atDNA Under the Hood: Segments and Crossovers

In this blog, I would like to look at segments and crossover points. Kathy Johnston has a method in which she compares the DNA matches between 3 siblings. These matches are then used to show how the 4 grandparents contributed DNA to each of these 3 siblings. She states in her directions:

The goal is to use crossover lines in PowerPoint to determine segment matches with grandparents among siblings when no parents are alive.

In this blog, I will use Chromosomes 1 as an example. I would like to use this Chromosome because of the Frazer DNA project I am working on. In this project, there is a very clear Triangulation Group at Chromosome 1 that leads back to Richard Frazer, b. about 1777 in North Roscommon, Ireland.

The Technique Depends on Gedmatch

Gedmatch is used because gedmatch.com comparisons show not only matches, but which matches have Fully Identical Regions (FIRs) and those with Half Identical Regions (HIRs). HIRs are the normal matches that we have with everyone except our siblings. Half Identical means we match either on our mother’s side or our father’s side. According to the ISOGG Wiki,

Identical twins are fully identical at every point in their DNA. Other full siblings, including non-identical twins, share around 50% of their DNA, and have both half-identical regions and completely identical regions. The expected percentages for full siblings are 50% half-identical, 25% completely identical, and 25% not identical for an overall average of 50%.

Me and My Two Sisters

This is how I look compared to my two sisters and how they compare to each other at Gedmatch. The green indicates FIR, the yellow area is HIR and the red is no match. The blue is FIR or HIR match and is also represented in the table above the graph with start and end points. Vertical lines are drawn through each crossover where there is any change from HIR to FIR to no match.

Chromosome 1 HJS

According to Kathy Johnston,

Only one sibling can own a crossover point (with few exceptions); identify that sibling. The owner of the crossover is the one who is in-common with
that point for two comparisons.

For the first vertical line, I have an ‘S’ for Sharon as the owner of that crossover point. She has a crossover, from a HIR match to no match in the comparison of Sharon to Heidi and a crossover from no match to a HIR in the comparison between Sharon to Joel. Note that the 3rd ‘S’ should be a’?’ as Sharon has only one crossover out the the 3 comparisons. I had originally left it as an ‘S’ for simplicity. The segments between the lines will represent the DNA the 3 siblings got from our 4 grandparents.

How Do We Get 4 Grandparents’ DNA Out Of This?

Let’s look at the next to the last segment and the one before it. Sharon and Joel have a green area there. This means that Sharon and Joel got their DNA from both their mother and father. However, this DNA was from one of the mother’s parents and one of the father’s parents. These 2 Fully Identical matches are represented by green and blue. These are 2 of our 4 grandparents, but we don’t know which ones they are yet – except that one is a maternal grandparent and one is a paternal grandparent. They are put on Sharon and Joel’s horizontal lines (represented by an ‘S’ and a ‘J’) as Sharon and Joel match each other on both parent’s sides represented here by green and blue.

Phase 1

By the way, I picked the most difficult Chromosome to look at. Chromosome #1 is the longest chromosome and has the most crossovers. Note that the green and blue are bounded by 2 J’s on my line. However, the green and blue are not bounded by an ‘S’ on Sharon’s line. So I will extend her DNA match to where it is likely to change (either at Sharon’s crossover point or at the end of the Chromosome – whichever comes first).

Phase 2

Looking For Two More Grandparents

This is starting to fill out, but Heidi has no DNA and we only have 2 grandparents. Now look again at the Gedmatch chart. Heidi has no match with Sharon or Joel on the next to last segment as indicated by grey in the bottom part and red in the top. We know that she does have DNA at all parts on her chromosomes, so she must have gotten her DNA from neither of the green or blue grandparents on her paternal and maternal side.

Chromosome 1 HJS

Our new grandparents are going to be orange and purple. I hope they don’t mind.

