I-L22-241510 Sturdevant Cluster Update #1

I-L22-241510 SturDEvant Update #1

February 20, 2017

This is an early analysis identifying important information related to the I-L22-241510 cluster of Sturdevants and to put that information into genealogical context. Three new haplotypes (persons) provide new data from which we may infer information about I-L22-241510 ancestry and internal cluster relationships. The following table provides the 37-marker data that is currently available for analysis. The most important features of this table and related inferences are discussed below.

Thirty-seven is the greatest number of markers common to all 241510 project members. At 37 markers, each haplotype has only a single mutation which is not sufficient to determine time estimates. Of the 37 markers, only five have mutated and are, therefore, meaningful. The remaining 32 markers have not mutated so they are the same value at each location and, therefore, provide little useful information. H1011 at DYS570=22 represents an exception that may be disregarded for this analysis since a genealogical record exists to prove descent from a DYS570=21 father. At greater test levels, there are likely more mutations and greater differentiation is possible. All project members are encouraged to upgrade tests to at least 67 markers (preferably 111) so that inferences and time estimates can be fine-tuned.

DNA analysis is based on statistical probabilities and the principle of parsimony (think “Occams razor” or the “KISS” principle). Simply stated: when more than one outcome is possible, the simplest outcome with the fewest assumptions is usually correct. Because of this necessity and because of our limited data, we must be careful because we cannot be certain that inferences or conclusions are correct. But we may be able to determine which are more likely than others.

The “Ancstr” haplotype  is an artificial construct used to provide temporal direction. For this analysis, “Ancstr” is an artificial haplotype that may represent an ancient and unknown I-L22-241510 MRCA. “Ancstr” establishes a direction in time so that analysis tools are able to determine relative mutation ages among haplotypes. Ancstr may be thought of as being modal although that may not be strictly true for the overall Sturdevant cluster (more data needed). In the table, the green, yellow and pink cells provide the genetic distance of each member from Ancstr, the “modal” MRCA.

New data: “544748” is a recent project member and a descendant of Justus Sturdevant of Vermont. “AnFred” is a Sturdevant reported by Ancestry.com as a genetic match to project member B4062 who is thought to be a descendant of Jonathan Sturdevant b. 1790. V2JDG is an early project member tested by DNA Heritage who did not transfer results to FTDNA. The latter two haplotypes are not available at FTDNA.

Haplotype position within the table is not relevant. I have organized the table rows to more conveniently reflect that there are two distinct lineages defined by DYS570 (see MRCA, below). Descendants of Justus are DYS570=20 and I have designated them “Lineage 1.” Descendants of John II are DYS570=21 (or 22) and I have designated them “Lineage 2.” Further, B4062 and ancestors are members of Lineage 1 because they are DYS570=20; but, they probably do not descend from Justus because of the mutation at DYS449=29.

The following lineage diagram depicts the known and inferred genealogical relationships of the I-L22-241510 cluster. It may be helpful to right-click the image and open it in another tab or window.

There are four triangulated, common ancestors; they are the MRCA of triangulated groups that are important to this analysis.

• MRCA (at the top) is the Sturdevant cluster patriarch, William’s unknown ancestor, the father of both Lineage 1 and Lineage 2, perhaps the first bearer of the Sturdevant surname. Due to the principle of parsimony, he is most probably DYS570=20 and not DYS570=21. Also probable because  H1011’s ancestral DYS570=20 (Big-Y report (members only)), inferring  that he must be an ancestor of William who remains unknown at this time. The diagram shows a single generation to William, but it may be one, two or more.
• MRCA21 is John II. He is the genealogical progenitor of H1923, 194626, H1943, H1011, and V2JDG. We know he is DYS570=21 because (1), (2), (4) and (5) are also DYS570=21.
• MRCA20b is John born 1783, a son of Justus and the known ancestor of AnFred and 544748 who are both DYS570=20. However, B4062 does NOT belong to this group because known genealogy traces to Jonathan (6) who is not directly related to (7) or (8).
• MRCA20a is an unknown common ancestor of (6), (7) and (8). He is the progenitor of Lineage 1.

Additional observations and inferences:

• Grey cells (bottom) identify the the tested members and their mutated markers.
• Yellow cells reflect the earliest possible generation that a mutation may have occurred. For example, the DYS576=17 mutation must have occurred at (2) or later. The mutation could not have occurred before (2) because (1) cannot have that mutation.
• Blue cells identify common ancestors, as above.
• Dashed lines and cells are entities that may be inferred, i.e., best guess given available information.
• Vertical and horizontal cell placement attempts to align generations across lineages.
• Cell (9) may represent a third lineage but we have no additional information at this time.

