Small chromosome segments

[GenieLogical]

Using Family Finders chromosome browser, I changed the default segment to 1cM & downloaded all the segment data from my matches to a CSV file. There's is a limit of 5 matches, so I combined all into one file. I sorted the entire file by chrom#, beginning seg#, and end seg#. I have a total of 89 matches in Family Finder, and 846 segments.

Matching segments less than 5cM are supposed to be statistically insignificant, but some of the smaller matching segments are hard to ignore:
on chromosome 5, I have 20 different people who match me from 129643803 to 132242229 (2.39cM) and 8 matches on chr10 from 95868647 to 97209409.

Questions: Shouldn't these matches be considered significant based on statistical odds? Do they indicate some racial origin or one common ancestor or geographical area? why are these matching segs area so stable from generation to generation? I am very curious and don't really know enough about the science of genetics to explain this.
Randi Petrone
H3ah

[mixedkid]

What you have noticed I have also noticed, especially with FF cousins who are Finnish or Jewish or French-Canadian. In other words, those common little chromosome segments you mentioned pop up between Finnish people. Others at different locations pop up with Jewish people. Still others appear when comparing French-Canadian cousins and perhaps even Acadian descendants.

[JSW]

Small Chr Segments Real or Random

RE: Questions: Shouldn't these matches be considered significant based on statistical odds? Do they indicate some racial origin or one common ancestor or geographical area?

I have also noticed that and I would like to do some research into just
what these segments are.
I have also noticed them in the Lost Colonies Project and have the
project of looking at them on my to do list.
They might be related to the data used in the deep ancestry tools
or they could be more recent.
It is definitely worth looking into

As a starting point can anyone point me to a few segments
on data in gedmatch.com ?

[lwsmiser]

Significance of small segment matches

I am on both FTDNA and 23andME sites. Some of my matches are at both sites. I have also uploaded my files from both sites to GEDMATCH.com which improves my matching results. I have found that when you have one match greater than 7 cMs the significance of all other matches is increased. These smaller segments are the result of more distant ancestry which have been passed down unchanged for many generations well beyond the reach of paper documents which may be as many as ten generations or more, about 300 to 400 years ago. Unfortunately few lines of descent have information beyond three or four generations.

[GenieLogical]

per JWalden

RE:As a starting point can anyone point me to a few segments
on data in gedmatch.com ?
On GEDMATCH, if you use my kit# fn6135 and the 'Chromosome Segment Comparison Entry Form' using the following parameters: seg start 95000000 seg end 97000000 and min seg size as 2. See all my matches in the same general area (actually 2 areas - 95000000m-97000000, and 93000000-95000000. When I put in random kits numbers, still get many matches, but not with me.
Is there anywhere that explains or identifies the properties of different chromosomes? Such as actual genes being on a certain chromosome, eye color, etc.? Or which chromosome pattern is from an actual geographic area?
That would be a helpful start in sorting all this out.
Regards!
Randi Petrone

[mkdexter]

Is there anywhere that explains or identifies the properties of different chromosomes? Such as actual genes being on a certain chromosome, eye color, etc.?

http://www.ncbi.nlm.nih.gov

http://snpedia.com/index.php/SNPedia

[mkdexter]

Questions: Shouldn't these matches be considered significant based on statistical odds? Do they indicate some racial origin or one common ancestor or geographical area? why are these matching segs area so stable from generation to generation? I am very curious and don't really know enough about the science of genetics to explain this.
Randi Petrone
H3ah

If you are seeing many people matching you at small 1cM areas then most of those 1cM segments could be one of many things but what they probably will end up being are segments from isolated populations somewhere back in your past.

So when you ask are they from a common ancestor or geographical area, yes probably an area in the 1600 and 1700s where many people intermarried because there was no one else to marry for example; an isolated population.

