Overlapping Longest Block Segment Matches

[Ericjle]

I entered into a excel spreadsheet all my 226 matches longest block segments. Then I sorted them by chromosome and segment and matched up the overlapping segments.

The first thing I noticed was that the question was not what ones overlapped, but how few did not overlap.

Some of them can be eliminated by having a minimum of say 5 to 7 cM overlap.

The next question is how many of these overlapping segments are circumstantial (False Positive). If you take the total cM of the 22 chromosomes at 3352 cM and divide it by 7 cM = 479. This means that you can have 479 matches (people) without any one overlapping.

If you have a matches with the longest block of 100 cM you can only have 33.5 people without overlapping. Do I have my math right? This look's like too few people and a high degree of circumstance.

My point here is that you can not use the overlapping longest block as the sole criteria for matching. Many of these matches, especially the small segments are going to be circumstantial.

Any thoughts? Any suggestions on how to refine these overlapping segments?

[T E Peterman]

To whom were you comparing your autosomal DNA?

Timothy Peterman

[tomcat]

Why are you discounting overlap of 5 to 7 cM?

Are all four of your grandparents of the same ancestry from the same geography?

Did you test your conclusions via the In Common feature?

[Ericjle]

To whom were you comparing your autosomal DNA?

Not sure what you are asking about. I coped and pasted just my longest block from my Family Finder matches file, using the chromosome browser.

Did you test your conclusions via the In Common feature?

Yes, that's what started me on making a Excel file of my longest block matches. What I found was that there are many more than the "In Common" feature shows.

Are all four of your grandparents of the same ancestry from the same geography?

No. The only ethnic like group is from my great grand mother who has two Colonial lines; one from Connecticut, these are easy to match up because the records are excellent back to the 1600's. The other is Virginia, Virginia in the 1600's are impossible to match and they all married their 1st cousins.

Why are you discounting overlap of 5 to 7 cM?

Let me give you an example. On chromosome #7 I have 18 matches, only one match does NOT have overlapping longest block segments. They are grouped 3 people match, 2 people match, one person match, 2 people match, 3 people match, 7 people match.

Here are the 7 people match data:

7 97,781,645 105,456,869 8.92 1800 5th Cousin - Remote Cousin
7 100,655,364 106,842,546 7.70 1500 5th Cousin - Remote Cousin
7 101,388,981 111,619,752 11.17 2300 4th Cousin - Remote Cousin
7 101,630,597 110,827,173 9.64 2100 5th Cousin - Remote Cousin
7 101,630,597 110,827,173 9.64 2100 5th Cousin - Remote Cousin
7 105,275,566 130,231,616 18.30 4800 3rd Cousin - 5th Cousin
7 105,275,566 130,231,616 18.30 4800 3rd Cousin - 5th Cousin

If you take the last two pairs 110,827,173 minus 105,275,566 = 5,551,607 Divide my 1,000,000 ( 1 cM is about equals to a million SNP's)

Equals about 5.5 cM. If I set the minimum at 7cM those 2 people above do not qualify as having overlapping longest block segments.