DNA matching: Half siblings

My DNA quests continue.  My mother's half brother was kind enough to submit his DNA specimen to 23andMe.com.  I am busy analyzing the results.  They share 26.6% of their DNA.  Full siblings share about 50%.  To interpret, each of them has about 75% of unshared DNA.  25% is from their mother, and 50% from their respective fathers.  If a genetic distant cousin matches one half sibling and not the other, we still do not know which parent holds the relation.  But if a genetic distant cousin matches both half siblings, then the relation is through their common parent.  Here is what my mother's comparison to her brother looks like.

23andMe.com
Autosomal DNA comparison of half-siblings (sister and brother sharing mother)

My uncle matches me 14.2%.  This is roughly half of the match between him and my mother.  This is because approximately half the DNA is lost from one generation to the next.

23andMe.com
Comparison between my mother, her half brother, and me.

This information is very valuable because anyone who matches both my mother and her brother matches on my maternal grandmother's side of the family.  The dark blue areas in the above graph are areas of my DNA that are identical to my maternal grandmother.  You can see how some segments break into tiny pieces quickly, while other stayed intact through a few generations.  The blank areas contain some more of my maternal grandmother's DNA along with my paternal grandfather's DNA.

The difficulty with identifying the common ancestor of genetic distant cousins is that you have two possible paths to follow at each generation as you travel back.  I know which parent to credit my matches to because both of them are in the database.  But beyond them, I do not know which of their parents share the relation, and so on.  At least now I have a better chance of identifying common ancestors by determining which ones match my maternal grandmother.  I will be explaining more and sharing examples in upcoming posts.

Roots into the Future, part two

Last year I received a free autosomal DNA testing kit from 23andMe.com through their Roots into the Future program.  According to their website, the purpose of the program is to increase understanding of the link between DNA and disease in African Americans.

The person who ended up using the kit was someone whose background I had researched in the past with little success.  She was raised in an orphanage in Genesee County, New York and had only the name of her mother.  The name was common, of course, and I was not lacking in finding thousands of women with the same exact name.  She submitted the kit in December and the results took about six weeks.

She was matched up with 44 other genetic relations.  This does not provide much to work with.  (For perspective, the other accounts that I manage have far more matches.  My mother and I have about 1300 each, while my father and his third cousin have around 600 each.)  The closest relation shares 0.50% over four segments.  This is outside the parameters for an accurate prediction of the closeness of the relation, but it can be characterized as distant.  The "Show Close Relatives" icon appeared and pressed it.  No other matches appeared.  I am not sure if this means that there are close relatives lurking in the database.  I initiated contact with all relations and just a few have responded.  They are related beyond a third cousin level, so we cannot map out this elusive family tree yet.

Autosomal DNA testing will reveal relatives and can be very useful in cases such as this when parentage is unknown.  But unless a very close relation surfaces and agrees to communicate, we will only have vague notions of possible ancestors.

Your identity is concealed from "close" relatives- first cousins and nearer.  Both parties must consent to be revealed.

23andMe.com provides an additional warning before you consent to view close relatives.
It is possible to discover that your parents were not biologically related to you,
or that a parent or aunt/uncle had children you were not aware of.

Indexing Projects, Part Two

Last fall I helped index Newark, New Jersey naturalization records for the Italian Genealogical Group.  (That database is not online yet.)  I am now working on a new indexing project for the group:  deaths in New York City 1949-1963.  I have the M surnames for the year 1952.  I am entering the data from a typed index.  The current death index on the site covers New York City (all boroughs) 1898-1948.

Page from the New York City death index, 1952