
Centimorgans: how to use DNA for genealogy using shared cMs
Centimorgans (cM) are units of measurement used in genetics to represent the probability that two locations on a chromosome are inherited together over the course of a single generation. A 100 cM segment indicates that, on average, there is one crossover event between the start and stop locations of the segment. Because some portions of the human genome are more likely to recombine than other portions, there is no direct conversion between the length of a segment in base pairs and the centimorgan value of the same segment.
Genealogists frequently use the total number of centimorgans that a test taker and a genetic cousin share with each other, to estimate and evaluate the generational distance between them. Closer genetic relatives share higher total centimorgan amounts in more clearly defined and unique ranges than more distant genetic cousins, where there is more ambiguity and overlap in amounts of shared DNA.
While each DNA testing company offer broad estimates on relationship levels, development of additional tools and resources has enabled more refined estimates based on total amounts of shared centimorgans.
The Shared cM Project
In early 2015, Blaine Bettinger sent out a call for submissions of known genealogical relationships and associated centimorgan values. Derived from more than six thousand relationship and centimorgan submissions, the first version of the Shared cM project was released in May 2015.1 Since the initial version, the Shared cM Project has continued to grow. The most current update to the project, released in March 2020, includes more than sixty thousand submissions from members of the genetic genealogy community.2 This version of the project also includes charts showing the distribution, range, and average of amounts of shared DNA for specific relationships. The shared cM project continues to accept submissions of known relationships and associated amounts of shared DNA, which may be incorporated in future updates of the project.
The Shared cM Project enables evaluation of amounts of shared DNA within the context of real, observed data. However, this data has some shortcomings. It is self-reported (introducing possibilities of misidentification of relationships, user error in data entry, and submission of the same relationships multiple times). Also, the dataset includes data from each of the major DNA testing companies, which may affect the analysis given these factors:
▪ AncestryDNA down weights cM totals lower than 80 cM with its Timber algorithm.
▪ 23andMe includes X-DNA in the calculations of its totals (which no other company does).
▪ FamilyTreeDNA, until recently, included very small segments (from 1 to 5 cM) in the calculations of its totals.
▪ Each company applies different algorithms for calculating centimorgan values as well as different thresholds for which segments to include in the calculations of totals.
This figure from the AncestryDNA Whitepaper reports the probability of particular relationship levels (in meioses or generational steps) based on total shared cM. The chart forms the basis for the underlying probabilities reported by DNA Painter’s Shared cM Project Tool. Courtesy of AncestryDNA.
The AncestryDNA Whitepaper
In March 2016, AncestryDNA published a whitepaper detailing its autosomal DNA matching algorithms and process.3 As part of this paper, a graph shows the probabilities of different relationship levels based on the total shared centimorgans between two individuals. Using this information, it is possible to estimate the probabilities of a relationship level, given an amount of total shared DNA.
In 2019, AncestryDNA began reporting probabilities of different relationship categories based on amounts of shared DNA with genetic cousins. This table is visible by clicking on the reported total centimorgans and percentage of shared DNA in the main match list or on individual genetic cousin profiles. “Possible DNA relationships,” sharing 201 cM, https://ancestry.com, private database.
The AncestryDNA whitepaper created a large and consistent dataset based on simulations, but it too has shortcomings. Simulations are only as good as the underlying assumptions guiding them. These simulations are most immediately applicable to AncestryDNA data (which includes the effects of the Timber algorithm) and may not be as widely applicable to calculated totals from other companies. Further, the chart on its own is difficult to interpret and determine exact probabilities of relationship based on amounts of shared DNA.
The Shared cM Project Tool at DNA Painter
In December 2016, Leak Larkin published a detailed analysis of the AncestryDNA whitepaper figure, using a plot digitizer to determine the probabilities associated with different amounts of shared DNA.4 This data formed the basis for probabilities of relationship reported at Jonny Perl’s Shared cM Project Tool at DNA Painter.5 This tool not only reports probabilities based on the AncestryDNA whitepaper, but also interfaces with data from the Shared cM Project to highlight relationships where an amount of shared DNA has actually been observed.
Conflicts existing between the two datasets are annotated and identified in the probability table. Clicking on a relationship level shows the associated chart of observed amounts of shared DNA for that relationship. Recent updates also show where a total centimorgan value falls in the chart and the percentage of observed cases below and above that value.
