PDA

View Full Version : Evidence that very large segments can be passed on through several generations



geebee
09-07-2017, 12:00 PM
I have a 2nd cousin who tested at Ancestry and recently posted his data file to GEDmatch. The exact relationship is that our maternal grandfathers were brothers, and each of the brothers had a daughter. One became the mother of our 2nd cousin; the other became the mother of my siblings and me.

I wasn't at all surprised by the total amount of DNA my siblings and I share with our cousin -- it's a fairly typical amount for 2nd cousins -- but I was somewhat startled to see just how large some of the shared segments were.

Of particular note is our sharing on chromosome 20. I might mention that I've re-checked the genome comparison of the siblings at 23andMe, and the sharing with our 2nd cousin and among the siblings seems to be consistent.

18598

Remember, all of the DNA the siblings share with this 2nd cousin has to have been passed down from our great grandparents on both sides. The two large segments I share could easily have been passed down for year another generation, since my daughter's paternal copy of chromosome 20 is almost entirely from just one of my parents. However, in this case the chromosome I transmitted to her is mostly from my father, so my daughter didn't inherited the segments.

The same thing is true for my niece, who is Bernie's daughter. Her paternal copy of chromosome 20 gives no indication of recombination at all -- but as with my daughter, it's from her grandfather rather than her grandmother.

Still, this should remind anyone who thinks that recombination is some sort of "halving" process that it isn't. Over time it seems likely that such large segments will be broken into smaller segments -- but it clearly can take a few generations.

EDIT: Possibly a different category would have been a better choice for this thread. I put it here simply because of the connection to GEDmatch. If an admin wants to suggest a better location, that would be fine by me.

randwulf
09-07-2017, 09:29 PM
The husband of my second cousin (neither of whom I have met) loaded her results to Gedmatch and we were able quickly to find the common ancestry. We have some pretty good matches:

Minimum threshold size to be included in total = 500 SNPs
Mismatch-bunching Limit = 250 SNPs
Minimum segment cM to be included in total = 7.0 cM


Chr Start Location End Location Centimorgans (cM) SNPs
3 65,848,146 72,193,268 8.2 1,004
4 61,566 5,642,349 9.1 745
6 105,175,781 154,504,070 53.1 7,056
8 3,812,181 21,256,652 31.6 4,540
13 18,178,540 43,540,464 47.2 4,878
14 95,729,403 100,866,880 14.2 1,112
15 95,353,461 100,198,883 14.6 979
17 14,764,347 34,309,341 21.8 2,260
18 27,281,399 43,779,618 13.1 2,084
Largest segment = 53.1 cM
Total of segments > 7 cM = 213.0 cM
9 matching segments
Estimated number of generations to MRCA = 3.0

Also, I have my father tested - so the distance being first cousin, once removed - it is interesting to see the much greater matching:

Minimum threshold size to be included in total = 500 SNPs
Mismatch-bunching Limit = 250 SNPs
Minimum segment cM to be included in total = 7.0 cM


Chr Start Location End Location Centimorgans (cM) SNPs
1 29,740,264 55,272,794 27.8 3,245
1 224,776,496 237,991,428 25.0 2,451
2 20,583,718 28,609,759 7.4 880
2 29,639,183 103,946,935 66.0 9,557
3 65,308,093 136,175,742 55.7 7,819
3 175,901,232 195,155,128 34.3 3,025
4 61,566 57,613,002 74.5 7,876
4 85,496,814 126,113,552 32.1 4,404
5 41,303,085 101,146,382 47.7 6,391
5 135,232,585 151,270,918 19.1 2,331
6 104,337,508 154,504,070 53.8 7,168
8 3,905,526 22,119,383 33.3 4,676
12 43,842,926 52,671,866 9.2 1,197
13 19,808,048 43,540,464 46.1 4,701
14 80,590,013 100,968,993 38.0 3,817
15 95,353,461 100,198,883 14.6 981
17 14,768,612 34,309,341 21.8 2,256
18 27,322,066 43,779,618 13.1 2,081
Largest segment = 74.5 cM
Total of segments > 7 cM = 619.3 cM
18 matching segments
Estimated number of generations to MRCA = 2.3

Another interesting thing is that she matched my father on a large, not-very-Western-European segment (and a couple of others like it), which has allowed me to narrow my search to one-half my family tree for the ancestor.

We can't quite top your 80+ cM match, though!

geebee
09-08-2017, 12:44 AM
Thanks for the response.

I've seen genome comparisons in which entire chromosomes have escaped recombination for more than one generation, so the occasional huge shared segment even beyond parents or siblings shouldn't be too surprising. We speak of recombination as being random -- and it is, more or less -- but there are some "conditions" on that randomness.

One of the most important is the variability of the crossover rate. For one thing, on average men have a much lower rate than women do. This is off the top of my head, but I think the average is close to 27 for men and 42 or 43 for women. This means that the chromosomes a father passes on to his children are likely to consist of segments that are both fewer and larger. This can affect how close to equal the DNA inherited from each grandparent will be.

I remember reading about someone who didn't inherit much more than 15% of her (?) DNA from one of her father's parents, which also meant she inherited 35% from his other parent. Between my parents and my daughter, the split it more like 22.5% from my father and 27.5% from my mother.

JMcB
09-08-2017, 03:43 AM
Thanks for the response.

I've seen genome comparisons in which entire chromosomes have escaped recombination for more than one generation, so the occasional huge shared segment even beyond parents or siblings shouldn't be too surprising. We speak of recombination as being random -- and it is, more or less -- but there are some "conditions" on that randomness ..

