dna, Finding John Costello, puzzling

Finding John Costello – A DNA Journey: Connecting Four MORE Cousins

Finding John Costello, cfc-01


Oh BOY! do I have an update for you.

But first, answer these questions for yourself.  What happens when a new discovery means that the information you started with cannot be correct?

What do you do when it seems like everything you’ve been told is . . . dare I say it? . . . a lie?

And if it is all a lie, or set of lies, whose lie is it?

How many people know the truth?

Who lied to me?


Did they even know they were lying or is the lie so old, that the perpetrator of the lie managed to pull off an all-encompassing act of deception that has been perpetuated through the generations of my family by unwitting accomplices?

Is it possible to identify the source of the lie now, after all these years?

Chew on those bad boys for a minute while I update you on some new DNA data.


We Have Contact


Months ago, I messaged a DNA match.  Let’s call her Sarah*.  She matches my Mom and Uncle C with a fairly significant amount of shared cMs – 129.3 and 160.6, respectively.  I didn’t hear back from her.

Then I was granted access to Cousin Mack’s* DNA as well.  He matches Sarah with 182.4 shared cMs.  Wow, that is a pretty close match for an unknown cousin.  I really wanted to talk to Sarah!

Well, four-and-a-half months later, Sarah messaged me back.  We started working together.  She is awesome!  She started reading all of my posts about John Costello and got her whole family fascinated in my quest.

Guess what happened next?

Her Mom and Dad, who had already tested but not gotten results, retested, and her aunt and uncle also tested.

Well, those results have been rolling in, my friends!  And IT. IS. FASCINATING.

Seriously, fascinating.

It was definitely time for more data tables.


Working With the Details


So, last week, I dug in.  I started making some tables.  I started looking at the possible ways we connect based on the number of shared cMs between everybody.  The great thing is that I have data about SEVEN people who descend from John Costello, and FOUR people in Sarah’s family to work with.  That is a lot of good data.

Even better?

Sarah’s family also match the last cluster of four cousins.  We just don’t know exactly how for sure.  Yet.

So, back to the charting.  I’m building tables, and tables, and tables.  Every time I think I’m done, I build another table – just to check.  I went through all of the possible relationships that were also logical.  Obviously, Sarah’s Mom is not my Mom’s Great-Great-Aunt – or vice versa.  There are several reasons why that just does not work.  Aside from the obvious no-gos, I checked it all!

In a fury of work over a few days, I ended up with a 39-page report filled with relationship hypotheses and tables of data that prove that hypothesized relationship possible OR not possible.  This time, I even added in the probability that the tested relationship was correct.  I was getting fancy trying to wrap my mind around the data.  Those tables look a little something like this:


Hinde Fried as John Costello's Aunt - relationship chart


We have Sarah’s family on the left of the table, my family on the right.

I have a few things to explain about the differences between my family members in this table and my family members in the last tables.

First, Sarah has shared with me how many cMs her family members share with James*.  I haven’t been able to communicate with James so I had no cM data for him before now.  Second, I had totally forgotten to include my surprise first cousin, Bob* on the last go ’round.  And third, I mistakenly listed Mary* as the daughter of Mack*.  She is not, she is his niece.

{Let’s just take a moment right here to acknowledge that when a family has decades-old broken places post-divorce like mine does, these types of mistakes are MUCH easier to make.  I don’t know my John Costello second cousins.  But, all is fixed now with Mary is her proper place.  Phew!}

This relationship chart is lined up to reflect the possibility that John Costello is the nephew of Anna, born Hinde, Fried.  Everyone who has taken a DNA test is highlighted with green.  Now I can quickly see what relationship that gives all of the DNA test takers to plug into this table:


Hinde as John Costello's Aunt, data table


I got a little bit fancier with this table.  You’ll notice that the various family members listed on the left and along the top have two different blue colors to separate them into the two generations.  The older generation is listed first and is highlighted with darker blue and the younger generation is listed second and is highlighted with lighter blue.  This is a visual reminder to me that the older generation’s numbers matter more.

