Wednesday, June 18, 2014

It's all in the Recombination

So I blogged recently about the McCurdy cousins and our DNA results. I found a few other cousins that were from another branch of the family, and the results seem interesting to say the least. It speaks volumes as to the recombination of DNA, and the many possibilities. Now if we can only figure out which DNA is which.

The ancestor's in question are
John Sunday and his wife Barsheba Collins.

John Sunday was most likely Sondes, Sontag or Sondag and from France or Belgium. He emigrated to the United States prior to his 1809 marriage to Barsheba Collins in Baldwin County, Georgia.

We have no clues as to Barsheba Collins family except at the same time two other Collins married in Baldwin County, Georgia, Absey Collins married Brittain Pearman and moved to Coffee County, Alabama, and Mary or Polly Collins married a William Owens. I have been unable to trace her.

John was born 1780-1790 and Barsheba about 1784.

The other ancestor is  Elijah McCurdy born 1793 in Pendleton District area of South Carolina. Elijah married a Martha "Mattie" Bowen prior to his marriage to Barbara/Barsheba Sunday in Pensacola, Florida.

The cousins are as follows.

LM and JH who are 4th great grandchildren  with the common ancestor being John Sunday and Barsheba Collins. (5th cousins to most of the participants)

Elijah McCurdy is my father's 3rd great grandfather, and John and Barsheba are his 4th great grandparents.

GK shares the same relationship to them but through a different brother. LH shares the same set of 2nd great grandparents as my father. SM is a 3rd great granddaughter of Elijah and his first wife, and of John and Barsheba. SF is both the 2nd and 4th great granddaughter of Elijah, and the 3rd and 5th great granddaughter of John and Barsheba. CB is the 3rd great granddaughter of Elijah and his first wife and is not related to the Sundays. CV hasn't shared her info, but based on her age and what I know of her genealogy, she is either the 3rd or 4th great granddaughter of Elijah and the 4th or 5th of John and Barsheba.

Thus, Dad, GK, LH, and possibly CV are 5th cousins to LM and JH.
SM is fourth cousin once removed to LM and JH.
SF is fourth cousin once removed and 6th cousin to LM and JH.

The relationships of the cousins as McCurdy's is a bit more complicated.

CB and SM are full fourth cousins, and half fourth cousin to my father, LH and GK, and CV (or half fourth cousin once removed). To SF they are half 3rd cousin once removed twice on the McCurdy side. My father and LH are 3rd cousins.

Since SM is also a Sunday through a different line, she is through the Sundays, a fourth cousin once removed to my father, LH, GK and CV (or fourth cousin twice removed). To SF she is 4th cousin and fourth cousin twice removed.

Complicated isn't it?

GK's great Grandfather is also SF's great great Grandfather. And her other Great Grandfather was the Uncle to that one on a totally different line. Thus not only are both GK and SF McCurdy's twice, but they are also a Morris.

SF tested at a different company and so we can only compare her to those who are on gedmatch.

Aside from a few  matches we have in common, we have some interesting results.

None of the matches match CR except a Sunday, a family she does not descend from. CR matches my father and SF, but so does her mother, which indicates an unknown relationship for us both to her mother's side of the family. Although I match CV, LH, GK and SF, I do not match CR or SM.

Of the remaining cousins
LM is matching SM, GK, and SF.
JH is matching SM only. We can't test against SF.
SM matches my father, GK, LH, LM, JH, and CV. We can't test her just yet against SF.
GK matches my father, LH, CV, SF and SM.
LH matches GK, CV, my father and SM. We can't test against SF.
CV matches GK, LH, SM and my father. We can't test against SF.
My father matches GK, LH, SM, CV and SF.

Of the matches it is even more odd.I have omitted GK and SF because of the complexity of their relationship.

 On chromosome 1
GF, myself and my father
SM and JH also share a match
SM and LH as well.

Chromosome 3 has several sets of matches.
Dad, myself and CV share a few matches on one segment.
SF, GK, and LH match on another.
On another, Dad, LH, myself and CV all match
And on one more Dad, LH and myself match.

On Chromosome 6
SM and LM

On Chromosome 7
SF and LM
GK and SM have two different segments

On Chromosome 10
SM, GK, LH, CV and my father all match
SF has no matches on that Chromosome with my father or the others.

