Search billions of records on

Home F.A. Questions Clan Member Photos Ancestral Photos Dennstedt Treatise 1st Settlement Family Groups FTDNA Family Finder Observations Spreadsheet

by Patrick Childress, Project Administrator

Find anybody, any place or anything mentioned on this website
by using the Website Search Engine at the top of the Home Page.

For those viewers who really want to get into the "nuts and bolts" of how Family Groups are determined, using the extended marker Y-DNA tests, this is the page for you.  Click on this link to bring up the open source Excel spreadsheet on your computer, then use the following notes to follow the logic of defining family lines.  Please note the date of this spreadsheet as it is updated on an irregular basis.  For the conventional, up-to-date and comprehensive Y-DNA results (without the comments shown on the below spreadsheet) for all Participants, click here.

The observations and conclusions on this page and on the spreadsheet above have been reviewed for content by Elizabeth Lindsey Britton (Project Administrator of the Britton surname project, the Kevan surname project and a Co-Administrator for the Sandidge Project) and by Michelle Fiedler, Information Specialist for FamilyTreeDNA.  Our appreciation to them for their time.

Notes on Childers/Childress Spreadsheet dated 12.26.2013
“Worksheet for Group Analysis, showing only Mutual Mutations

A.   General discussion of conventions used in construction of worksheet:

I.      Initially, all mutations against the proposed ancestral Y-DNA "signature" were noted and marked with either a bright YELLOW highlight (indicating a unique mutation showing up for only one tester) or “softer” color highlight (indicating a “mutual mutation”) that exists for two or more testers in the same allele location.

4.    UNIQUE MUTATIONS are of course useful in determining how far one must go back in time to find a Most Recent Common Ancestor.  However, once that determination has been made, the UNIQUE mutation has largely served its purpose and that data is not necessarily helpful in determining probable family lines within the overall group.  For that reason, all columns that had zero mutations and/or UNIQUE mutations in the column were eliminated from this spreadsheet.

5.      MUTUAL MUTATIONS (more than one individual having a marker value that is different from the overall group value) in allele locations enable us to speculate on possible family lines within a Family Group, particularly if the mutual mutation occurs in a non-fast mutating marker.

IN SUMMARY, instead of showing all 111 marker values and columns, this spreadsheet has been condensed from 111 alleles values down to only 16 columns (some with multiple values).


No major mutation activity within the first twenty five markers results in little value being assigned to the basic kit resolutions of 12- and 25- marker kits.  While we can use these basic results to definitively place the testers  into this sub-project, no further delineation of family groups can be made; consequently, all ten of these participants with 12- or 25-marker tests are put together in Group 09, Limited Markers.

1.      The first major division of families is based on CDY a and b (first column highlighted with green).  However, as noted by several researchers, this marker is highly susceptible to mutation and may experience changes much more rapidly than others.  Consequently, we can label this mutation as “subordinate” to other allele mutations such that when there are conflicting results for a tester (kit #8287, Cody Childress, for example, whose CDY value is “35-37” which would normally exclude him from his assigned Family Group 04), the other mutual mutations prevail (in Cody’s case, his mutual mutation at DYS 576 of “15” puts him into this group).  There has been a tendency by some researchers to suggest that this primary split of the population at this marker can be traced back specifically to a mutation at this marker between the paternal historical figures of Abraham I (b. 1622 England; d. VA) son of William (b. 1599 England; d. VA) and his cousin or nephew Philemon (b. abt. 1630 VA; d. VA).  This is possible and if proven true, makes for a very convenient delineation of basic family lines.  Thus far, the three testers (kits 12758, 284830 and 19389)whose lineages have been conclusively documented back to the two patriarchs above have validated this convention.  Using CDY as the first primary division results in the creation of Family Group 01, descendents of Abraham I (CDY=35,37) and Family Group 02, descendents of Philemon (CDY=35-38).

2.      Family Group 03 was eliminated and the earlier testers were assigned to other Family Groups, based on extended testing.

3.      Family Group 04 consists of those testers whose marker value at the fast-mutating DYS576 = 15.  The eight individuals in this group likely are a subset of Family Group 02, based on their CDY=35-38.

4.      Family Group 05 consists of two testers whose mutual mutation at DYS456 and whose documented family lineage coincides with one another and commences in the Spartanburg, SC area about the same time.  Common given names exist in both families and instances of mutual movement from the East Coast toward the South suggest a strong genealogical connection.  This Family Group likely is a subset of Family Group 01, based on their CDY value of 35-37.