Speaking of the variability of recombination, I recently came across what seemed to be a fairly large segment between my 3rd cousin once removed and myself. According to Ancestry we share a segment that's 94 cM long, out of 100 cM total. FTDNA says it's 90 cM long, out of 104 cM total. Our relationship goes back to my 2cd Great Grandfather (her 3rd) who was born in 1799 and died in 1877.

At any rate, not being well versed in these things, I meant to post a question about it but never got around to it. Plus, after looking around for awhile, I figured it was just a slight anomaly in a process that is prone to such things.

geebee
09-08-2017, 02:14 PM
Well, as I've been saying -- or at least trying to say -- whenever a chromosome is passed on with no significant recombination, the chance that the same thing will happen in the next generation remains the same.

It many ways it's like a coin flip. Just because you flip a coin once and get "heads", that doesn't make it more or less likely that you'll get "heads" on another flip. As an aside, when you're doing coin flips, there's a stipulation that you're using a "fair" coin. If you know what you're doing, how you flip actually can affect the outcome.

Recombination is supposed to be a more-or-less random process, but there are "hot spots" and "cold spots" on chromosomes -- places where a crossover is either more, or less, likely to occur. In addition, just as the average crossover rate varies significantly between the sexes, it also varies among different individuals.

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0020321

Dewsloth
09-08-2017, 04:25 PM
Well, as I've been saying -- or at least trying to say -- whenever a chromosome is passed on with no significant recombination, the chance that the same thing will happen in the next generation remains the same.

It many ways it's like a coin flip. Just because you flip a coin once and get "heads", that doesn't make it more or less likely that you'll get "heads" on another flip. As an aside, when you're doing coin flips, there's a stipulation that you're using a "fair" coin. If you know what you're doing, how you flip actually can affect the outcome.

Recombination is supposed to be a more-or-less random process, but there are "hot spots" and "cold spots" on chromosomes -- places where a crossover is either more, or less, likely to occur. In addition, just as the average crossover rate varies significantly between the sexes, it also varies among different individuals.

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0020321

One for my dad and me -- we split off from these folks in the chromosome browser around 1750
(the dad was Valentine Cook, a probable English colonist and the mother was Susannah Baughmann born 1732 in Zurich; they had a bunch of kids):

18617

^Largest contiguous block is 39.3cM; the two blues are 13.7cM and 27.4cM


Edit:
Also, here's a woman who might be descended from a sibling of my 2nd Great Grandfather:
18618
54.9cM, 13.2cM and 14.6cm all on the same chromosome.

geebee
09-08-2017, 07:52 PM
This is exactly the sort of thing I was hoping to see.

There's another poster here who is convinced that it's virtually impossible for large segments to be passed on for several generations, and that any apparent such transmission must really be some sort of "reconstruction" from smaller segments passed along multiple lines of descent from the common ancestor.

I wouldn't say the mechanism is impossible, but that it is both improbable and unnecessary. While pedigree collapse does exist, in most cases it isn't recent enough for there really to be that many pathways from any common ancestors. It's also easy to show that entire chromosomes are often transmitted for one or two generations with virtually no recombination. So one would certainly anticipate that on occasion this can happen over an even greater number of generations -- as a number of folks can now attest.

If anyone has any screenshots of genome comparisons for even three (or four?) generations, please post them. Meanwhile, I have one that shows my father and two granddaughters (1st cousins to each other):

18624

The one shown in purple is my daughter; the orange in my brother's daughter. My daughter shares 1673 cM in 24 segments with my father, while it's 1887 cM 25 segments for my niece.

This next graphic is a comparison of the two cousins to each other. Neither one shares anything on chromosome 17 with their grandfather, so their paternal copies of these chromosomes must have come from their grandmother. They share nearly all of the chromosome with each other -- except for a small break which may or may not be real. (The break shown between my daughter and her grandfather on chromosome 9 does not exist, as I've been able to confirm using 3rd party tools.)

18625

geebee
12-03-2017, 10:46 PM
I'll have to see if I can track it down, but I remember Dr. Ann Cousins posting somewhere to the effect of there being an entire chromosome in her family that was passed on without apparent recombination* for three or more generations. I can't recall whether this was an X chromosome, but I think it might actually have been an autosome.

*I say "apparent" recombination because in fact recombination does occur. However, in this case it occurs between sister chromatids . Because the crossover(s) involve(s) chromatids which were identical to begin with, their product is also identical. (See https://en.wikipedia.org/wiki/Sister_chromatid_exchange.)

EDIT:

Found it here: https://dna-explained.com/2014/01/23/that-unruly-x-chromosome-that-is/, in a comment dated May 7, 2015 at 2:54 pm. The chromosome in question is chromosome 4, which is actually relative large, and she says it was "inherited intact through at least five and maybe six transmission events". (Which doesn't necessarily mean that number of generations, since she might be talking about a chromosome passed down to multiple offspring.)

This clearly says that very large segments can be transmitted through numerous generations, since entire chromosomes can sometimes be passed several generations without any -- or with very little -- recombination.

Each time a chromosome "escapes" apparent recombination, there is absolutely no disminuition of any matching segments on that chromosome. The best illustration in my family is between my daughter and my oldest brother's daughter. The two of them are very nearly identical on chromosome 17.

Since my brother and I are also identical on our maternal copies of this chromosome (and nearly identical on our paternal copies), it seems likely that this is basically an unrecombined copy from our mother -- matching either her own maternal or paternal copy, but not being a recombination of both.

If so, this would represent four recombination events without a detectable crossover on chromosome 17.

2nd EDIT:

I might note that, so far, all shared segments for either my brother or me on our maternal copy of chromosome 17 appear to trace to our maternal grandfather -- or rather, to ancestors on his side.