GEDmatch comparisons

[buckyb]

Greetings
I am trying to get my head around dna comparisons.

My question is this:
I did a Gedmatch Autosomal Comparison between two close cousins,  the comparison results show a match at Chr 21 as, start 9,849,404 end 31,833,963 at 34.5 cM.
A comparison between myself and one of the cousins shows a match of, start 9,849,404 end 15,177,413 at 5.2 cM. However a comparison with the other cousin shows start 13,758,260 end 14,708,142 at 3.4 cM.
If the match comparison between the two cousins shows 9,849,404-31,833,963, why doesn't my comparison show a start point of 9,849,404 for each?
Any help with the above question would be most appreciated, likewise would be any comments on the chances of a dna match. The paper trail suggests a link at 2nd Great Grandfather level.
Many Thanks
Rod

[hurworth]

Hmmm...good question.

So you think they're descended from your gtgt-grandfather, but they have a closer relationship (more recent common ancestor) to each other than to you?  How closely are they related to each other?

[medpat]

From what my DNA links have shown me you shouldn't forget they are getting DNA from both parents in each generation and 1 of the parents will not be related to you. With each generation the link becomes smaller. You do not say whether they are male or female or both, females link with their maternal line and males their paternal line. Variations should be expected.

You don't make it clear but are you suggesting that the 2 cousins are your cousins but not related to each other - or thought not related to each other but you've found a link.

Have you checked their families and go back 1 generation past your idea of the link and if you can't find any factual basis of the link perhaps someone isn't the child of the registered father.

[Spike H]

None of the chromosomes are read from zero. Some start close to zero. Chr 21 does not start until 9,849,404 (start points can be checked by comparing your own kit with itself). I'm sure there is a reason, maybe can't be read or nothing to report.
So your "cousins" start at the earliest point of Chr 21. You still have that start point but at some stage that part of the segment you received from your ancestor in common has been swapped out to be replaced by some other ancestor's DNA. There's no rule that says a match must have exact start and end points, matching segments are said to overlap and that's what they do more often than match exactly.
I don't know how you are defining close. 34.5cM's is not all that close. Second great grandparents  would give third cousins. The average DNA shared by third cousins is 79cM's. 3C2R is 36cM's and 4C is 31cM's. Of course, they may well have matches on other chromosomes that have not been mentioned.
Chr21 is an autosomal chromosome and we get one from each parent. Gender of the cousins would not be relevant in this case.

[buckyb]

Hi All
Many thanks for the replies.
Just to clarify the relationships, the cousins are both female and share gt grandparents. Their total segments>7 cM =166.4cM.
I have no doubt about the relationship of the two cousins only whether I fit in and whether the three of us share common gt gt grandparents. Or as medpat suggests perhaps there is a stranger in the mix.
I have learnt something from Spike H re the starting points of  Chromosones and I thank him for that, it all helps in gaining more information on a complex subject. I'm still not sure why the starting point between the two cousins is the same at 9,849,404 and yet my starting point with one is 9,849,404 and different with the other.
Would a match of the largest segment of 5.2 cM, with a total of segments>4cM= 35.7 together with a paper trail constitute  a match, or is it just another positive towards establishing a match?
Once again thanks to all for your inputs.
Regards
Rod

[hurworth]

It's not uncommon for people with mutual gtgt-grandparents to not match.  And if you don't match them your sibling might.

My mother matches a 4C1R (my mother's gtgt-grandfather was the brother of the match's gtgtgt-grandmother).  My aunt matched him slightly better. The match's aunt tested (so a 4C) and was a better match than her nephew to my mother, but not a match to my aunt.

The matching segment probably is from the speculative gtgt-grandparents, but not enough to prove anything.

Did all three of you test with the same company?  That may influence things as Ancestry no longer uses the same chip as FtDNA and it might explain a misread.  Have you tried lowering the SNPs on one-to-one?

[buckyb]

Thanks Hurford, I'm starting to understand the randomness of DNA.

All three tests were done through Ancestry and I have lowered the SNPs to the minimum level on a one-to-one in GEDmatch. I guess I will leave it at that, perhaps in time some other match will come up to confirm the paper trail.

Thanks again everyone for your inputs.

[Jan_A]

Rod,
 This is how random it is: you can have two siblings do the DNA test and the percentages of ethnics can be different for each sibling.  The sample is a random sampling.  I am still learning how to figure out to what degree (5th, 6th etc) in distance for relations, but because my father never knew his family, I am just happy at this point in FINDING his relatives.

Good luck on your findings/searches!
Jan

[buckyb]

Thanks Jan

It sure is random, I am only just starting to get a grip on it. According to one source I read there is a 50% chance of matching a fourth cousin, while the chances drop to 15% (FTDNA says 10%) at 5th cousin level and less than 5% at 6th cousin level.

We just need to keep looking and hoping for more relatives to get tested.

Best Wishes
Rod