Group 3 Update (June 2019) - part 1

Description of the Group

There are 32 people in Group 3, making it the biggest group in the project. It also has the largest number of surname variants (see below), the most people who have undertaken SNP testing, and is the most genetically diverse.

Characteristics of Group 3

And almost immediately we hit a major problem - many of the people in this group are "Chance Matches" and misleadingly appear more closely related than they actually are. The reason for this is Convergence - in other words, as the Y-DNA signature of the people in this group has mutated slowly over the centuries, it has come to approximate the Y-DNA signature of people on nearby branches of the Tree of Mankind to the extent that those neighbours start turning up as matches in their match list, even though they are on a completely separate branch of the Tree of Mankind, separated by more than 1000 years.

We can see this reflected in the SNP results for the group, which help to identify on which sub-branch of the Tree of Mankind the various project members sit. So far, there have been 6 subgroups identified for Group 3. All the subgroups in Group 3 test positive for the SNP marker M222. This particular marker is associated with Niall of the Nine Hostages (the legendary Irish warlord). And this branch of the Tree of Mankind is notorious for Chance Matches due to Convergence.

The SNP Progressions for these subgroups are in the footnotes [1] and are summarised in the diagram below. From this you can see that the MRCA (Most Recent Common Ancestor) for the entire group sits on Branch Z46375 and lived >1900 years ago, and the earliest MRCA between any two branches (DF85, shared by 3b & 3c) was born >1850 years ago.

So, even though members in the different subgroups match each other and have similar sounding surnames, none of these subgroups are related to each other since the foundation of surnames (about 1000 years ago in Ireland) and therefore the various subgroups have completely distinct genetic origins.

The 6 subgroups in Group 3 are unrelated within the last 1000 years
(Dates in black are from YFULL & dates in blue are calculated
from FTDNA's Big Y Block Tree allowing 100 years per SNP)

A further problem is that because not all project members have done SNP testing, we cannot be 100% certain that everyone has been accurately grouped in the correct subgroup. Only SNP testing (preferably the Big Y test) can categorically define to which subgroup a project member belongs.

Let's take a look at each subgroup in turn.

Group 3e (Molloy)

There is 1 member (a Molloy) in this group (DRM-0572). He has 9 matches at Y-STR-111, 1729 at 67 markers, and 693 at 37 markers. He is also in the Malloy-Molloy DNA Project (of which I am a co-administrator). He matches a Malloy with a Genetic Distance of 5/11 and together they may form a new genetic subgroup, but the match is in neither DNA Project currently. The SNP Progression for this member is:

• M222 > Z46375 > DF106 > DF104 > DF105 > A10528 > Y96240 > BY71053

His closest SNP-confirmed matches are people with the surnames Fullerton (under Y96240) and John, Fergus, MacDonald, Lindsey, McFadden & Cryans (under A10528). This suggests there may be a Scottish origin for this particular genetic group.


Group 3d (Malloy)
This group also contains a single member (JPM-2799) and this member is also in the Malloy-Molloy DNA Project. In this other project, he is placed in a group with 16 people (Group 4) where the dominant surname is Molloy (9 people) and many have tested with the Big Y (due to the high rate of chance matches due to Convergence). Seven group members have roots in Offaly and it is highly probable that the members of this particular group are descendants of the Molloy's of Fircall, King's County (Offaly). Further details of this group can be found on the Molloy Clan website here. This group is also discussed in a YouTube video of an update I gave on the Molloy DNA Project at their Clan Rally in 2017. 

• M222 > Z46375 > DF106 > DF104 > DF105 > ZS8379 > BY11707 > A11391 > A11427 > BY21143 > L170 aka ZS10825

Summary of Groups 3e, 3d, and 3c - two are possibly Scots-Irish, one is probably the Fircall Molloy's of Offaly
(click to enlarge)

Group 3c (Melloy)
There are two members in this group, both with presumed origins in Northern Ireland / Ulster and both testing positive for SNP marker A10680. The SNP Progression is:

• M222 > Z46375 > DF106 > DF104 > DF105 > DF85 > S673 > S668 > BY11548 > A10680. 

Nearby genetic neighbours (on The Big Tree) include Kilcoyne, Craig (A10680); Lloyd (it should be Claycomb), Dunaway (BY93206); Galloway, Gallagher, Bain (BY25887); and Galyean (BY11548). Additional genetically-related surnames were harvested from FTDNA's Big Y Block Tree and the M222 Haplogroup Project and include Claycomb (BY93206); & Curran (BY25887, but MDKA was a Gallagher). These can be seen in relation to Group 3c on The Big Tree below.

The position of Group 3c on the Tree of Mankind

Surname Distribution Maps [2] indicate that the deeper origins of this particular group could lie either in the Galway / Mayo area or the Down / Antrim area. They may thus have Scots-Irish origins as the names Malloy, Melloy and Milloy all occur in the Scottish Lowlands (see below).

Analysis of Surname Distribution (mid-1800s)
(based on Griffth's Valuation at www.swilson.info/sdist.php)
(click to enlarge)

Malloy, Melloy & Milloy may have originated in the Scottish lowlands (SW Scotland)

Until more people join this particular group, and undertake Big Y testing, there is not much else we can say at this point.

