Longest segments in dna

[AlanBoyd]

There are fewer crossovers during the formation of male gametes than there are for female gametes. It’s those crossover events that chop up the shared segment lengths at each generation. A long segment like the one you are seeing is simply one crossover that “didn’t happen”. If you and your cousin are the children of brothers this could tilt the odds further towards longer shared segments.

[Albufera32]

In terms of genealogy research it shouldn't really matter if 600cM of dna is distributed in long chains across a few segments or short chains across several segments, it's still 600cM of dna.

According to Ancestry themselves, longer segments are characteristic of a closer relative.

For example a DNA match with say two long segments totalling 150 cm, is likely to be a relatively close cousin, whereas a 150 cm match but made up of 8 segments might well be a 4th or 5th cousin but with several links to your tree as a result of cousins marrying, or just the pure coincidence of someone related to both paternal and maternal lines.

[LizzieL]

My match with my 1st cousin's daughter is  546 cM across 16 segments, Longest segment: 115 cM

Match with mother's first cousin: 412 cM across 20 segments, Longest segment: 49 cM

Match with father's first cousin: 248 cM across 14 segments, Longest segment: 31 cM
Match with daughter of above (my 2nd cousin): 128 cM across 8 segments, Longest segment: 32 cM
Total match between mother and daughter to me is half (as expected) but the longest segment is almost identical. So one of the 8 segments the daughter retained must have been the longest of her mother's 14 segments.

Match with another 2nd cousin: 406 cM across 18 segments, Longest segment: 71 cM.
High, but this 2C is the grandson of my grandfather's identical twin brother, so in DNA terms it's as if we were half first cousins.

Not sure of my relationship to this pair, but they are sisters to each other
Sister 1: My match is: 49 cM across 1 segment, Longest segment: 54 cM
Sister 2: 112 cM across 2 segments, Longest segment: 58 cM
And their first cousins (The father's of these two pairs were brothers)- who are brother and sister to each other
Cousin 1:Shared DNA: 34 cM across 1 segment, Longest segment: 37 cM
Cousin 1's brother: Shared DNA: 34 cM across 1 segment, Longest segment: 37 cM
Another first cousin to the above. His mother was sister to the fathers of the above four.
23 cM across 1 segment, Longest segment: 27 cM

All results from Ancestry so some results will be timbered

[Glen in Tinsel Kni]

In terms of genealogy research it shouldn't really matter if 600cM of dna is distributed in long chains across a few segments or short chains across several segments, it's still 600cM of dna.

According to Ancestry themselves, longer segments are characteristic of a closer relative.

For example a DNA match with say two long segments totalling 150 cm, is likely to be a relatively close cousin, whereas a 150 cm match but made up of 8 segments might well be a 4th or 5th cousin but with several links to your tree as a result of cousins marrying, or just the pure coincidence of someone related to both paternal and maternal lines.

But characteristic is no guarantee, 'likely a close cousin but might well be a 4th or 5th', it's not what could be called a definitive answer. As with all things DNA it's down to interpretation by the testing company and the random nature of what dna is passed along and how it is distributed.

Only identical twins/triplets should have identical DNA but even then the results the will differ for each sibling, the same data added to another site will have different results again. There 's plenty of

[Glen in Tinsel Kni]

I did a lot of reading  last night and although l don't understand all of it does make sense that we can get varying amounts of dna coming down the line -
The person who came up with long segment only on Myheritage but l will look at other matches l have on Ancestry & Livingdna

Before you started the thread I requested membership of a forum and it was accepted an hour or two ago. I've done a bit of searching through threads for queries I have but came across this in relation to Ancestry raw data matches added to My Heritage, the initial question asked;

I'm wondering if anyone could advise as to why I match someone on MyHeritage at 34.6cM but somehow don't match them on Ancestry or Gedmatch at all - despite them saying they are on both of these sites?

