Go Back   Family Tree DNA Forums > Universal Lineage Testing (Autosomal DNA) > Family Finder Advanced Topics

Family Finder Advanced Topics Advanced discussion about Family Tree DNA's Family Finder Product.

Reply
 
Thread Tools Display Modes
  #1  
Old 19th March 2018, 09:50 PM
mickreed mickreed is offline
FTDNA Customer
 
Join Date: Oct 2014
Posts: 12
Full-Sibling shared DNA

Hi,

I know this has been covered before but I am still not quite clear. If I understand correctly:

1. Full siblings share around 3400 cMs of their autosomal DNA

2. FTDNA typically reports a much lower figure around 2500-2600 or so. Me and my sister share 2573 cMs according to FTDNA. I call this 'under-reporting' of the actual figure, but this may be wrong of me.

3. The reason for this discrepancy is the way that FTDNA handles (or at least reports) fully-identical regions.

I have asked FTDNA support about this, but they say that this is proprietary information and so cannot elaborate. Fair enough.

So, what I'd like to know is:

1. How can I work out what the real figure for shared cMs is?

2. What is the rationale behind this 'under-reporting'? I assume that FTDNA's analysis reveals the actual figure and that it is simply a question of reporting.

Any help welcome.
Reply With Quote
  #2  
Old 19th March 2018, 10:33 PM
bhemph bhemph is offline
FTDNA Customer
 
Join Date: Feb 2018
Posts: 6
FTDNA tests about 3400 cM, so a match of 3400 cM is either a parent, child, or identical twin. The 2600ish cM is a sibling who is not an identical twin.
Reply With Quote
  #3  
Old 19th March 2018, 11:04 PM
jimbirk jimbirk is offline
FTDNA Customer
 
Join Date: Apr 2017
Posts: 40
mickreed:

It all has to do with how full match DNA is reported.

Read the section "Distribution of shared DNA for given relationships" in the ISOGG write up on "Autosomal DNA statistics".

https://isogg.org/wiki/Autosomal_DNA_statistics
Reply With Quote
  #4  
Old 20th March 2018, 08:22 AM
josh w. josh w. is offline
FTDNA Customer
 
Join Date: Mar 2005
Posts: 2,523
Quote:
Originally Posted by mickreed View Post
Hi,

I know this has been covered before but I am still not quite clear. If I understand correctly:

1. Full siblings share around 3400 cMs of their autosomal DNA

2. FTDNA typically reports a much lower figure around 2500-2600 or so. Me and my sister share 2573 cMs according to FTDNA. I call this 'under-reporting' of the actual figure, but this may be wrong of me.

3. The reason for this discrepancy is the way that FTDNA handles (or at least reports) fully-identical regions.

I have asked FTDNA support about this, but they say that this is proprietary information and so cannot elaborate. Fair enough.

So, what I'd like to know is:

1. How can I work out what the real figure for shared cMs is?

2. What is the rationale behind this 'under-reporting'? I assume that FTDNA's analysis reveals the actual figure and that it is simply a question of reporting.

Any help welcome.
I am not sure that there was any under-reporting, The figures reflect average findings which may be viewed as what to expect or 'theoretical'. Your 'actual' mixture with your sister may be above or below the average. It is a matter of chance.
Reply With Quote
  #5  
Old 20th March 2018, 08:52 AM
John McCoy John McCoy is offline
FTDNA Customer
 
Join Date: Nov 2013
Posts: 629
If you did not test with a vendor that adjusts the cM count for "fully identical regions", you can upload the raw DNA data files to GEDmatch and then use the One-to-One comparison tool there, with the graphics option. The "fully identical regions" will show up in green, and you can, if you wish, interpolate as necessary and get a good estimate of the extent of the fully identical segments, and add that figure to the reported shared cM (which counted all matching segments as half-identical) to obtain the "correct" total.

The same tool, with the graphics option, is very helpful for the perennial question of how to spot half-siblings.

For all the vendors and web sites, reporting a separate count of fully identical segments would be a nice improvement. It would give everyone a clearer view of how genetics works and what the matches mean. Also, it would be a good excuse for the web developers and marketing specialists to design pretty graphics.
Reply With Quote
  #6  
Old 20th March 2018, 11:22 AM
Frederator Frederator is offline
FTDNA Customer
 
Join Date: Jul 2010
Posts: 801
Quote:
Originally Posted by mickreed View Post
Hi,

I know this has been covered before but I am still not quite clear. If I understand correctly:

1. Full siblings share around 3400 cMs of their autosomal DNA

2. FTDNA typically reports a much lower figure around 2500-2600 or so. Me and my sister share 2573 cMs according to FTDNA. I call this 'under-reporting' of the actual figure, but this may be wrong of me.

3. The reason for this discrepancy is the way that FTDNA handles (or at least reports) fully-identical regions.

I have asked FTDNA support about this, but they say that this is proprietary information and so cannot elaborate. Fair enough.

So, what I'd like to know is:

1. How can I work out what the real figure for shared cMs is?

2. What is the rationale behind this 'under-reporting'? I assume that FTDNA's analysis reveals the actual figure and that it is simply a question of reporting.