Phase 3

I extended the colors to the right as Heidi has no crossover point in that direction. Also note as a check that in the last segment Sharon and Heidi have no match, so they have to have opposite grandparents contributing to their DNA there. Now in my 3 Gedmatch comparisons, there is a spot where all 3 siblings have a FIR shown in green. These means all 3 siblings got their DNA from the same 2 blue and green grandparents. I’ll extend the blue and green down and to the right and left to the appropriate crossover points.

Phase 4

Looking good for my green and blue grandparents. In the first J-H segment, Heidi and Joel have no match. This means opposite colors/grandparents. I blasted through the ? crossover as it doesn’t involve Heidi.

Phase 5

Next, I’ll fill in the rest of the no match areas with opposite colors/grandparents. I only see one more segment to add.

Phase 6

Time for the HIR’s

So  far, we have only added opposites. Now it is time for the Half Identical Regions (HIRs). That would be a green on the top with a purple on the bottom or orange on the top with blue on the bottom. On the Sharon to Joel comparison below there is a long yellow HIR. I will give that a shot.

Chromosome 1 HJS

That means that I need to keep one color the same between Sharon and Joel and change the other. I randomly extended the blue and put the orange on top. I don’t know what to do at the ?, so I stop there. Up until now all the colors have been relative. However, once a half identical color is chosen, it locks in where the grandparents are (even though we don’t know which grandparents go on which segments yet).

Phase 7

Also notice that I put numbers on the bottom. Those are the positions of some of the crossover points. Sharon’s first match with Heidi stops at about 8 million. Sharon’s first match with Joel starts at the same spot.

Now Sharon to Heidi has no match in the large middle segment. That means I need to add the opposite colors/grandparents. Joel to Heidi is HIR, so we can continue Joel’s colors to the left. Sharon to Joel needs to be opposite in the first segment.

PHase 8

That seems to be all the easy segments. Actually, the upper left space can be filled in as Sharon and Joel match with a FIR. That leaves the lower right hand side. I’m not supposed to guess here, but I’m tempted to go with orange/blue to fill in my missing middle segment. The tiny segment below that is opposite. I guess again and pull the orange to the left and the blue to the right to fill the larger segment. My last guess is on my own segment where I extended the blue and added some orange.

Phase 9

That fills everything in, but with a bit of guessing. Note that the tiny crossover point at 205 has not been used, but that is a fine point.

Now For the Reality Check

I would be nice to put some names on these grandparents and check to see if I got these segments straight. Here are some of my sister real life Sharon’s matches.

Sharon to All Chr 1

Catherine is on my mother’s father’s side and James is on my father’s father’s side. The rest are Frazer relatives, but the matches are small. My results are somewhat similar to my sister Sharon’s. The difference is that I tested at 23andme and found a relative on my mother’s mother’s side (Judith).

Joel to All Chr 1

Let’s see if Heidi’s matches shed any further light.

Heidi to All Chr 1

Now we are getting somewhere. We see the Triangulation Group that I was interested in. This is on the right between Frazer relatives Michael, Paul, Jane and Bill. This is good because the Frazers are on my paternal mother’s side and James is on my Hartley side which is my paternal father’s side. Heidi matches James between the locations of 107 and 155 million. And she matches the Frazers from about 205 to 237 million.  Now all I have to do is find an open segment on Heidi’s line from 205 to 237 that is a different color on Joel and Sharon’s line. That is orange.

Will the Real Grandparents Please Step Forward?

Here is my first shot. I can already see it will need adjustments.

Frazer Segments

My paternal Frazer/Hartley Frazer side isn’t too bad. The bottom right in orange is where the Triangulation Group is that indicates the Richard Frazer family (b. 1777). This explains why Heidi was in the triangulation group and Sharon and I weren’t. We inherited the opposite set of DNA on the paternal side. Also note on my line I have James who represents my Hartley side. I match him from 204-233. So that green segment needs to move to the left a smidgen. I had trouble the first time I tried to figure this out also. After a little fiddling, this is what I come up with:

Chr 1 Segments Adjusted again

In a future blog, I’d like to try the same technique with the DNA results of Joanna and her 2 siblings on the James Frazer line.