5.      Family Groups 06, 07 and 08 each consist of testers whose results placed them in this subproject, but suggests that a common ancestor lived between 7 and 15 generations ago, assuming that the probability of such exceeds the 50% level (using the 67-marker test results).  While there is some small hope of finding the elusive common thread for the Jackson and St. Clair family to the Childers/Childress line, it’s likely that records validating the relationship between the Creed family and the Childers/Childress line are lost to time (between 300 and 400 years ago).

C.  CONFLICTING MARKER VALUES – It might be helpful to examine some of the most remote relationships between the “Speculative” ancestral signature and contemporary values in order to estimate when some of these marker values may have mutated.  Specifically, refer to DYS504, column “U.”  This is not a fast-mutating marker and is considered relatively stable.  If we look at Family Group 07, we see that the family of St. Clair carries a value = 18.  Likewise, kits 8284 (Family Group 02), 43870 and 197794 (Family Group 01) also carry a value = 18, even though they’re in different Family Groups.  The explanation for this conflicting situation could be that “18” was the initial value of the MRCA and subsequent mutations occurred at this marker before mutations occurred at other markers that determined family groups.  This scenario could explain some of the conflicts between marker values that might place the testers in Family Groups other than those to which they currently are assigned (another example of this is DYS534 in which several testers carry the mutual mutation of “16” against the ancestral value of 17).


The spreadsheet includes extensive notes and is largely self-explanatory.  The reader is invited to submit any questions or corrections to Patrick Childress, Project Administrator, via e-mail.


Archival Observations, dating back to August 27, 2009

The following observations on the DNA spreadsheets are made after considerable consultation with Bennett Greenspan, President of FamilyTreeDNA.  Mr. Greenspan has been most patient and helpful in analyzing the DNA results of the Viking Group:

  1. The most outstanding feature of the entire spreadsheet is the "tightness" of the values of the markers.  Bennett noted that it was very unusual to have this many matches on these markers, indicating that we can be very confident that everyone on this sheet are is a descendent of the same ancestor - we're still not sure who, however.

  2. The "random" mutations on the various markers can largely be ignored - except for those on which more than one donor matches.  So, we ought to be looking specifically at three markers: 576, 570 and CDYb (all of which are shown in the column headings).

  3. We can do a general split on this Clan (assuming a 37-marker test was ordered) by using the CDYb marker, as everyone is either a "37" or a "38."  However, this is a very volatile marker and one that can be a bit "unstable."  Thus, we should group individuals who match on the other markers, in preference over this CDYb marker.

  4. In view of the above, William Childress (8284) and Benjamin Childers (8772) likely are closer kin to each other than to the rest of us.  Likewise, Paul Allan Childers (19389), Clyde O. Childress (8267), Cody Childress (8287) and Daniel Ray Childress (8269) are most probably a "twig" off the main branch.  The same may be said for Robert E. Childers (8365) and Clayton Robert Childress (8295).

  5. It's interesting to note the current location of the donors, as this relates to the migration of their ancestors.  So far, all of us "Texans" are in the same general CDYb group.

  6. The DNA testing company for this project, FamilyTreeDNA of Houston, Texas, has conducted tests and associated analyses on hundreds of thousands of Y-DNA samples.  One of the many conundrums that has arisen in these tests is the situation of different surnames with matching DNA characteristics.  Statistically speaking, this anomaly occurs in about 5% of all those males tested.  And, while matches on 12-marker DNA tests between different surnames are routinely seen, near perfect matches on the extended 25- or 37-marker tests usually point to one of four scenarios: (see Miscellany, B., above).  In the case of this Childress project there have been several matches with non-Childress surnames.  The individuals reflected on this spreadsheet have indicated a desire to explore the possible links with the Childress lineage that most likely is their ancestor.

    In the case of John and Richard Michael Jackson, the "spokesperson" for the family (Trellys Erwin) speculates that a Childress orphan was adopted by the Jackson family.  She further speculates that the best candidate for this occurrence was one Jesse Jackson, born 1825 in Madison County, Tennessee and who died in 1879 in Delta County, Texas.  Please contact Trellys direct if you have any information that might shed some light on this relationship.  It should be noted that, while there is a very close match between both Jackson males and the "ancestral signature" of the original Childres progenitor, both Jackson males reflect mutations at 464b and 464c, so a match from a Childress participant in the future on these two markers will narrow the Jackson/Childress search down in a hurry!

