exploring DNA matches by total cM and cM in the longest strands #dna


In screening DNA matches for close relatives ( ~ second cousins),  I typically pursue matches with total DNA of at least 125 cM along with longest single strands of at least 30 cM.  I often come across identified matches of much less than 125 cM total, say 80 cM, but with longest strands of say 20 or 25 cM.  In other words, with an unusually high percent of matching DNA in the longest strand.  Can anyone enlighten me as to whether these matches (characterized by relatively low total cM but high cM in the longest strand as a percent of the total) are also worth pursuing?

Steve Diamond
New York City


I think you are using unrealistically high numbers for longest single strand and total DNA. I only have three matches out of the many thousands that have more than 130 cM total and only two those have single strands larger than 30 cM. I have found 4th cousins that share as little as 24 cM total with as little as 10 cM largest single strand.

I think the most important thing is to have a well flushed out paper trail, a tree that includes not only direct line ancestors but also as many uncles and aunts as can be found and all their lines. This allows for things like name recognition in matches to target for research.
David Belton


MyHeritage frequently sends me a match to someone with 155.5cM across 12 segments, longest 41cMs. He is also matched to one of my  known 1st cousins once removed but we cannot find any real connection to him. None of his ancestor names are known to me or my cousin. 
Any suggestions please?
Alan Cohen

Steven Turner

You raise an interesting question. If you had to pick one value that would be most important for Ashkenazi Jews I believe it would be the longest segment. However if you are going to dismiss all matches less than 125cM total you might be missing out. If you or your match is 100% Ashkenazi this could work for you but if you have a match that is let's say 50% Jewish an 80cM total would be significant because there is less endogamy in play.

The one rule is that there are really no rules. Each match needs to be judged on its merits. If you had a match with a person - 30cM longest and 80cM total you should find out what % Jewish they are and then proceed accordingly.

Good luck.

Steven S. Turner
President, Gesher Galicia

Bob Smiley

The problem for investigating relationships through DNA for Jewish ancestry, is the problem of endogamy. Your ancestors married others of the same faith who lived nearby, meaning that the same DNA was being passed back and forth for generations. You may match someone but would have to go back 10 generations and then find out that you were related in three different ways. The hallmark of endogamy is having a large number of very small (less than 7 cM) matching segments which together make up a large number.

For Ancestry, which does not have a chromosome browser, I look not at the total or largest fragment, but look at the average size (total divided by the number of segments). So, for someone that matches you at 125 cM, but has 15 segments, that means that the average size of your matches is only 8 1/3. Also, a single large segment can be passed (by happenstance) over many generations, so if you remove that segment from the total and then calculate the remainder as averaging less than 10 cM, that is also a match who is likely to be related many generations back.

A useful resource is the Facebook group "Jewish DNA for Genetic Genealogy" at https://www.facebook.com/groups/DNAHelpJewishJourney

They have a files section that includes many explanations on endogamy and how to properly look at matches.

Personally, unless the average size is over 10 cM, I will not look too closely at a match. The exception is if they have a large tree and I find a possible link.
Bob Smiley
Kirkland, Washington USA

Joe Bratspis

Among the hundreds of suggested matches provided by the DNA testing companies, I have found two distant cousins; one with a total 27 cM match and the other with a total 11 cM match. These matches were confirmed, not because of the DNA result, but because each of the matches had provided their Family Tree to the testing company. Why would someone, interested in their genealogy, have their DNA tested but not provide a Family Tree to help in confirming a relationship?

Joe Bratspis
Weston, Florida, USA   

Robert Hanna

Perhaps he was adopted.  Or his father is not his biological father.  There are many reasons this can happen.  Have you been in contact with him?  Perhaps he has the answer.

Robert Hanna


JoAnne Goldberg

After ten years of looking at DNA relationships and trying to figure out
which are real and which random endogamy, there is no single answer DNA
is certainly useful for finding closer relatives, though even third
cousins may share no DNA.