I also added a probability column for each relationship.  After plugging each number into the Shared cM Tool, I listed first the grouping this number of shared cMs is a part of – 1, 2, 3, etc.  Below that, I listed the percentage of probability for that hypothesized relationship.  Then I color coded that box – green is a percentage from the first grouping of relationships on the probability table, yellow is second, orange is third, light grey is fourth, darker grey is fifth, light red is sixth, and 0% probability relationships are red.

I ran through all kinds of numbers, all kinds of relationships.  Then I ran numbers for Sarah’s family and the other four cousins from my previous cluster.  Then I re-ran my numbers with the four cousins from my previous cluster to add in some additional data and correct a mistake I noticed.  I had mistakenly used the longest shared segment number instead of the total number of shared cMs between my Mom and one of her matches.  Oops!

Finally, 39-pages later, my data masterpiece was officially – a rough draft.

Rough draft because, well, I’m human and we humans make mistakes.  I’d already found two to fix, plus the omission of Bob.

But even more importantly, it’s a rough draft because there are LOTS of empty boxes in those tables!

I want more data.



Because after running all the numbers and looking through my charts and then printing all of my charts so I could physically move them around into “if this & this – then this” groupings, I noticed something.

The grouping with an overwhelming amount of green is this one:


Hinde & John Costello as siblings relationship chart


Plus two others that make Isidore Fried, Anna Fried, & John Costello – SIBLINGS.

Siblings, people, SIBLINGS!

Sorry I keep shouting at you, but, siblings?

What the heck is going on here right now people?!

The Fried family were not from Spain, or near Spain, or Castillos.

You probably want to see the data table though, right?

Here is one of them:


Hinde Fried & John Costello as siblings data table


There are only three other possible relationships based on the current data:

  • Anna Fried and Isidore Fried – full siblings, are the half-siblings of John Costello.  However, this one is troubling because John was born in between Anna and Isidore, supposedly in Spain, as opposed to – Russia.  Well, the part of Russia that is sometimes Poland, and sometimes Ukraine, or so it seems.   That is where Isidore and Anna were born.  Russia, not Spain.
  • Anna Fried and Isidore Fried are the Aunt and Uncle of John Costello.
  • John Costello is the Uncle of Anna Fried and Isidore Fried.

That’s it.  Nothing else works.

I had to take a bit of a breather, because, well, something is not right here.


A Question


I called my Mom.

There was this story niggling at me.  A story about family members trying out their high school Spanish on Grandpa Costello and him not understanding a word of it.  I’d heard it a handful of times, but it had been a while.  I was thinking it was members of my Mom’s generation that were the high school Spanish speakers, but as the phone was ringing I knew that couldn’t be it.  They were much younger than that when Grandpa Costello died.

I asked my Mom about the story.  Without even a nanosecond of hesitation, she answered my question.

It was my Grandma Deane and Vince who had done it.  They had tried to talk to Grandpa Costello in Spanish and he knew nary a word of their basic efforts.

Not one word.

I asked my Mom if she remembered her Grandpa speaking any language other than English.

Nope.  She did not.


I did the very thing I asked you to do many words ago.  I chewed on all of it for a bit.

And then I slept on it.

And then I decided I needed to study up on the one factor I hadn’t been able to put into my data tables yet – endogamy.


Endogamy Study


Endogamy makes comparing shared cMs far less precise.  What is endogamy?  Well in a few words, it is groups of people that have intermarried because of geographical, cultural, or religious limitations.  This intermarrying causes people in that group to be related in multiple ways, inflating the number of shared cMs.  You can read more about endogamy here.  The European Jewish community – from which John Costello definitely descends – is an endogamous group.

Were all of my tables still valid?  Or close to valid?  Or not at all valid?  Or a hot mess of some of all of that?

I started reading.  And reading, and reading, and reading.