    In the case of Dover family, three individuals (Truman, David and Albert Dover) have taken the 25-marker test and two of the three have an exact match with the Abrah Childres on all 25 markers.  Carol Dover, spokesperson for the family, indicates that the family is realistic about these situations and their primary interest is finding and following the proper genealogical connections, wherever that may take them.  Please contact Carol direct if you have any information on this relationship.

  7. This posting date of August 21, 2005 finds an almost "dead heat" in the number of Viking participants split between Group A and Group B.  Amber Dalakas had her father Carlos Otha Saint Claire's DNA tested and the resultant match with the Abrah Childres Group A was a surprise to her family.  There is no known connection between the two surnames, so if any of the rest of the participants recognize the Saint Clare surname, please contact Amber direct.

  8. The reader will note that the results of Robert Brooks Childress (last line) is a bit different.  That's because his test was undertaken with a different testing company, DNA Heritage.  They provide a 43-marker Y-dna test that is identical to the FamilyTreeDNA test on only 31 of the markers, hence the "X" in the columns that their testing does not provide.  Unless his results are retested by FamilyTreeDNA, it seems unlikely that we'll be able to move beyond the matching on the 25-marker scale, the extra six markers notwithstanding.

  9. Kent Ross Childress' (kit #47845) results for the 37-marker test are shown on line 36 of the attached Excel worksheet.  The DNA results confirm the family's belief that their oldest known Childress ancestor, Thomas J. Childress (b. 1797) likely was the grandson of William Childress, born 1740 in Goochland, Virginia.  This confirmation is due to the fact that Kent's DNA results were identical to those of Ronald Brent Childress whose marker results are immediately above Kent's on the spreadsheet.

  10. The reader will want to closely examine the marker values for the three Creed donors and the two Jackson donors.  You'll note that both family names contain the same mutations on the 464 markers, both of which are different from the Childress/Childers Ancestral Signature.  After discussion and analysis by FamilyTreeDNA experts, we concluded that, while there are many possible explanations for this anomaly, perhaps the most likely historical event that brought about this scenario is this:

    Somewhere back in pioneer times (not sure when), a Childers/Childress family may have lost both mother and their Childress/Childers father at the same time.  As a result, the children (including at least two male siblings) were packed off to either friends or relatives in undocumented adoptions.  The two adopting family surnames were "Creed" and "Jackson."  What we'll be looking for in the future will be some Childers/Childress male whose DNA markers match the mutations shown by the Creed/Jackson donors.  We'll then be able to more closely discern the Childress/Childers line for the Creed and Jackson names.  What the contemporary Creed/Jackson researchers will now be looking for will be census records and other documents wherein their family name ancestors were in close proximity to a Childress/Childers household.

  11. The extended 67-marker results are posted on this website for the participants who signed up for the extended tests and whose results are now available on the FamilyTreeDNA Public website.  By clicking on the links at the top of this page, the reader may review the details of the new testing.  There remains a significantly tight clustering of the results, to the extent that there are no new marker mutations which give rise to various family lines.  In fact, a FamilyTreeDNA representative states:

    "Looking at your spreadsheet, ... these 67 marker results re-emphasize the closeness of the group as a whole.  It is very clear that these different Childress-Childers lines are all very closely related.  Even people who had 3 differences at the 37 marker level are identical on the 38-67, meaning they are probably closer than you thought at first...

    This is really what strikes me the most when looking at these results.  The results may also help map out some of your various Childress branches more clearly."

    Hopefully, as additional 67-marker results come in, we'll be able to better refine our assessment of the various family lines.

  12. Jerry K. Childers, Kit #68316, has his 37-marker results posted on this update.  His results place him squarely in Group A (CDYb=37), whose markers are exactly equivalent to those which our earliest common ancestor may have had.  Put another way, Jerry's marker values on the first 37 markers likely have no mutations at all when compared to our earliest common ancestor.

    A new feature added with this posting is the "Mutated DNA" Excel Spreadsheet and associated HTM Spreadsheet.  We believe that the reader will find it easier to visualize the mutations (only) in the DNA signatures of each participant, as opposed to the full listing of all DNA marker values.  In actuality, as long as the alleles value for ALL participants are equal on any specific marker, then we can effectively discard that DNA marker for purpose of attempting to align sub-branches of the family tree.  In lieu of the 67-marker spreadsheet and htm page, the reader may now view the compressed spreadsheet featuring only the 21 markers (out of 67) in which at last one DNA Participant had a mutated value.