Since there are very few mutations within 37 markers, the I-L22-241510 mutation rate appears to be slow. That may be because 241510 DNA is young and relatively stable so the expected average mutation rates may not apply. At 37 markers, each of the eight tested haplotypes (except 544748) has experienced only a single marker mutation since the MRCA. The exception, 544748, has an additional mutation at DYS385a that does not match the other haplotypes. This may change when we introduce of more markers.

In reality, average STR mutation rates vary greatly and a single rate is rather unrealistic.  Here are the “latest” (Heinilia, 2012) mutation rates for our mutated markers. For a base understanding, recognize that 0.002 represents the 500 generation average mentioned above. The latter two (DYS446 and DYS712) are not used in the cladogram, below.

• DYS385     0.002645
• DYS448     0.001243
• DYS449     0.008337
• DYS576     0.011087
• DYS570     0.008929
• DYS446     0.003435 (not tested at 37 markers)
• DYS712     0.016378 (not tested at 37 markers)

We now have more available information from the descendants of Justus so we can make a couple of additional inferences. Member 544748 has a mutation at DYS385a that must have occurred at or after (8) because (7) retains the ancestral value at that marker. Similarly, AnFred has a mutation at D448=21 that must have occurred at or after (7) because (8) retains the ancestral value. Further, although AnFred was not tested at DYS446=12, we can infer that value because it is the value of (8). Because of the mutation rates, it is highly unlikely that DYS446 would be ancestral at (7). It then follows that Justus and ancestors were likely DYS448=20 and DYS446=12. And, similar logic applies to B4062 at DYS446 which may be either 12 or 13 repeats.

The following cladogram was created by Network, the Fluxus Engineering tool that tries to create a parsimonious tree (network diagram) by analyzing  related haplotypes. A cladogram is a form of phylogenetic tree where nodes (yellow, below) are member haplotypes and lines between nodes represent lineage descent from left to right (in this case). Network considers markers and mutation rates then arranges nodes in all possible combinations. Finally, it selects the most parsimonious network as the most likely. Using Ancstr as the MRCA, I have rearranged the output network so that nodes are displayed old-to-new; that is, ancient to modern, left to right. Each node is labeled with the ID of a haplotype and each link is labeled with the STR mutation. Multiple, identical nodes are combined into a single node of proportionate size. For example, the node “H1923” represents “MRCA21” ancestor of Lineage 2, described above. It also includes the haplotypes “194626” and “V2JDG” because at 37 markers they are the same (without mutation); similar to the triangulation with MRCA21 on the above lineage diagram. Also notice that the haplotypes of B4062, AnFred and 544748 originate at Ancstr which supports the inference that MRCA is older than William. Line length between nodes is a rough estimate of time. In this cladogram, the “Ancstr” node is the MRCA of all descendant nodes of this cluster, both lineage 1 and lineage 2.

The following tables provide the genetic relationship between haplotypes. Obviously, Lineage 1 and Lineage 2 are very closely related. The diagonal in each table identifies the number of markers included in the comparison.

The Genetic Distance table simply provides the total number of mutations between any two haplotypes. The TMRCA (Time to Most Recent Common Ancestor) table is more complex and provides an estimate of the 5% and 95% probabilities of finding a MRCA within a number of generations (25 years). Using the 5%-95% numbers, a 50% probability may be easily calculated. For example, B4062 and H1943 have about a 50% chance of finding their MRCA within about 18 generations: one-half of the difference between the two numbers (½(31-5)=13)) added to the smaller number (5+13=18).

In summary, after analyzing the limited 37 marker data that is available, we can deduce the following:

• Descendants of Justus Sturtevant b.1741 and descendants of John Sturdevant II b.1710 belong to different different lineages, both lineages having an older MRCA.
• Both lineages descend from an unknown father, probably two or more generations earlier than John II or Justus.
• Justus is not a descendant of either John II or his father, William.
• Haplotypes are unusually stable; 7 of 8 experiencing a single mutation.
• Not shown, the calculated age to the MRCA of both lineages is 437 years or about 15 generations (based on the 111 marker tests of 544748, H1011 and H1923) or about 5 generations before William.
• We know that John b. 1783 probably carried the mutations of Daniel P. and Alfred. Justus, the father of John b. 1783, is likely the same haplotype as John. More information needed.