Genealogy Help & Advice

                 AUTOSOMAL DNA TESTS
So, you are intrigued by the TV commercials for DNA testing, and you’ve taken the plunge.  Your results are back but you feel like you know less now than when you first took the test.  Before you can understand your results, you need to know some basics about DNA tesTESTting and you need to know what is realistic. 

FIRST What is an “autosomal DNA test?” 

A human’s  genome consists of twenty-four chromosomes.   Chromosomes come in pairs, two each of chromosomes number 1 through  22 (one ½ of each came from your father and the other ½ from your mother).  The we all have two “sex chromosomes” (X- chromosome #23 & #24  in females and Y – chromosome #24 in males ) plus twenty-two autosomal chromosomes.  Females have two X chromosomes (X X) and males have an X and a Y (XY).  Chromosomes 1-22 are the AUTOSOMES. 

An autosomal DNA test analyzes chromosomes 1-22. 

Now, follow me – As I stated earlier, you “get,” or inherit ROUGHLY ½ your autosomal-DNA from your mother and ½ from your father –BUT IT DOESN’T END THERE.  Your parents get ½ from both of their two parents, which means you get ¼ from each of your 4 grandparents.  This pattern continues with 1/8 from each of your 8 grandparents, 1/16 from each of your great grandparents, 1/32 from your great great grandparents and on-and on.  With each successive generation the portion you inherit from any one particular ancestor is halved.  In other words, we are a mosaic of our ancestors. 

SECOND What does this reveal?

RELATIVES/RELATIONS

Once your results are analyzed, they are compared to the results others who have tested looking for segments (the area on any given chromosomes measured in cM or centiMorgans) that match.  The larger the matching segments, the closer you two are related.   The further back in time your common ancestor lived, the small the piece you inherit from him/her.  Theoretically this could take you back to someone with a common ancestor that lived many hundreds of years ago, but practically this won’t work.  We reach a point where the size of the common segment is so miniscule that we cannot separate a matching segment from “genetic-static” or “genetic noise.”  This point is basically at 5 to 8 generations back, meaning  you two are basically 5^th cousins with a common 4^th-great grandparent.  At this level you two have about 7 cMs out of the roughly 3600 cMs, or  0.0194444%  of the human genome. 

Below is a chart the shows your estimated relationship with another who shares a given amount of cMs. 

An autosomal test is great for telling HOW two individuals are related – BUT  an autosomal test cannot easily tell you WHO your common ancestor was.  The common ancestor could be from any branch of your tree.  The only way an autosomal test can confirm who was your common ancestor is for you to speculate who your common ancestor was then find one of his/her descendants, have THEM do an autosomal test and see if you have any common DNA segments.

Traditionally when you take an autosomal test your results are a list of people with whom you share cM segments and how large those segments are, BUT NOT FROM WHOM THEY DESCEND.

Another little glitch is that a pair of siblings do not necessarily inherit the same DNA from their parents.  You may have inherited parts A,B,C,D,E & F while your sibling may have inherited parts B, C, D, F, G & H. 

THIRD: ETHNIC MAKE-UP

Can an autosomal test ACCURATELY identify your ethnic heritage?

Be honest – most people purchase an autosomal DNA test with the goal of learning “What they are,” (English, Irish, Native American, German, etc).  Unfortunately, this is the weakest feature of ANY autosomal DNA test.  It has been the nature of humans since long before we kept records be nomadic and to intermarry with those they met along the way.  Because of this, none of us are purely from one group / ethnically un-mixed.  SOOOoooo, how do the testing companies come up with your ethnic background?  They make an “educated guess” based on comparing your DNA to the DNA of a set of control-populations. They do their best to form their control groups from a group of Log-term residents whose ancestors traditionally didn’t move.  Unfortunately, they cannot guarantee that the ancestors of those in this control group were not mobile in the last millennium or longer.  That they were not invaded.  That they were not the result of intercultural mixing. 

These control populations are usually very small compared the total population.  Any statistician worth their salt knows that a study based on to too small of a sampling is NOT a valid study. 

I personally have tested with four different companies, and had my downloaded raw results analyzed by several more companies.  Note the disparity between companies.  (Unless and until DNA testing companies can standardize the results from all companies, ETHNICITY ESTIMATIONS SHOULD BE TAKEN ONLY AS A GUESS AND NOT CONSIDERED ACCURATE.