Total cM versus Average over the Segments

[Mudgeeclarke]

Anyone?

Chromosome Browser. Am I right in thinking that the TOTAL of cM for another match is not as important as the AVERAGE of cM which fall over a number of segments?

For example, the match I have with the highest total cM (60.29 over 22 segments) has an average less that my FTDNA presumed closest cousin, whose 51.59 cM fall over 15 segments.

The higher is shown as 4th to Remote, while the other example is shown as Third cousin.

Both have names and locations which are feasible as connections.

So I'm guessing it is the average, when you see the table and pull it into an excel file. Why isn't an average shown in the chromosome browser?

Am I off track here?

[Mudgeeclarke]

cM in Chromosome Browser - largest, or average over segments?

Can anyone enlighten me, please? Total, or Average ...

[MMaddi]

I've never heard that FTDNA or 23andMe would use the average of shared segments in their calculations to estimate how close a relationship is. The key numbers are largest segment and total amount of shared DNA. Those are the keys to estimating a relationship between matches, not average size of segment.

For instance, if you take the figures you provide in your original post, here's what you get for averages. For 60.29 cM over 22 segments, the average is 2.77 cM. For 51.59 cM over 15 segments, the average is 3.44 cM.

Most genetic genealogists would not not even take any segments less than 5.0 cM into consideration. 23andMe doesn't either. Family Finder counts them, but only when the largest segment is at least 7.7 cM (I think I'm correct about that) and the total cM shared is 20.0 cM.

So, if you're discussing Family Finder, there will be matches with segments counted below 5.0 cM, but only if they meet the other criteria for matching. What are the estimates for relationship that Family Finder is giving you for the two cases?

[Mudgeeclarke]

Hello!

Thanks for replying. The higher number total cM match 60.29 in my quotes example is shown as a 4th-Distant, and the lower number match 51.59 cM is 2nd-4th cousin. (Neither of which I can verify at this stage).

Somewhere, then, I'm getting confused by the just posted tables showing now on the ISOGG Wiki http://www.isogg.org/wiki/Autosomal_DNA_statistics because the numbers there seem to almost exactly equal the averages I derive when I look at the total cM for a FF match, divided by the number of segments where numbers fall. (I see this in segment detail when I look at the table option in Chromosome browser).

So, I must be reading the latest posted Wiki incorrectly, and the averages I get are just coincidentally (and weirdly) almost the same. Is that how you see it?

Many thanks for your interest.

Colin

[MMaddi]

Hello!

Thanks for replying. The higher number total cM match 60.29 in my quotes example is shown as a 4th-Distant, and the lower number match 51.59 cM is 2nd-4th cousin. (Neither of which I can verify at this stage).

Now that's telling you something important. There's a big gap between matches estiamted to be 4th-distant versus 2nd-4th. My guess is that more than one of the 15 segments in the 2nd-4th cousin match are above 5.0 cM. In fact, there are probably more than one above 10.0. How many segments for that match are above 5.0 cM and how many above 10.0 cM?

Somewhere, then, I'm getting confused by the just posted tables showing now on the ISOGG Wiki http://www.isogg.org/wiki/Autosomal_DNA_statistics because the numbers there seem to almost exactly equal the averages I derive when I look at the total cM for a FF match, divided by the number of segments where numbers fall. (I see this in segment detail when I look at the table option in Chromosome browser).

So, I must be reading the latest posted Wiki incorrectly, and the averages I get are just coincidentally (and weirdly) almost the same. Is that how you see it?

Many thanks for your interest.

Colin

I'm sorry, but I don't see anywhere on ISOGG's Wiki page which you linked to that indicate that an average of about 3.0 cM per shared segment is meaningful. Where do you see this?

I'm thinking that perhaps you're referring to the section titled "Ranges of sharing percentage." This table shows percentages used in DNA Relatives in the 23andMe test. Family Finder does not use the same measurement as Family Finder. So, if you're referring to that table, it's not relevant to your Family Finder results.

[Mudgeeclarke]

You're absolutely right ... I was (wrongly) looking at the 23 and Me percentages. Oops. Thanks.