The DNAPainter Shared cM Project Tool reports probabilities of different genetically equivalent relationships. Probabilities in the original tool (shown here) are based on AncestryDNA’s whitepaper, figure 5.2. Probabilities in the beta tool are based on AncestryDNA’s updated 2019 relationship probabilities associated with different amounts of total shared centimorgans. Courtesy of Jonny Perl, “The Shared cM Project 4.0 tool v4,” 201 cM, https:// dnapainter.com.
DNAPainter’s Shared cM Project Tool highlights possible relationships in the Shared cM Project chart. This example shows highlighted relationships for an individual sharing 201 cM. Courtesy of Jonny Perl, “The Shared cM Project 4.0 tool v4,” https://dnapainter.com, CC 4.0 Attribution License.
In 2019, AncestryDNA began reporting probabilities of relationship for genetic cousins of test takers based on total amounts of shared DNA (weighted amounts after application of the Timber algorithm) in a pop-up information window. These probabilities presented a different distribution than the probabilities originally presented in the 2016 whitepaper. They form the basis for probabilities reported in an updated beta version of the Shared cM Project tool at DNA Painter.
DNA for Genealogy: Genetic equivalence
While the Shared cM Project treats each relationship level separately, AncestryDNA’s Whitepaper data and its updated probabilities (and by extension the probability estimates provided by DNA Painter’s original and beta tools) combine multiple relationship levels into groups of genetically equivalent relationships.
The amount of DNA that two individuals are expected to share with each other is partially dependent on the number of generational steps between them (each generation from the test taker back to the common ancestors plus each generation down to the test taker’s genetic cousin). Some relationships are genetically equivalent because they have the same number of generational steps. Thus, second cousins and first cousins twice removed are expected to share similar amounts of DNA since both levels of relationship include six generational steps.
Clicking on any box in the interactive Shared cM Project chart shows the associated chart of observed values along with indications of where the value of interest falls in the observed range, with percentage of cases above and below that observed value. This same popup also reports the standard deviation of each chart. Courtesy of Jonny Perl, “The Shared cM Project 4.0 tool v4,” https://dnapainter.com, CC 4.0 Attribution License.
Second cousins (left) share two common ancestors and are separated by six generational steps. First cousins twice removed (right) also share two common ancestors and are separated by six generational steps (two steps for one individual and four steps for the other). Despite having different generational levels, these relationships are genetically equivalent due to the fact that the total number of common ancestors and the total number of generational steps is the same. Diagram by author.
In many cases, genetic cousins are related through two common ancestors: an ancestral couple. When two individuals share only one ancestor, the amount of DNA they are expected to share in common is cut in half, which is equivalent to adding another generational step. As a result, some full relationships are genetically equivalent to half relationships with one less generational step.
Half siblings (two generational steps with one common ancestor) share about 25 percent of their DNA. An aunt and her nephew (three generational steps with two common ancestors) also share approximately 25 percent of their DNA. A grandparent and a grandchild (two generational steps with one common ancestor – the grandparent) also share approximately 25 percent of their DNA.
A first cousin (four generational steps with two common ancestors) shares about the same amount of DNA as a half-uncle and his half-nephew (three generational steps with one common ancestor). A first cousin once removed (five generational steps with two common ancestors) shares about the same amount as a half first cousin (four generational steps with one common ancestor).
Second cousins (left) share two common ancestors and are separated by six generational steps. Half first cousins once removed (right) share one common ancestor and are separated by five generational steps. Despite having different generational steps, these relationships are genetically equivalent due to the half-relationship. Diagram by author.
Consideration of shared DNA in genealogical context
When evaluating the amount of DNA shared with an unknown genetic cousin, leverage additional
information regarding the DNA match including family trees attached to the results, age, and shared genetic cousins, to determine which of several possible genetically equivalent relationships is the most likely. When evaluating the amount of DNA shared with known relatives, determine if their amount of shared DNA would be more typical of a half relationship.
Conclusion
By using observed data from the Shared cM project and simulated data from AncestryDNA’s whitepaper and later updated probabilities (available through DNA Painter’s Shared cM Project Tool), it is possible to estimate the probability of different relationship levels based on amounts of shared DNA. From there, researchers should consider additional genealogical information such as ages of a test taker in comparison to genetic cousins and family trees attached to genetic cousins’ profiles, to distinguish which of several genetically equivalent relationships is the most likely.
If you'd like help interpreting your DNA results to help you with your genealogy research, you can schedule a consultation here to work with one of our genetic genealogists.
I’m looking for any information on Jean Joseph voisin
Hi Melody! Someone on our team will be reaching out to you shortly!