Group 3b (Maley)

This group consists of 4 members - two Maley, one Malley, and one Mellen. But there is considerable doubt regarding whether or not they should all be grouped together in the same group. Below is a screenshot of their first 37 STR marker values together with a Genetic Distance (GD) Matrix summarising how closely they are all related to each other:

• The 1st and 4th members (Maley & O'Malley) seem correctly grouped because of the closeness of their match (Genetic Distance [GD] 3/37) as well as the Unique STR Pattern (USP) that they share (see the mutations in blue in the screenshot below - 30, 17, 17). 
• And the 2nd and 3rd members seem like a close enough match (GD 4/67) ...
• ... but what about members 1 and 2? ... the GD is probably too big (16/111) ... and member 4 has only tested to 37 markers so the GD of 6/37 with members 2 and 3 seems too big ... so the only feasible connection is between members 1 and 3 (Maley & Maley) as they share a GD of 6/67 (which in turn is just above the threshold for declaring a match to be present).

Group 3b - are the project members correctly grouped?
(GD, Genetic Distance; grey squares indicate number of STR markers tested)

In short, the connection between members 1 & 4 seems valid, and the connection between members 2 & 3 seems valid, but the connection between the first two and the second two is tenuous.

The only way that we can establish if the connection between all four members is valid is for everyone to do the Big Y test - this will prove definitive. Even if everyone upgrades to 111 STR markers, this may not answer the question, so the Big Y is the preferred test under these circumstances.

There is also evidence of "Chance Matches" due to Convergence in this group. The 1st member (RLM-7064) has 69 matches at the 67-marker level of comparison, and this is quite a lot. Of these, 27 have done downstream SNP testing. Analysis of these matches (see footnote [3] below) reveals that at least 20 of the 27 are Chance Matches i.e. they sit on separate branches of the Tree of Mankind. And if this is extrapolated to the entire list of 69 matches, it means that at least 74% of them are "false positives".

Despite these limitations, we can draw some tentative conclusions about the origins of Mellen-1317. He has tested to 111 STR markers and has done the Big Y test. His position on the Tree of Mankind is illustrated below. His SNP Progression is:

• M222 > Z46375 > DF106 > DF104 > DF105 > DF85 > BY35719

He sits on branch BY35719 (>1200 years old), one of the many downstream branches of DF85 (>1850 years old). The surnames of other people sitting on this branch include: Kerr, Graham (from The Big Tree); Lamont, McBride, & McBrayer (from FTDNA's Big Y Block Tree). These names are all clearly Scottish and point to a Scottish origin for this particular project member. If the other group members do the Big Y test and match Mellen-1317 on the same SNP (BY35719), then they too can assume Scottish origins.

The position of Mellen-1317 on the Tree of Mankind

Group 3a1 (O'Malley)

This group also has 4 members. And once again we are faced with the question: have these people been grouped appropriately?

• From the screenshot and Genetic Distance Matrix below, only the 3rd & 4th members seem to have a highly probable close connection (GD 0/37 i.e. they are exact matches).
• The 1st member has a reasonable GD to the 4th member (GD 5/67) and this suggests he is accurately grouped, but his GD to member 3 (GD 5/37) and especially member 2 (GD 10/37) goes against this.
• It seems likely that the 2nd member (0126) does not belong in this group - his GD to the other 3 group members indicates a more distant connection, but conceivably this could still be possible within a genealogical timeframe (1000 years for our purposes here).

Group 3a1 - are the project members correctly grouped?
(GD, Genetic Distance; grey squares indicate number of STR markers tested)

There is also a strong suggestion of chance matches due to Convergence in this group:

• the 2nd member (-0126) has 468 matches at the 37-marker level of comparison
• the 4th member (-7761) has 323 matches at the 67-marker level of comparison

If we conducted a similar exercise as the one we did for Group 3b and compared the SNP Progressions for those SNP-tested matches among the 468 and 323 above, we would probably find that the proportion of chance matches is again about 70-80% (at least).

Only member 1 (-3727) has done the Big Y test and this has identified his Terminal SNP as BY35297. His SNP Progression is as follows:

• M222 > Z46375 > BY35297

BY35297 is relatively upstream and there are 48 branches downstream of this particular SNP marker (per FTDNA's Public Haplotree). Member -3727 sits on a very isolated branch of the Tree of Mankind and shares a common ancestor with his nearest genetic neighbours at least 1700 years ago. Many of these neighbours have Irish origins - see The Big Tree here (member -0727 is on the far right).

To clarify if all members are accurately grouped together, and where the deeper origins of this group might be, the other members of this group would need to do the Big Y test (and upload their results to the Big Tree).

In the next post we will look at the last and largest subgroup within Group 3, namely Group 3a.

Conclusions & Next Steps

Group 3e (Molloy) is probably Scottish in origin. Members should join the M222 Haplogroup Project and the Scottish Geographic Projects (see links below).

Group 3d (Malloy) probably represents the Molloy's of Fircall, who originated in Co. Offaly (Kings County). Join M222 & Ireland Y-DNA Project.

