Running successful atDNA projects in Ireland
2:30 p.m. Saturday 20 October 2018
Dodder Suite 5A, Main and Industries Halls, Royal Dublin
This is my attempt to answer Maurice Gleeson's 24-part questionnaire to panel members in the allotted 8 minutes!
My FamilyTreeDNA projects
My first public project was the Clare Roots project:
I am now administrator or co-administrator of the following
projects (member counts as of 17 October 2018):
[There are inconsistencies in the count of FF kits on the Order Summary page and the Project Statistics page
(above). The latter appears not to count any autosomal transfer kits.
The former appears not to count Y-DNA or mtDNA kits which subsequently
add an autosomal transfer. It is unclear whether each page is counting
orders submitted or results received.]
Roots (co-administator Terry Fitzgerald, WA, USA; 1153
members; 678 FF)
Y-DNA Project (administrator John Hallissey, Waterford; 69
members; 41 FF)
(administrator Fergus Clancy, Dublin; 90 members; 57 FF)
(administrator Mike Durkin, NY, USA; 28 members; 13 FF)
(administrators Cory Marinan, WI, USA; Cindy Wood, MI, USA; Greg
Marrinan, CT, USA; 33 members; 20 FF) (This project has an associated Facebook group.)
(rescued from worldfamilies.net;
102 members; 50 FF)
(administrator James O Dea, Dublin; 96 members; 40 FF)
I welcome members with Y-DNA, mitochondrial DNA and/or autosomal DNA
results to all of these projects, provided that they qualify for
membership on either geography or surname grounds as appropriate.
My standard e-mail invitation, sent, for example, to people who contact
me because they match a kit I manage:
I am co-administrator of the FamilyTreeDNA Clare Roots project which
you can join by going to https://www.familytreedna.com/sign-in?ReturnUrl=%2Fmy%2Fgroup-join%3Fgroup=ClareRootsSociety,
logging in to your FamilyTreeDNA kit if necessary and following the
The "About" page outlines both benefits of
membership and responsibilities
If you want to see how many of your Family Finder matches are in this
project (or any project that you have joined), then just go to the
Advanced Matches page at https://www.familytreedna.com/my/advanced-matches.aspx,
login if necessary and tick the Family Finder checkbox and select the
project in the "Show Matches For" dropdown.
In order to achieve the full benefits of project membership, you must
take action to give the project administrators "Limited" or "Advanced"
access to your kit.
This can be done when you join or can be done later at https://www.familytreedna.com/my/project-preferences.
For each project that you have joined, click the orange Edit button and
if necessary select "Advanced" or "Limited" in place of "Minimum" from
the relevant dropdowns, then click Accept and then click Confirm.
If you belong to more than 10 projects, then they will be split across
The Clare Roots project probably has an emphasis on the old Kilrush
Poor Law Union because of my involvement in the Kilrush
& District Historical Society and my recruitment
and swabbing campaign is also largely in that area (318 FTDNA kits processed).
The Clare Roots project has outgrown the 1000 individuals shown on the
(private, unhyperlinkable) member information web page and
the 500 lines available on the (public, hyperlinkable) Y-DNA results page.
A county is probably too big an area for an effective DNA project or
genealogy database; the Poor Law Union, more recently known as
Superintendent Registrar's District, is a more appropriate size; the
wide diaspora of a single parish alone or even of a single townland
could constitute a very useful and interesting project.
The public DNA results pages make no
reference to autosomal DNA, and the Y-DNA Colorized Chart is the only
one that I use.
Project members are encouraged to copy their Family
Those of us involved in geographically-based or surname-based
genealogical projects have come up with informal ad hoc ways of using
the other DNA websites for such projects.
Marrinan descendants: only 33 members of the FTDNA project, but over
105 GEDmatch DNA kits, with over 30 more on
Ancestry but not on FamilyTreeDNA or GEDmatch.
143 people have shared AncestryDNA match lists with one AncestryDNA
account, effectively an
autosomal project with no tools or sophistication: requires
- a tool to avoid clicking individually on all kits on the
dropdown to find all matches to existing project members;
- a chromosome browser; and
- triangulation tools.
I manage 46 GEDmatch kits.
GEDmatch tag groups are its nearest
equivalent to projects.
Shared tag groups not yet implemented.
No facility to reorder members of saved tag groups, e.g. by family
branch or by location.
Marrinan results are displayed using GEDmatch tag groups, broken down
geographically - see printout.
My personal genealogy database:
I also keep a Microsoft Excel spreadsheet listing whichever of the
following details I can find for autosomal
DNA kits of interest, to enable quick switching back and forth between DNA and genealogy databases:
- edited with Ancestral Quest on my Windows
- copied periodically to Reunion on my iPhone
- copied periodically (excluding details of living
individuals) to TNG behind a password on my
- started in 1987
- 135,445 individuals as of 18 October 2018
- I have found 1,752 of those individuals in one or more
online autosomal DNA databases
- I export ancestors of GEDmatch kits to a GEDCOM file if the
- I export ancestors, and 3C or closer also in DNA databases,
of Family Finder kits to a GEDCOM file if the user wishes
- Should the GEDCOM database include all one's research, or
have multiple GEDCOM databases for different surnames or different
- My philosophy is that I should never have the same
individual in two different master databases.
- But good software should be able to create a single-surname
database, by exporting all end-of-line ancestors with a particular
surname, plus all of their descendants, as a GEDCOM, for use in an autosomal surname project.
- known relative of my own?
