PDA

View Full Version : Technique to find Y-DNA matches using GEDMATCH: Help me test its validity



MitchellSince1893
08-13-2016, 04:02 PM
For those of you with known paternal line matches on Gedmatch, I'm requesting your assistance in determining the validity of a technique I'm experimenting with to identify paternal line autosomal matches...i.e. father's, father's, father's etc line.

The technique:

1. Use the gedmatch.com "One-to-Many matches" tool using a gedmatch ID that has confirmed paternal line distant matches (father's father's father's etc line). Set the cM to 3. and hit "Display results. It will take a while for the results to appear at 3cM setting.

2. Once the results display, sort by Y haplogroup, by clicking the left blue triangle under the "Y"

3. This step may require some research to get all applicable matches, but in the "Select" column check those matches that are in the same haplogroup as your paternal line sample.

4. Click the "Submit" button at the top of the page.

5. On the next page click the top option "Chr Browse" 2-D Chromosome Browser comparison.

A screen will appear showing where the matches you selected match you on chromosomes 1-22.

What I'm seeing after selecting the various nomenclature for my own haplogroup and pertinent subclades (e.g. U152, L2, R1b1b2a1b4c / R1b1b2a1a2d3, R1b1b2a1a2d3*,R1b1b2a1a2d etc) is that there are some segments that quite common among my fellow U152 matches.

Examples from chromosomes 4, 14, 15, 20 respectively. Pink segments are < 5cM, Blue segments are 5-10 cM in length, Green are 10-20 cM.

https://s-media-cache-ak0.pinimg.com/originals/17/a6/15/17a615f33a60e4616051f1911027baf2.png
I'm wondering if these common segments are inherited from my paternal line. As I don't have any known paternal matches (Non paternal event for my father's paternal grandfather...his father is unknown), I have no way to verify if these are indeed paternal line segments.

Can those of you with confirmed distant paternal y-dna matches, determine if your known paternal matches are appearing in these types of commonly shared (among y haplogroup matches) autosomal segments?

MitchellSince1893
08-13-2016, 05:05 PM
Background.

Using gedmatch and the genome mate software to keep me organized, I've been able to identify most of my father's maternal and paternal matches thanks to autosomal dna testing at FTDNA, ancestry.com, and 23andme (and subsequently adding them to gedmatch) by his paternal 1/2 sister, maternal 1/2 brother, and several known relatives on both sides.

In many cases I've been able to determine if the paternal match was via a paternal grandmother line. But my challenge has been to distinguish between matches to my father's paternal grandfather's mother (my Mitchell line), and his paternal grandfather's father (my mystery man line). As it stands I have a 50-50 chance a match on one of these segment is via the mystery man line.

My hope is this technique will help identify the paternal grandfather's father (mystery man) line segments in chromosome 1-22...enabling me to distinguish these segments from a paternal grandfather's mother match.

Additional details on my paternal line mystery and research found here. http://www.anthrogenica.com/showthread.php?3987-Help-me-to-solve-a-family-Mystery-My-paternal-line-before-my-great-grandfather

MitchellSince1893
08-13-2016, 10:40 PM
I went back and added even more U152 matches. Here are the relevant segments

Chr 1. 15,217,523 to 18,781,452. 4x U152 matches located here. Currently my father has no good matches on gedmatch
Chr 4. 68,299,960 to 78,012,399. 4x U152 matches located here. This is a Paternal grandmother segment so it's not valid
Chr 4. 108,860,749 to 114,698,102. 5x U152 matches located here. I have 4 paternal gedmatch matches >10cM located here
Chr 4. 142,417,367 to 154,991,635. 4x U152 matches located here. I have 2 paternal gedmatch matchs >12cM located here
Chr 12. 81,828,162 to 95,141,155. 6x U152 matches located here. I have 1 paternal gedmatch match located here with 13.8cM
Chr 14. 28,344,225 to 32,450,556. 9x U152 matches located here. This is a Paternal grandmother segment so it's not valid
Chr 14. 22,344,671 to 26,304,106. 3x U152 matches located here. This is a Paternal grandmother segment so it's not valid
Chr 15. 24,879,996 to 29,165,751 . 18x U152 matches located here. I have 2 paternal gedmatch matches >10cM located here
Chr 15. 27,711,033 to 31,568,615 . 7x U152 matches located here. Currently no good matches found here on gedmatch
Chr 18. 6,481,727 to 8,208,296. 5x U152 matches located here. I have 1 paternal gedmatch match >10cM located here
Chr 20. 57,061,634 to 59,648,619. 10x U152 matches located here. I have 5 paternal gedmatch matches > 10cM located here
Chr 21. 14,240,640 to 17,825,725. 4x U152 matches located here. This is a Paternal grandmother segment so it's not valid
Chr 21. 23,732,459 to 31,384,852 13x U152 matches located here. I have 1 paternal gedmatch match located here with 12.8cM

jbarry6899
08-14-2016, 12:18 AM
Mark, I tried this with the autosomal kit for my closest Y match, who also is a 5 generation match to my grandfather's Lazarus kit. I couldn't get it to run at 3 cm but at 5 it returned a cluster of U152 and/or subclades on ch. 18 from 56 to 63. I have no identified matches on that segment but interestingly it is adjacent to a large segment that I share with a person who lives in the same Cork parish where my paternal ancestors lived. We've never been able to find a common ancestor, but it is clearly a paternal line match. Also, several of the men in the 56-63 cluster have a very common Cork name, McCarty, but they are the only U152 McCarty's I've ever seen; they are invariably L21. So I would suspect an NPE in their line involving the introgression of a man from mine.