    In summary, the reader may view the original, full Excel spreadsheet with all 67 marker values; the same full HTM spreadsheet; the new, compressed Excel spreadsheet with the only the 21 markers that have mutations; or, the same abbreviated HTM spreadsheet by clicking on the appropriate banner at the top of the home DNA Page for each of the four versions.

    When viewing the Mutations spreadsheet, note that the yellow highlight on any marker indicates that the mutation has occurred in only one single DNA Participant.  (Where this is the case, little can be discerned from that occurrence.)  On the other hand, where multiple Participants have the same mutation, the mutated marker values are highlighted in a salmon color.  These mutations are the one that should give us more ample clues as to family lineages.

  13. This posting date of November 9, 2006 reflects an additional three individuals whose extended, 67-marker results have been received.  An interesting development of a mutation at marker 534 (the 55th marker) shows up for Samuel Hall Childress, (earliest ancestor is Judge John Childress, m. Elizabeth Robertson) and William A. Childress (earliest ancestor is Robert Childress, b. circa 1758).  These two individuals would do well to compare ancestral notes.  Carlos St. Clair also shows this marker mutation.

  14. The posting date of December 7, 2006 reflects a new DNA donor and his 67-marker results.  Terry Childress (Kit #710110) is listed second on both spreadsheets.  Notice that on the "Mutations" XLS and HTM spreadsheet, Terry reflects two single step mutations from our Ancestral "Probable" Signature.  His first mutation is at 464b; this mutation is also reflected in the marker values of the Creed and Jackson entries.  However, their mutation at 464b is accompanied by an adjoining mutation at 464c, both fast mutating markers.  Because these markers oftentimes mutate at the same instant, it is likely that Terry's mutation at marker 464b did not occur in the same generation as the others, so the match of mutations between Terry and others at this marker is a "coincidence."  However, Terry also shows a mutation at marker 456, a slower mutating marker.  His mutation here is matched by this author, Joel Patrick Childress (Kit #12544) and we are the only two of the entire James River Viking Clan who reflect this mutation.  In comparing notes with Terry, we find that our oldest known ancestors (Terry's David Childress, born 1768 in Spartanburg County, SC and my William Childress, born ca. 1750 in Spartanburg County, SC) both were born in SC and lived in Mississippi at about the same (within a three county radius of one another).  Further, several "unusual" given names appear in both family trees.  Lastly, one of Terry's ancestors (Meredith Childress) was married by this author's ancestor, Justice of the Peace Obediah Childress, when both were in Jefferson County, Alabama.  We're now working on the assumption that David Childress and William Childress could well be brothers.  This is a great example where the benefits of a paper trail and the DNA research have borne fruit!

  15. December 21, 2006 reflects the addition of several new DNA Participants, as well as the housekeeping of adding several participants whose results have been available for some time.  Unfortunately, none of the four new postings have DNA results available beyond the 25-marker level, so no conclusions can be readily drawn.

  16. December 22, 2006 updates the spreadsheet with the extended 67-marker values for Samuel Graham Childers, in Group A.  Samuel and Perry Creed both have a mutated value of "21" in the fast mutating marker 570.   Are these two more closely related than the others, or did both have ancestors whose marker value at 570 mutate coincidentally with one another, perhaps at different generations?

  17. April 7, 2007 posting includes the inclusion of Ralph Loyd, an adopted descendent of Silas Childers, born circa 1857 in Arkansas.

  18. This October 2, 2007 posting once again includes the dna profile of an adopted child.  However, this time the Participant is James  Daugherty McCullough of Rockville, Maryland.  His birth father was musician Janus Hall, born 1908 in Sedalia, Missouri and his birth mother was Frances Carter Hall, born 1907 Guthrie, Oklahoma.  Jim has been unable to find any information at all on his birth parents.

    There is a considerably higher proportion of non-Childers/Childress surname matches in our project as nearly 30% of the matches have surnames other than our own.  McCullough's situation is a bit unique, however, in that even though he has a non-Childers/Childress surname, his DNA signature perfectly matches three known Childers/Childress participants; Ralph Carl Loyd (Childers), Mark Daniel Childress and James Ronald Childress.  Jim will do best by first confining his search to these four participants' family histories.