On Chromosome 12
Dad and SF match
CV and SM match

On Chromosome 14
Dad, CV and SF

On 17
Dad and LH

On 19
Dad and GK

On 20
Dad and CV

And finally on the X chromosome
Dad, GK and I all have a match
LM and GK have two small segments
LM and SF also have a small segment.

Given that she has double the X data from the Sunday family (two of four great grandmother's, both on her mother's side), it makes sense that GK is sharing the most X chromosome DNA. That the last common female they had in common (my Dad and GK) is Barbara Sunday is I think pretty special. Of the 182 cM of X chromosome I share with Dad, 37.3 cM is shared with Gloria (20 percent). I think that is a pretty significant chunk of DNA to share. Add that to the 36.3 cM she shares with LM and 40 percent of her X chromosome is from either John Sunday or Barsheba Collins.

The total DNA shared is also of interest.

SM with Dad 13 cM, with GK 15 cM, with LH 11 cM, with CV 20 cM and with LH  19 cM, and 6 cM with JH.
LM is additionally matching 27.3 cM with SF and 36.3 cM with GK.
GK with Dad 60 cM, with me 46 cM, with LH 49 cM, with CV 8 cM
LH with Dad 42 cM, with me 22 cM and with CV 27 cM
and finally CV shares 100 cM with Dad and 38 cM with me.

SF is harder to compute, as we can't compare her to everyone. Of the above not mentioned already, we show a match of 70 cM with Dad, 32.9 cM with me, and 256 cM with GK.





Thursday, June 5, 2014

Chromosomal mapping of Ancestry from 23 and me

Yesterday I posted how I made a spreadsheet utilizing the ancestry composition file I downloaded from DNAgedcom. I was thinking of an easier way to use the information to determine if the matches were a possible connection to my grandfather's Choctaw relatives. Then it hit me. I could mimic the table format used by 23andme.com and put the data into genome mate.

The process I came up with was a bit labor intensive. In the fact that you could more easily use the minimum default values for cM and SNP totals, but I wanted it to be more accurate. Even if you don't do what I will describe below, you will get a map of the Chromosome's that will look something like this.

I searched and read for quite a while before I found the necessary instructions to find the cM data. I finally found the data on this blog. Note your eana and American results will overlap, there is no need to do the same section twice.

To figure the cM units for a segment
1.Go to Rutgers Map Interpolator.
2. Open your spreadsheet with your ancestry composition information from 23 and me.
3. On Rutgers Map Interpolator, select the Chromosome from the drop down menu.
 4. Select physical location only.
5.  Put the start number in on one line, put the stop number in the next. Either copy and paste or type.
6. Select get map positions.
7 Take the sex averaged cM's, subtract the start from the stop. This is your cM value. For the X chromosome there is no other option than female cM position. (round to the nearest tenth)

Thus on a result such as this
Map Interpolator Results
 
No.  Chr Query_Posn Sex_Ave_cM Female_cM Male_cM
1   9   10411338    25.6840   17.0798 34.4570
2   9   87589165    87.0749  107.9594 67.5525

The cM value would be 61.4 cM

To ensure that I had the right format I copied and pasted a result from family inheritance advanced on 23 and me  in view as a table to a blank document. I then entered the data into the table exactly  as follows. (using the above example)

Jane Doe vs. Native American 9 10411338 87589165 61.4 cM 700 

I used the default value of 700 SNP because I couldn't find a way to figure out the total of each segment. I did run across several matches where I got a NA for a result. In that case I used the default of 7.0 cM (which is why I learned this was all unnecessary).

I then took the table and copied it. 

What worked for me was to first convert the table to text separated by tabs. I then did the manual entry for 23andme data on the genome mate program. Basically you select the 23andme family inheritance advanced option on import and paste the data and hit format. Don't forget to save or backup your data before and after you do your upload.


Wednesday, June 4, 2014

23 and Me and Ancestry Composition

While I prefer the admixture programs available at Gedmatch.com (hint, donate, I am not affiliated with them, but it's a great site and I would hate to see it go away), I found a useful tool to find the specific Chromosomal information of the Ancestry Composition in Start and Stop points.

Since my mother has Native DNA, I am hoping (and sadly found out with this, we have none yet) that I can break down some brick walls by matching DNA with some of my Choctaw kin. When using DNAgedcom to export the information I needed for Genome mate, I noticed in the files an AC or Ancestry Composition CSV file.