I'll try running at 3 cm tomorrow morning when there are fewer users.

Jim

MitchellSince1893
08-21-2016, 08:12 PM
Mark, I tried this with the autosomal kit for my closest Y match, who also is a 5 generation match to my grandfather's Lazarus kit. I couldn't get it to run at 3 cm but at 5 it returned a cluster of U152 and/or subclades on ch. 18 from 56 to 63. I have no identified matches on that segment but interestingly it is adjacent to a large segment that I share with a person who lives in the same Cork parish where my paternal ancestors lived. We've never been able to find a common ancestor, but it is clearly a paternal line match. Also, several of the men in the 56-63 cluster have a very common Cork name, McCarty, but they are the only U152 McCarty's I've ever seen; they are invariably L21. So I would suspect an NPE in their line involving the introgression of a man from mine.

I'll try running at 3 cm tomorrow morning when there are fewer users.

Jim

Jim,
Sorry for the late response...didn't get a notification you posted here.

Thanks for taking the time to do this. I had no takers when I asked for the same assistance on gedmatch Facebook page...just critiques of the premise...I fully understand this technique may not work...just want test it out.

Hopefully my shared terminal branch match's ftdna family finder results will be available soon (1000 AD MRCA).

Anyone else want to give this a try?

jbarry6899
08-21-2016, 08:34 PM
Mark, I did try running it again and still couldn't get below 5 cM. So at this point I found it to be of limited utility, though it did help me to identify a segment that I will keep an eye on.

MitchellSince1893
08-21-2016, 09:44 PM
Mark, I did try running it again and still couldn't get below 5 cM. So at this point I found it to be of limited utility, though it did help me to identify a segment that I will keep an eye on.

Jim,

I failed to mention earlier that I ran this test on a lazarus kit I created for my dad's father.

Creating a "good" Lazarus sample was a lengthy process...was getting large shared segments to my father's maternal half brother...kept tweaking it until I eliminated this match (there are no shared segments between his half brother his paternal half sister). I used all 100 available entries with known paternal matches to create it. I had to change out a few that coincidentally had matches to my father and his maternal half brother.

Actually, I have 3 Lazarus versions. One at 12cM setting, one at 7.5cM and one at 6cM. When comparing them to each other there are matches one has that the other doesn't. It is interesting to run these lazarus kits on the various admixture tools.

The 3cM setting I mentioned in my first post was to maximize the number of y-dna matches. If I left it at the default settings I don't get any where near 2000 y-dna matches (more like 1000 after sorting for y-dna). Others may get more than 2000 y-dna matches at the default setting.

jbarry6899
08-21-2016, 09:54 PM
I created two Lazarus kits, one for my father and one for my paternal grandfather. I'm fortunate to have results for three paternal first cousins and more than two dozen distant cousins on the paternal side. The kit for my father is particularly useful for checking new matches. The kit for my grandfather is less reliable but can be helpful, as in the case of my closest YDNA match who has a measurable match to my grandfather's lazarus kit, although that Y match's results were not used in its creation.

Haven't done much with admixture on Lazarus kits but will give it a try.

MitchellSince1893
08-21-2016, 10:08 PM
I created two Lazarus kits, one for my father and one for my paternal grandfather. I'm fortunate to have results for three paternal first cousins and more than two dozen distant cousins on the paternal side. The kit for my father is particularly useful for checking new matches. The kit for my grandfather is less reliable but can be helpful, as in the case of my closest YDNA match who has a measurable match to my grandfather's lazarus kit, although that Y match's results were not used in its creation.

Haven't done much with admixture on Lazarus kits but will give it a try.

As I have no clue who my father's paternal grandfather's father was, for me the admixture was useful because it doesn't have anything out of the ordinary...unexpected. Based on the following it looks like my mystery man was NW European and probably British.

Eurogenes K15 6cM version
# Population (source) Distance
1 West_Scottish 4.56
2 Irish 5.24
3 Danish 6.11
4 Orcadian 6.13
5 North_Dutch 6.56
6 Southeast_English 6.75
7 West_Norwegian 8.22
8 Southwest_English 8.37
9 Norwegian 8.64
10 North_German 8.91

Eurogenes K15 7.5 cM version
1 West_Scottish 4.35
2 Irish 4.93
3 Danish 5.29
4 North_Dutch 5.92
5 Orcadian 6.31
6 Southeast_English 6.5
7 West_Norwegian 7.9
8 North_German 7.91
9 Norwegian 7.95
10 Southwest_English 8.17

Eurogenes K15 12cM version
1 Irish 4.9
2 West_Scottish 5.17
3 Southeast_English 5.97
4 Danish 6.47
5 North_Dutch 6.95
6 Southwest_English 7.18
7 North_German 7.65
8 Orcadian 7.73
9 Norwegian 9.8
10 South_Dutch 10.29

MitchellSince1893
08-28-2016, 04:21 AM
Jim,
...Hopefully my shared terminal branch match's ftdna family finder results will be available soon (1000 AD MRCA).
...

His results just came out. Unfortunately he only shared one of the above mentioned U152 clustered segments with my father and my father's paternal sister.

Chr 1 14,911,998 to 16,632,203.

So it's not looking good for this experiment.