Any help welcome.
As someone else posted earlier, I think the most useful basic guidance on this topic are the figures posted by ISOGG.

https://isogg.org/wiki/Autosomal_DNA_statistics

But I've given quite a bit of thought to this myself because I'm trying to create as complete a reconstruction as possible of one of my parents' DNA. To that end, I've done a rough calculation of the probability of several sharing outcomes among full siblings.

As somebody else mentioned, the basic sharing figures for siblings, like everyone else, are reported as if they were half-identical, even though in those regions where both siblings inherited the same segments from both parents, they will be fully-identical. So the totals I report below will follow this standard convention, reporting as if they were all half-identical.

The process really is completely random, so the basic calculation is something like:

[A/2^2]/2+[A/2^2]=total cM expected to be shared by 2 siblings

"A" represents the total # of cM tested on a given platform. Judging by my own results and those of my siblings, Family Finder must currently test about 6,678 cM total (i.e., we each report sharing 3,384 with our mother). I think other companies test a slightly different amount. Statistically unimportant lab discrepancies might even result in different individuals on the same platform sharing slightly different amounts with a parent, but the variation should be trivial.

So the basic prediction is that two siblings at Family Finder should expect a reported 2,538 cM shared. Remember, though, as was mentioned earlier, some of these are actually fully-identical, not half-identical segments.

But like I said, the process is completely random. It is technically possible--although extremely unlikely--that two siblings could share 0 cM. I mean, it's ridiculously unlikely. So unlikely that Excel had to return the odds out to 100 digits in scientific notation. But technically possible.

I'm not really sure the best way to calculate the odds of any one sharing scenario. It would all hinge on what the absolute minimum size of a completely reliable shared segment would be.

The industry standard minimum for reporting matches seems to be 7 cM, although as people have discussed elsewhere on these forums, the validity of a segment should really be measured in SNPs and not cM, and numerous false matches do appear to be reported at the standard minimum level.

If I had to guess, though, my personal experience tells me the minimum safe segment size would be around 14 cM. I plugged this into Excel's non-cumulative binomial distribution formula (i.e., 3,400 cM/14 cM=242.86 successes; 6,786 cM/14 cM=484.71 cM trials; 2,583 cM / 6,786 cM=38% probability of success).

So in my crude calculation,the odds of the case you report could be 1 in 47,399,068. Pretty rare. But obviously still within the range of theoretical possibility.

If you're concerned that there is something more fundamentally wrong with your results, I agree with the poster who suggested that you download your raw data to Gedmatch where you can use their enhanced chromosome browser options to determine who proportion of the shared segments are fully identical. I'd expect it to be around 1/3 segments for full siblings to be fully identical.

But I'm guessing this is just an isolated case falling within tails of the normal probability curve.

Last edited by Frederator; 20th March 2018 at 11:36 AM.
Reply With Quote
  #7  
Old 20th March 2018, 01:41 PM
ltd-jean-pull ltd-jean-pull is offline
FTDNA Customer
 
Join Date: Jun 2016
Posts: 477
This might be too simple an explanation, and I may have misunderstood, but it appears to me that the way the amount if shared DNA is calculated by FtDNA is by taking 3400cM and subtracting the unshared segments.

This doesn't account for fully identical segments having twice as much shared DNA.
Reply With Quote
  #8  
Old 20th March 2018, 01:58 PM
Frederator Frederator is offline
FTDNA Customer
 
Join Date: Jul 2010
Posts: 801
Yeah, I guess I misunderstood the original post, too. It seems to me that the author and their sibling are almost at the dead center of the probability curve. Just 35 cM over the average.

But, as has been said, that treats fully-identical segments and half-identical segments alike.

If you wanted to discover the true amount, reporting an additional share for each fully-identical segment, you'd really need to go to Gedmatch and use their enhanced tools to find the precise #.

On a population-average, though, full siblings should expect to be fully-identical on about 1/3 of the reported segments, bringing the grand total to around 3,384 cM, if the sharing of me and my 3 siblings accurately relfects the full volume tested by FTDNA.

Which is what I believe is exactly 1/2 of the total cM tested.
Reply With Quote
Reply

Bookmarks


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Half-Sibling versus Full-Sibling andyfoulk Family Finder Advanced Topics 14 28th September 2017 11:07 AM
Q about half vs. full sibling (sister and brother) pteg Adoptees Forum 8 21st September 2015 02:18 PM
full sibling for sure? MEKoc Adoptees Forum 23 27th July 2015 11:49 AM
Sibling relationship based on longest segment cM and total shared cM emyr Family Finder Advanced Topics 4 20th July 2012 09:10 PM
sibling shared matches LSSmith Family Finder Advanced Topics 2 3rd August 2011 12:13 PM


All times are GMT -5. The time now is 08:49 PM.


Family Tree DNA - World Headquarters

1445 North Loop West, Suite 820
Houston, Texas 77008, USA

Phone: (713) 868-1438 | Fax: (832) 201-7147
Copyright 2001-2010 Genealogy by Genetics, Ltd.
Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.