Group 3c (Molloy) is probably Scots-Irish, possibly originating from the Scottish surname Milloy or Melloy, and may have a similar background to Group 3e above. Join M222 and Scottish projects.

Group 3b (Maley) may not be accurately grouped. The one Big Y-tested member may have Scottish origins. Join M222 & Scottish. Do Big Y to clarify.

Group 3a1 (O'Malley) may not be accurately grouped. The group appears to be Irish in origin but more Big Y data is needed. Join M222 & Ireland Y. Do Big Y to clarify.

All group members should join the M222 Haplogroup Project. Just click this link and sign in to your FTDNA account.

In addition, Groups 3e, 3c, and 3b should join the Scottish Geographic Project (to join just click here). And Groups 3d, 3a1 should join the Ireland Y-DNA Project (click here to join). The Admins in these projects may provide some additional insights and advice on further SNP testing. I am frequently in touch with these Admins about project-related matters.

All members should supply a birth location for their MDKA - see item 2 on this page for full instructions.

All members should supply pedigree information about their direct male line - full instructions are on the Post your Pedigree page.

More people should do the Big Y test - because so many Group 3 members have Chance Matches, it is the only way to ensure that you are accurately grouped. All people undertaking the Big Y test should upload their data to The Big Tree for some additional essential analyses (provided for free by Alex Williamson and his team).

Maurice Gleeson

July 2019

Footnotes & Sources

[1] A SNP Progression is simply the sequence of SNP markers that characterises each branching point on the Tree of Mankind, starting "upstream" at the level of the Haplogroup (R in this case) and progressing all the way "downstream" (i.e. towards the present day) to the Terminal SNP. Comparing SNP Progressions helps us see exactly where each Terminal SNP sits on the Tree of Mankind relative to the others.

All 6 subgroups in Group 3 are on the M222 branch of the Tree of Mankind. The SNP Progression down as far as M222 is as follows:

• R-P312/S116 > Z290 > L21/S145 > DF13 > Z39589 > DF49/S474 > Z2980 > Z2976 > DF23 > Z2961 > Z2956 > M222 ...

The SNP Progressions for each of the 6 subgroups from M222 downstream are as follows:

• M222 > Z46375 > BY35297 (Group 3a1 - O'Malley)
• M222 > Z46375 > DF106 > DF104 > DF105 > A10528 > Y96240 > BY71053 (Group 3e - Molloy)
• M222 > Z46375 > DF106 > DF104 > DF105 > BY35730 > BY35759/60/61 (Group 3a - O'Malley)
• M222 > Z46375 > DF106 > DF104 > DF105 > DF85 > BY35719 (Group 3b - Maley)
• M222 > Z46375 > DF106 > DF104 > DF105 > DF85 > S673 > S668 > BY11548 > A10680 (Group 3c - Melloy)
• M222 > Z46375 > DF106 > DF104 > DF105 > ZS8379 > BY11707 > A11391 > A11427 > BY21143 > L170 aka ZS10825 (Group 3d - Malloy)

[2] Surname Distribution Maps for Ireland can be found on several websites:

• John Grenham's website (www.johngrenham.com) has maps based on Griffith's Valuation (from the mid-1800s). The first 5 searches are free and then you need to take out a subscription.
• Shane Wilson's website also uses Griffith's Valuation (https://www.swilson.info/sdist.php). No subscription needed.
• Barry Griffin's website has maps based on the 1901 & 1911 censuses, including maps identifying just the farmers (whose families are likely to have held the land for several generations at least).
• I used to use the PublicProfiler websites (WorldNames & GBNames) but these no longer seem to be maintained.
• I also used to use the forbears.co.uk but the new format provides less fine detail than previously.

[3] RLM-7064 has 69 matches at the 67-marker level of comparison. The table below shows the 27 of his 69 matches who have done downstream SNP testing, their Genetic Distance (GD) from RLM-7064, their surname, Terminal SNP, and SNP Progression (showing the series of SNP markers from M222 down to their Terminal SNP). These matches sit on many different downstream branches of the Tree of Mankind. Most are a GD of 7/67 (17/27, 63%), eight are a GD of 6/67 (30%), and single matches are a GD of 4/67 and 5/67 (4% each). This indicates that chance matches are more common at the higher levels of GD.

In the best case scenario, let's assume that RLM-7064 actually sits on Branch BY11723 (rows 10-13 below). Four of the 27 matches sit on this branch. In addition, 3 members have only tested as far down as M222 - let's assume that these would also have tested positive for BY11723 if they had tested further downstream SNPs. That means that at most 7 of the 27 matches sit on this particular branch along with RLM-7064. That leaves 20 out of 27 matches (74%) who sit on entirely different branches of the Tree of Mankind and cannot be related to RLM-7064 within the last 950-1100 years (the crude date for the formation of BY11723, based on YFULL and Big Tree data). In fact, most of these 27 matches (17/27, 63%) would not be related to RLM-7064 within the last 1500 years (because they do not sit on branches below A529, which is at least 1550 years old).

The 27 matches to RLM-7064 who have undertaken downstream SNP testing.