[Mudgeeclarke]

Now that's telling you something important. There's a big gap between matches estiamted to be 4th-distant versus 2nd-4th. My guess is that more than one of the 15 segments in the 2nd-4th cousin match are above 5.0 cM. In fact, there are probably more than one above 10.0. How many segments for that match are above 5.0 cM and how many above 10.0

The 2nd-4th has 4 segments at 3+ cM (none though at 5) and one at 19.87.

The 4th to Distant cousin with the 60 cM total has 6 segments at 3+ cM (none though at 5) and one at 12.1

There is one other 2-4th cousin and he has 3 at 3+ cM (none at 5) and one at 18.39.

BTW, in Table View both 2/4s show as 3rd cousin.

I just need to try and wrap my head around this ... Not easy at 3% Neanderthal ....

Thanks again.

[MMaddi]

The 2nd-4th has 4 segments at 3+ cM (none though at 5) and one at 19.87.

The 4th to Distant cousin with the 60 cM total has 6 segments at 3+ cM (none though at 5) and one at 12.1

You posted in your original post in this thread that the 2nd-4th cousin has 15 shared segments with you, yet you only mention 5 of the segments above - "4 segments at 3+ cM (none though at 5) and one at 19.87. "

Also, that the 4th-distant cousin match has 22 shared segments, yet you only mention 7 of the segments above - "6 segments at 3+ cM (none though at 5) and one at 12.1."

What are the cM ranges for the unspecified 10 segments in the first case and the unspecified 15 segments in the second case? Are the unspecified segments all under 3.0 cM?

[Mudgeeclarke]

... What are the cM ranges for the unspecified 10 segments in the first case and the unspecified 15 segments in the second case? Are the unspecified segments all under 3.0 cM?

Sorry if it was confusing ... Yes, all the other unspecified segments are under 3.0 cM. I was just trying to be brief, and show that apart from the one large segment each, no others were at 5 or higher. I'll see if I can paste in the table so you can see the "where"s.

Cousin A. 2nd to 4th. Table shows as 3rd. (Table won't stay formatted :-( )

Chromosome. Start Location End Location centiMorgans (cM) # of Matching SNPs
2 49442161 51844463 1.64 800
2 85938656 97699722 1.95 679
2. 134283795 138242472 3.35 1000
2 169362301 170916488 2.75 600
3 78891540 82441725 1.11 500
3 85212019 89493009. 1.29 600
5 141867735 143405983 2.35. 600
7 86892272 89170527 2.57 700
8 103926950 106245043 1.51 500
10 21118428 24043573 2.99 600
11 65388838 68173670 3.53 600
12 20668638 21525661 3 600
12 40267993 60521100 18.39. 4800

Cousin B. 2nd to 4th. Table shows as 3rd

Chromosome Start Location End Location centiMorgans (cM) # of Matching SNPs
1 50718114 53551492 1.09 500
3 87272149 96569526 1.08 600
3 130701634. 133141799 1.99 500
5 58608484 62060779 1.87 600
7 98754182 102725120 3.43 900
8 14523775 15689298 2.6 500
8 69355812 70887841 3.07 500
10 105378681 108354793 2.52 800
11 37162511 40048945 1.33 500
11 87132158 88879523 1.93 500
12 21063320 21882887 2.49 500
12 32611195 58236357 19.87. 5600
12 83951822 87831110 3.06 500
15 40899345 43796695 1.76 500
17 59030266 61313804. 3.06 500


These are my only two "3rd Cousins". They cannot be verified by paper trail (yet). (Paternal GGF was NPE so those family names are unknown to me ... (So far).

[Mudgeeclarke]

PS (edit time expired). I did not include the 4th-Distant table, but I can if you needed it.

[MMaddi]

PS (edit time expired). I did not include the 4th-Distant table, but I can if you needed it.

No, I just wanted to confirm that all the segments you didn't mention are below 3.0 cM.

I've tested at 23andMe, but not Family Finder at FTDNA. I'm not used to the idea that segments below 5.0 cM can be considered in looking for matches. So, it's a little shocking to me.

But I suppose that FTDNA has tested out their algorithm and it works in estimating how closely matches are related.