My general rule of thumb for investigating mystery matches is that you
need at least one segment of 30cM and another of 20cM. I have random
matches where I share one segment longer than that, and it's clearly not
a close match. With my biggest find -- since corroborated by records --
the overall total was only about 100, but there were matching segments
of 50cM and 25cM.

The advice to build out the family tree is great if you've got ancestry
from Germany or the UK or somewhere else where they had decent records.
Because those records go way back, I can find fifth/sixth cousins on my
German side with whom I share only a tiny amount of DNA (on Ancestry).
Whether that shared DNA comes through our known common ancestor or
elsewhere, impossible to tell.

Building out the family tree with ancestors in other places where
records are sparse -- good luck. People changed names, they moved
around, they had kids with someone who wasn't their known spouse. My big
Lithuanian finds have emerged through family documents such as letters
and photos.. Some of the oldest living relatives have documents and
photos that even their kids didn't know about until someone asked a
direct question!
JoAnne Goldberg - Menlo Park, California; GEDmatch M131535


Moishe Miller

Dear Group,

I agree with David Belton's comments. I would add that endogamy aside, there is  a very good tool for determining the potential relationship. See The Shared cM Project 4.0 tool at https://dnapainter.com/tools/sharedcmv4 (free). That will give you an idea of the range of relationship possibilities. 

Using your example of 24cM, the Tool gives a 
57%6C 6C1R 5C 6C2R 4C1R 5C1R 7C Half 3C2R 4C2R 5C2R 7C1R 3C3R 4C3R 5C3R 8C or more distant

and if it is a real 4c, then the likelihood of a match of just 24cM is only 16%

For a 2c, the range is 41cM-492cM, with 229cM being a mean number. 

It is great that you share your techniques and that we can have dialog about it. 

Moishe Miller
Brooklyn, NY
JGFF #3391

Lee Jaffe

One thing often lost in these discussions is that these metrics are meant to prioritize your DNA matches, to help you identify those most likely to be identified and located on your tree.  Hitting some tried-and-true threshold for total cM or longest segment isn't a guarantee that you'll be able to place someone in your tree and falling below those recommended numbers doesn't mean that the match isn't traceable.  I think that those who teach or promote such strategies are trying to help folks – especially newbies – to not be overwhelmed by large numbers of matches many of us have.  Unfortunately, too often the effort to get people to take these schemes seriously get them set in concrete.   Use these numbers to identify the matches most likely to fall within traceable family history.  But don't let others tell you to dismiss an interesting match just because it falls below their recommended thresholds.  You can always pursue these as 2nd or 3rd priority cases, but with correspondingly lower expectations.  

Lee David Jaffe
Surnames / Towns:  Jaffe / Suchowola, Poland ; Stein (Sztejnsapir) / Bialystok and Rajgrod ; Roterozen / Rajgrod ; Joroff (Jaroff, Zarov) / Chernigov, Ukraine ; Schwartz (Schwarzstein) / Ternivka, Ukraine ;  Weinblatt / Brooklyn, Perth Amboy, NJ ; Koshkin / Snovsk, Ukraine ; Rappoport / ? ; Braun / Wizajny, Suwalki,  Ludwinowski / Wizajny, Suwalki


EA Wurster

It helps to set up a non-destructive work flow to deal with the long DNA match lists, like from MyHeritage (used as example below).

  1. I start with that file in Excel, and add an empty column, say after column, Largest .
  2. Enter a formula using column references: Total cM shared / Number of shared segments (you'll have to discover the references on your sheet).
  3. Hit enter and select that cell with the formula. Double click the fill handle to fill down the entire column with the formula. Good idea to name the column, perhaps "Ratio".
  4. Apply Data Filter to all column heads.
  5. Click on the filter arrow for this new column, and enter a Number Filter, Greater Than = 9.9 (or something else you decide).
  6. Click on the Largest Seg filter arrow and enter a Number Filter, Greater Than = 29.9 (or something else you decide).
With these settings I now see just 67 matches, which is easier to work with than the 16,000 records I have in the beginning.
Of course this may temporarily hide some matches you need, but there are ways to overcome that. 
Ed Wurster
Voorhees, NJ
Leider (Leader) | Samowitz (Samuels)