Observations:

  • Kathy Johnston has developed an excellent tool for seeing where all your grandparents’ DNA ended up in you and your 2 siblings.
  • Phasing can determine whether a match is on your paternal or maternal side, but requires that you have a tested parent. This technique does not require a parent and goes one step further – to the grandparents.
  • If the procedure is done right, I will know which grandparent all my matches match and which grandparent all my siblings’ matches will match. This is a big help when looking at 3-4,000 matches.
  • The procedure requires other relatives be tested to put names on the 4 grandparents. I am glad to have at least one relative tested to represent each of my 4 grandparents.
  • The technique takes a bit of practice and can be a bit of a puzzle. However, the results are interesting, instructive and informative.
  • The results show the need for additional DNA testing. If I had only tested my sister Sharon and myself, I would not have proof of being in this particular Frazer Triangulation Group on Chromosome 1 going back to the late 1700’s.

December 2016 Update to my January 2016 Blog

The above attempt was an early try at visual phasing. Chromosome 1 is the longest Chromosome and therefore one of the most difficult ones to try to visually phase. Since then, I have had more experience at doing this. M MacNeill [prairielad_genealogy@hotmail.com] also has done some raw data phasing which is more accurate. Here is his result for Chromosome 1:

chr1rawphase

By comparing the two images, you can see the mistakes I made in my early try at visual phasing. There are some things that I see now on my initial try that scream out at me that they are wrong. The major mistake is the small segments that are internal to the Chromosome (i.e. not right at either end).

 

9 thoughts on “atDNA Under the Hood: Segments and Crossovers”

  1. Thanks for this clear post! However, I think there’s a mistake in this sentence: “The green indicates FIR, the green are is HIR and the red is no match.” I think it should be: “The green indicates FIR, the yellow area is HIR, and the red is no match.” (That is, the half identical area is shown in yellow, not green.)

  2. Looking at the figures for Sharon to Heidi, there seems to be a break between 206, 334,201 – 206,376,398. Doesn’t that mean there should be two crossovers relating to Sharon and Heidi in that area? (ie where they stop matching and then restart.)
    Does that help in making sense of the area around 203 – 205 where you currently have the three “J H J” crossovers, but have ignored the one at 205?

    1. Hi, and thanks for the comments. You are right. Since then M MacNeill has done some raw DNA phasing for me. Based on your comment, I have updated my post to reflect that and point out the mistakes I have made. His analysis shows the following Crossovers, Joel (M) 202.4M, Heidi (P) 204.7, Heidi (M) 206.3, Sharon (M) 206.5. I was in way over my head at the time apparently. You are better at this that I was in January and probably better than I am now!

      Joel

  3. Thanks for posting the update, Joel. I think this is a learning process for all of us. (I am just beginning to write up my phasing attempts – but I am starting with the easiest chromosome first 🙂 )

    1. You are welcome. I tend to keep my mistakes up (with corrections, hopefully), as I’m hoping that it is helpful for others to see where I’ve gone wrong.

  4. When visual mapping the X chromosome for three male siblings would you see ONLY three colors representing three of the four gps due to how the X is inherited? I’ve tried visual mapping the X as a start from Blaine’s blog and am missing something.

    1. Hi Karen,

      If I understand you correctly, no. The colors at gedmatch are different than the colors representing the grandparents. You need to come up with your own colors for grandparents.

      I started this process on the other chromosomes before I did the X Chromosome, so I find that one a little confusing. If you have 3 brothers, they all got their X from their mom. So that would only represent 2 grandparents on the mother’s side. If you want to send me what you have or Kit numbers, I can look at them.

      Joel

  5. After that map is, say, 50 percent done, or so, you can go back and add in the next smaller size segments and see if Triangulation among several Matches validates the smaller segments.

Leave a Reply

Your email address will not be published. Required fields are marked *