When I opened the file, which opened in a spreadsheet, it was the breakdown of the ancestry components by chromosome from the Ancestry Composition at 23 and me. I use Open office and not Microsoft Works so I can only tell you how I made myself a spreadsheet from there.

When I first opened the file, it was all in the first column, so I clicked the Data option at the top and clicked Text to Columns, and changed the parameter to commas. I then had a table but it wasn't sorted. So I then went to Column A, do not highlight whole column, and then chose Data and Sort. I sorted by Column A.

What I ended up with is a Table sorted by Chromosome (though not in order Ch1, Ch10, Chr 11 ect then Ch 2, Ch 21, etc. ).

I then took my matches (from genome mate or 529) and compared the matches to the numbers. You have two separate sets of data. One I think is for when you show the Subgroups and the other for when you don't, either way the data is the same. The native matches are designated eana and American.

Hope this helps

You've got (DNA) matches!

Okay, so I am dating myself a little, parodying "you've got mail" but I think it's the closest thing I could think of to describe the excitement (which we all had initially with AOL), of receiving my first autosomal DNA matches on my parents and my DNA.

Then I tried to figure out the matches. Oh, my. Out of the 2000 matches for my parents, I have figured out 13 easily. The rest has had me scratching my head. I also signed up for gedmatch, a free site run by donations (hint if you use it, donate), which is currently down while they upgrade their servers. I got several more matches (about 1400 per person), but had some bumps along the way as I learned just how to navigate triangulation of matches.

The first thing I learned is just because two people match you at the same point on the same chromosome, doesn't mean they match each other. You get half of that DNA from each parent, so if they don't match that means that one person matches DNA from your father and the other your mother.

The second thing I learned is that sometimes large segments get passed entirely from parent to child. What that means is that if my Dad has a 28 cm match and I got it, the relationship estimate for my relationship won't be accurate. I don't think this is so much in the matches for 1-3rd cousins but beyond, it can make it harder to locate the relationship. Mom has a 6th cousin she shares a pretty large segment with, that I don't match at all, but she has two 4th cousins from the same line that I match in the exact way.

The third thing I learned was how to use some nifty free tools available. For 23andme.com, you can use DNAgedcom to download a zip file with specific data. You can also install the 529 extension in chrome. Genome Mate a free program is a desktop database/spreadsheet program that allows you to track your matches across several venues (gedmatch, 23andme, ancestry.com, ftdna, etc). All these tools are going to be needed, unless you can do it yourself, because you must keep track of the matches.

 If you don't have surnames in common, you need to cross your fingers for some more matches in that segment. Chances are even if you have a fairly robust tree, you don't have adequate research on the cousins of your 4th great grandparent. Or you haven't researched all the lines from each sibling of that great grandparent. Some folks have a tree, but some don't. It is somewhat frustrating because I thought DNA would make it easier to break down my brick walls. So far, it's given me a lot of information that I have no clue how to use.

When you have a match, it is helpful, because anyone in that segment who matches your match shares the same common ancestor. Because a rather large group of my McCurdy cousins have tested, I was able to work with them and we have started compiling a list of specific segments that are "McCurdy" but here's the best part, we don't all match each other in the same place!

Because of intermarriage of cousins and more, we have a rather complicated tree on my McCurdy's.

Cousin A has two great grandmother's who are sisters, and daughters of my Dad's GGGrandfather's brother.
Cousin B shares my Dad's GGGrandfather, but is only a half sibling descendant. The other half on that same line is a niece of Cousin A's McCurdy's (and, lol, she shares the other Surname Morris).
Cousin C shares my Dad's GGGrandfather and GGGrandmother.
Cousin D shares only a 3rd G Grandfather of my Dad's with the rest.
Cousin E is a half sibling descendant of the 3rd GGrandfather, but we found out that Cousin B and Dad match her mother who isn't a McCurdy (she matches none of us on the McCurdy's).

Only one match is in common for more than two of us. Another match may in fact be a match to the Beck family (the wife of Dad's GGGrandfather) and not a McCurdy Match. We have several more cousins awaiting results, with each piece we will build it together. We are hoping we can get Cousin B's mom to test because she and her first cousins that are alive are the closest we can get to Elijah McCurdy (great grandchild) and we are hoping it will open up other McCurdy families.

As you can see, it's complicated. It's not easy, but it sure is fun.

Will keep you posted.