[JSW]

per JWalden

RE:As a starting point can anyone point me to a few segments
on data in gedmatch.com ?
On GEDMATCH, if you use my kit# fn6135 and the 'Chromosome Segment Comparison Entry Form' using the following parameters: seg start 95000000 seg end 97000000 and min seg size as 2. See all my matches in the same general area (actually 2 areas - 95000000m-97000000, and 93000000-95000000.
Randi Petrone

Here is my first round of analysis on the small segments

I started with FN6135 and Chr 10 96m to 97 m and down to 1 cM
I that gave me 88 matches

I then look at the matches for these 4 kits at CHr 10 96m to 97m
F213117 Hatte R Blejer
M425048 Ruth Herbert
F219960 *ARobinson
M162301 Karen Rose

The total for all 5 kits is 292 matches
The number of kits with just one match is 279
The number of matches on small segments to 3,4 or 5 kits is ZERO

The number of matches in common with one other kit is 13

I believe this says that of the 292 marches only 13 proved to
be a (deduced phased) match to one other kit
That is a match rate of 4.4%
because of the way gedmatch uses 23andME to FTDNA matching
there may be some loss of both good and bad matches.

Looking only at matches to FN6135 there are 8 Kits
that match other kits

The question is would full phasing reduce this further.

The way to get a better idea of if any of these are real
relative to kit FN6135 it to actually phase that segment
using these two kits to start with
M425048 Ruth Herbert
F219960 *ARobinson

and to follow that up with a beagle advanced phasing
and see how that changes the matches


One of the matches that came through was
F203568 *Daughter-Mkdexter [email protected] 10 95708027 97353334 1.1 595 FN6135

This must be a random match because there is no match to
either the mother or the father and those two kit are indeed in gedmatch.com

If mkdexter were to use the gedcom phasing tool on kit F203568
That would give me more data on how this segment maps out.

Any thoughts?

[GenieLogical]

small segments

JWalden..The two people you refer to are my mother, Ruth, and my son, A Robinson.

[GenieLogical]

The web site you noted take me to the main pages, but where do I go from there? I'm a novice, obviously, but I'd like to learn more - thanks

[JSW]

Small chromosome segments

JWalden..The two people you refer to are my mother, Ruth, and my son, A Robinson.

You can help with this analysis by phasing your data using your mother's data
(and your father if available)
and also phasing your sons data your data (and his other parent if available)

See

http://ww2.gedmatch.com:8006/autosomal/phase1.php

If you do that today we can use the data late tomorrow afternoon.

[GenieLogical]

To JWalden -
using m201342 as child (me, Randi, from 23andMe) and m452048 as mother (Randi's mother, Ruth Herbert on 23andMe) pat kit=pm201342p1
mat kit=pm201342m1

using f219960 as child (Randi's son, Austin Robinson on FTDNA) and m201342as mother (Randi, Austin Robinson's mother) pat kit =pf219960P1 mat kit= pm219960m1
generated 11:30 pdt

I also have uploaded on gedmatch my mother's 1st cousin f237150 and will have my father's sister in 3 or four weeks.
Thank you -
Randi Herbert Petrone

[JSW]

To JWalden -
using m201342 as child (me, Randi, from 23andMe) and m452048 as mother (Randi's mother, Ruth Herbert on 23andMe) pat kit=pm201342p1
mat kit=pm201342m1
<snip>

On this 2nd round of the analysis using only the data for M201342
That kit has 40 match on Chr 10 of length 1.0 cM to 2.4 cM
all at least touching the 96mb to 97mb segment of chr 10
Of these matches five remained after the partial phasing with one parent
Of those 5 there were two maternal matches and three paternal matches

Thus 90% of the small matches failed after one parent phasing

The two maternal matches do not match each other so there is
no common match on the maternal side

For the three paternal matches
There is a match (an unphased match) between M140302
and M191624 but kit ID M083357 does not match either.

We cannot at this time tell if the 1.2 cM (748 SNP) segment
is common to the three of them if it were to be completely phased.

At this point the small segment matches are not yielding much in
the way of common ancestry and I still do not know why
they show up - what is it that makes them different from other segments?

The analysis will continue.