- if so, closest relationship
- current surname
- first name
- birth name (if known) or "adopted"
- GEDmatch kit number
- RIN from my database, if connected
- GEDmatch alias
- DNA company
- e-mail address
- Facebook profile
- AncestryDNA profile
- AncestryDNA tree number
- AncestryDNA cfpid
- Formula to create hyperlink to Ancestry pedigree chart
The most useful autosomal tool provided to FTDNA project administrators
for private research
is the Family
Finder Illumina OmniExpress Matrix, the equivalent of
GEDmatch's Autosomal DNA comparison matrix.
My favourite GEDmatch tool is the Tier 1 matching segment
- Facebook discussion about the Hehirs of Cloverhill in IGP's County Clare Ireland Genealogy Group.
- Descendants of probably unrelated Hehirs from all over the
county chimed in.
- If Michael Brooks who m. 7 April 1901 in NY to Ellen Clancy
was Michael Brooks b. 25 March 1873 in Fountain to John Brooks and Mary
Hehir, then the first person in the sample matrix would be 2C1R to the
second person and 3C1R to the other four.
- In fact, Michael Brooks of Fountain m. 19 Feb 1901 in
Kilnamona to Mary McKee and never emigrated.
- The descendant of Michael the emigrant is not on
GEDmatch, but is in the FTDNA Clare Roots project, along with five
known relatives of Michael of Fountain.
- The display to five decimal places conveys a spurious sense
- Rounding down to 0.00000 is at variance with the FTDNA
policy of otherwise counting half-identical regions of 1cM/500SNPs.
- Administrators needs facility to import the known
relationships from a GEDCOM file and scale the shared centiMorgans by
the expected shared centiMorgans for the known or hypothesised
relationship: when looking at different relationships in a single
matrix, "percentage of expected centiMorgans" would be a much more
meaningful scale that plain centiMorgans.
Communication with project members
Project administrators should be able to anticipate and answer questions before new
members have figured out what to ask.
- Different novices will want to be pointed towards different parts of the online information (FAQs).
- Project activity feed is more appropriate for broadcasting to members than mass e-mails.
- E-mail is probably better for one-to-one advice to
- Sometimes phone or Skype is the only way to persuade
members to do something.
- We meet visiting diaspora project members somewhere like
the "genealogy office" in the corner of the bar in the Stella
Maris Hotel in Kilkee.
- For many projects, Facebook posts have supplanted the
activity feed (and Ireland Reaching Out parish pages)
as the easiest way to facilitate mass communication to a wide audience.
Unanticipated consequences of GDPR:
It can be hard to persuade project members and others to:
- Before GDPR changes, administrators of multiple projects could switch quickly between kits in any of their projects via
- Since GDPR changes, there are no longer the same benefits from having separate projects (or having a
private project for people who don't fit anywhere else, of which the
main benefit was avoiding the need to repeatedly log out of one kit and
log in to another);
- FTDNA's interpretation of GDPR is that mailto: links to the
e-mail addresses of all project administrators must be displayed
publicly on the website to spammers and non-FTDNA customers, where
pre-GDPR e-mail addresses were available just to matches and optionally
to project members;
- my long-standing guidelines on e-mail etiquette forbid anyone from publishing my e-mail address online, in order to protect it from spammers.
- answer e-mails or AncestryDNA or MyHeritage messages;
- add pedigree charts;
- add most distant known patrilineal and matrilineal
- upload to GEDmatch and other databases;
- order Y-DNA upgrades;
My major successes
- Adoption reunions;
- O'Brien/O'Dea surname/DNA switch;
- llinking branches together;
- verifying or
refuting the results of traditional research;
- predicting a fourth cousin relationship purely on the basis
of surname and townland, and subsequently finding a 0/67 Y-DNA match;
an ChlŠir (Memories of Clare) slogan: 'When an old man dies,
a library burns'.
We owe it to our descendants:
Those running DNA databases owe it to their stakeholders (and those
participating in the databases owe it to fellow participants) to
best possible service, information and tools, as in the following wish
- to record the memories, the oral traditions and the DNA of
our older relatives, all of which often go back beyond the history can be found in surviving
- to do unto our descendants as we wish that our ancestors had
done unto us;
- to leave our DNA to our descendants as we are trying to reconstruct that
of our ancestors; and
- to record and keep alive the history of local areas and
Wish list (cf. Challenges above)
- Easier sharing and comparison across DNA databases.
- Every DNA kit to have a public pedigree chart and use the full 50 characters to describe MDKAs.
- Women to always use their maiden names (but not to call
themselves Mrs Maiden-name!).
- Better algorithms to identify ancestors who must be dead, instead of
calling them "HIDDEN" or "Private".
- Every member of every DNA database to have enough public information on ancestors that there is never a
need for follow-up questions.
- Separation of paternal and maternal autosomes.
- Couple the DNA database to the genealogical database by embedding DNA kit identifiers in the GEDCOM, allowing more automation:
- Combine the GEDmatch matching segment search with FTDNA's Family Matching:
For each half-identical region in the matching segment search output,
- the relevant most recent common ancestral couple if
is a known relationship;
- the ancestral couple through whom the relationship
appears to come if there is a triangulation with a known relative;
- a paternal/maternal indicator if phasing is possible;
- a "relationship-undetermined" indicator otherwise.
- Search my genealogy database to identify:
- all end-of-line ancestors with a given surname;
- variant surname spellings which might be included;
- all descendants of these end-of-line individuals;
- those of the descendants who are in the DNA
- their lines of descent from the end-of-line
- pull all the DNA kits into GEDmatch;
- create a tag group with members sorted by lines of
descent (closer relatives grouped together);
- generate an autosomal comparison matrix;
- scale the observed shared centiMorgans by expected shared centiMorgans.
Thanks to Cindy Wood of the Marrinan project for many of the thoughts
and suggestions above.