The reply;
There is an added issue with My Heritage in the way they do imputation if you’ve uploaded there, you’ll match more than you should or match by coincidence.
Ancestry is the most scientific matching algorithm so if they do not match you there then take that as correct.

Further replies mentioned how the cM figure would usually increase by 50 to 100% on MH compared to the Ancestry figure.

[Ruskie]

Seems odd that there would be NO DNA match with someone on Ancestry yet match a reasonably high 34.6 on My Heritage.

I can understand some difference but no match at all? Or is this due to Ancestry deciding that the lower cM matches are useless and removing them, so the match could be below Ancestry’s cut off?

Even if this is the case on Ancestry, surely they would match on Gedmatch because that pulls results from several testing companies. 

Do Ancestry have the “most scientific matching algorithm”? What does that even mean? 

[melba_schmelba]

Seems odd that there would be NO DNA match with someone on Ancestry yet match a reasonably high 34.6 on My Heritage.

I can understand some difference but no match at all? Or is this due to Ancestry deciding that the lower cM matches are useless and removing them, so the match could be below Ancestry’s cut off?

Even if this is the case on Ancestry, surely they would match on Gedmatch because that pulls results from several testing companies. 

Do Ancestry have the “most scientific matching algorithm”? What does that even mean? 

Two things really. Ancestry and MyHeritage use a different chip - the chip MyHeritage uses is a GSA chip, that 23andme and FamilytreeDNA also uses. It is somewhat problematic as it has a low crossover to the chip formerly used by all those companies, but those SNPs (of the old chip) are what form the basis of matching databases such as GEDMATCH (and I think others). Ironically the Ancestry chip is a better match for the old chip than the new GSA chip. After seeing some very dubious results on MyHeritage more and more, even matches just over 40cM, which directly mirrors the very high % false matching I had discovered on GEDMATCH for any GSA chip based kit say below 50cM, I have come to the conclusion that MyHeritage uses the same matching database parameters as GEDMATCH, so when they switched to the new GSA chip, they did not expand their database to match the new SNPs covered, as I previously thought they would have. So there is a rather ridiculous situation that if you do MyHeritage's own DNA test, a large amount of that data analysed is not actually used to match other users and is discarded. So they have to use imputation to try and guess matches, but as far as I can see, this is basically smoke and mirrors and just creates matches that for the most part are not real.
   And the equally ridiculous situation (if my strong suspicions are correct) is that uploading an Ancestry test (or a pre 2019 MH, FTDNA or pre 2017 23andme test) will give you more accurate matching than doing MyHeritage's own test. I suspect FTDNA in fact did expand their database, as did 23andme (although you can't upload to the latter) so matching may be slightly less inaccurate there. I've also seen the same inaccurate matching on Geneanet DNA so I think it also uses the GEDMATCH database and matching algorithms.

[Ruskie]

That sounds complex and confusing but I’m sure you are right melba. 

Your suspicions might be correct. I uploaded from FTDNA to My Heritage quite a few years ago - I find my matches to be fairly accurate. I’m relatively happy with it at least. 

[melba_schmelba]

That sounds complex and confusing but I’m sure you are right melba. 

Your suspicions might be correct. I uploaded from FTDNA to My Heritage quite a few years ago - I find my matches to be fairly accurate. I’m relatively happy with it at least. 

Ah, if it was a pre 2019 FTDNA test, that may well be why Ruskie . All my post 2019 23andme uploads to MH are producing really unlikely matches even in the 40-50cM range (i.e. people of entirely a completely different genealogically traced European ancestry, which is exactly the same pattern I see on those kits on GEDMATCH and Geneanet). For the really family history mad (perhaps certifiable), if you wanted to get the maximum match accuracy on GEDMATCH/MH/Geneanet, you could do a Whole Genome Sequence, with a GEDMATCH compatible file, but it will set you back £400+ .

https://www.rootschat.com/forum/index.php?topic=861091.msg7302573#msg7302573