The Importance Of Using Your Arsenal

There are only a handful autosomal DNA testing companies in existence and each company brings something different to the table.  Testers that are actively engaged in genetic genealogy still hope for the holy grail of testing companies.  Often times we have to use a combination of resources to get results.  One classic example of where I had to use 23andme, AncestryDNA, Ancestry.com and Gedmatch to see just how different each company as it pertains to identifying genetic relationships among its testers.  Another peculiar thing that may occur is what I like to call phantom segments. These segments may appear in a child's profile and not the parent's profile. Situations such as these require further analysis that the DNA testing companies do not provide.  The following example demonstrates why you must use your arsenal for analysis:  

May 2015, I received a request to share genomes on 23andme from a DNA match whom I’ll call A.H.  After communicating back and forth via 23andme, I had somewhat of an idea regarding our kinship but wasn’t sure if I was on the right track.  We were able to establish a connection through our respective Askew/Cherry lines.  Her grand uncle and my grandmother’s cousin actually married in the 1950's.  Although we had not been able to establish a genetic connection, we were able to establish a connection through this marriage.  As far as we could ascertain there was no known genetic connection between the two families, although it is highly plausible and highly likely since both families were from the same area and  there were a lot of relationships and intermingling before the turn of the 20^th century between the two families.  A.H. started to wonder if our connection went further back, perhaps as far back as the 1700’s or even the 1600’s.  The only thing we could do was use what we had and analyze the DNA, which brings me back to my original point that you can’t solely rely on one autosomal testing company to connect the dots.  So as I began to dig deeper, this is what I discovered:

According to 23andme, A.H. and my father shared a 16cM’s of DNA on chromosome 13.

I checked my profile and discovered that I received a smaller portion of the segment on chromosome 13.  I also have a segment shared between me and A.H. that is not on my father's profile.  This warrants further investigation as I may also be related to A.H. through my maternal line or the segment may be IBS. Since 23andme ‘s chromosome browser does not display the amount of genetic material shared between two individuals on a chromosome, I had use Gedmatch for further analyzation. 

Please note that 23andme will display the total amount of genetic material shared across all chromosomes; however it will not display individual metrics for each chromosome.  Also if you add the 23++ Google Chrome extension, the extension will convert the segment measurement of Mb (Megabase) to cM (Centimorgan).   

According to Gedmatch, A.H. and I shared 9.5 cM’s on Chromosome 13.  Please note that I used the default threshold of a minimum size segment of 7 cM’s with a minimum threshold value of 700 SNP’s.

That still leaves the phantom segment on Chromosome 16.  The only explanation that I could deduce is that it was either an IBS segment (identical by state) or the segmented was passed down through my mother (IBD), which means I could be related to A.H. through one of my mother’s lines in addition to my paternal lines.   Please visit http://www.isogg.org/wiki/IBS or https://www.familytreedna.com/learn/faq-items/identical-state-ibs/ for an explanation of IBS and IBD.

So in order to get to the bottom of this phantom DNA segment, I felt confident that I should lower the minimum threshold in Gedmatch to 500 SNP’s and 5 cM’s.  5 cM’s seems to be the absolute drop dead minimum threshold that the testing companies use to determine a genetic match within a genealogical timeframe.

I did the same process when comparing my father’s kit against A.H. and determined that the segment shared on Chromosome 16 was probably a combination of IBD and IBS.  It appears that I share 1 cM more of DNA at slightly higher SNP count in regards to this segment.  That  additional 1 cM is probably IBS.  I’m sure there could be an argument made for or against my analysis, but it’s what I could deduce from my understanding and why the segment appeared in 23andme’s genome browser.  If you have tested at least one parent, you can phase your kit to determine if a phantom segment came from another parent or if the segment is IBS.  In a nutshell, phasing allows you to determine if a match is related through your paternal or maternal line, provided that you have tested at least 1 parent.  

Now that I have established where the extra DNA came from, I began to wonder why she did not appear on my match list on AncestryDNA.  A.H. had confirmed that she had tested with AncestryDNA and that she matched a few of my DNA matches on AncestryDNA as well.  So what gives?  If the minimum threshold for the 1^st segment is 5cM’s at 500 SNP’s, then surely we should be a match on AncestryDNA, correct?  The answer is yes, however we are not. I'm pretty sure that the 2014 updates to eliminate IBS matches using aggressive phasing measures may be a contributing factor as to why A.H. appears as a match one one service and not the other.  If she and I had never tested on 23andme or uploaded the raw data onto Gedmatch, we would have never known that we were genetically linked.

So as you can see, the importance of testing and uploading your data to Gedmatch is crucial.  The testing companies all have varying methods in determining DNA matches.  You can't rely on one testing company to help you connect the dots. You have to use everything in your arsenal.