Endogamy wasn’t a new concept for me, but for the first time, I was highly motivated to try to understand if there was any way to quantify its impact on my data.

After lots of reading, there were three tables that were the most helpful for my question.

Lara Diamond has compiled several incredible tables based on data she has gathered regarding endogamy in Jewish populations.

This first bit of table comes from her post about 2nd cousins (I cut it down considerably):


Screen Shot 2019-03-25 at 7.58.57 PMScreen Shot 2019-03-25 at 7.52.40 PM crop


While the expected “average” number of shared cMs between 2nd cousins is 212.5, according to the ISOGG wiki, she was seeing an average of 240.75.

Ummm, the average of the shared cMs between Rachel and my Mom, Uncle C, and their cousin Mack is 225.67.

Okay.  How about her next table?

This one comes from her compiled data of all responses.  Check out the three 2nd cousin relationships I have circled:


Screen Shot 2019-03-25 at 7.53.07 PMScreen Shot 2019-03-25 at 7.53.25 PM


My family’s average of 225.67 definitely falls into those ranges.

So, feeling like an endogamy numbers newbie, and wondering if I was even in the right ballpark, I asked for some input on my numbers in this Facebook group – Jewish DNA for Genetic Genealogy and Family Research.  Lara Diamond totally reinforced my own conclusions, that the numbers look like my Mom’s generation of Costellos and Sarah’s Mom’s generation of her family are either 2C, 2C1R, or 1/2-2C.

Oh boy!

Then I read through the pdf file about the data in the Shared cM project.  There is a section on endogamy and how it impacts the numbers.  You can look at that pdf here.  The endogamy portion begins on page 22.

While there are several interesting numbers you can look at, I’m just going to list the averages for the three relationships in question on that table:

  • 2nd cousins in endogamous groups average 274 shared cMs in the Shared cM project.
  • 2C1R in endogamous groups average 153 shared cMs in the Shared cM project.
  • 1/2 – 2nd cousins in endogamous groups average 117 shared cMs in the Shared cM project.

While those numbers are not a perfect match for Lara’s data project, they are similar enough that I started to feel more comfortable with considering endogamy’s impact on my own numbers.


What Next?


So, so, so much is next.

First – data.  I need more data.

We have lots of DNA testers who are not in the same pools of data.  Now before you go suggesting that we all use gedmatch, I’m gonna stop you right there.  That is a non-starter in my family.  And since I want to get MORE testers, not fewer – as in existing testers requesting their results be deleted, I’m not even going to consider it for a second.

But – we can all transfer ourselves over to MyHeritage and get some more comparisons going on.

I also have a handful of DNA kits for Ancestry and 23andMe sitting in my office waiting to be mailed out to anyone in my family willing to test.  How am I going to lure them in?

Well, you know that magnificent rough draft report?

My genius brother was overwhelmed by it, so I definitely can’t use that.  But I am going to synthesize it and present a short and to the point report of the possible relationships between our family and Sarah’s family.  Hopefully, that will cause some family members to bite.  Whether to help prove or disprove what I am seeing.

I’m not just going to send it to a few family members.  I am going to send it to every single John Costello descendant.  With an invitation to test.  With no great-aunt filter between me and the request.

{wish me luck – I need it!}


Second – info.  I need more info.

I know a handful of story bits that I want to trace back to the source.  I want to document every last scrap and whisper of a story about Grandpa Costello, his immigration story, his memories of the old county, what he said about his family, his business in Spokane, his friends, ALL OF IT.

I want to document the heck out of ANYTHING that ANYONE remembers.  No matter how ridiculous.  No matter how much they question their own memory.  I want it all.


{Where is Dumbledore’s Pensieve when we need?  Oh, and while we are at it, the ability to draw lovely shimmery wisps of memories from people’s minds with a wand?  Oh wait, we genealogists have to make our own magic, dang it!}


I want it all because this is what I’m thinking.

John Costello wasn’t named John Costello at birth.

John Costello wasn’t born on 14 February 1893.

John Costello wasn’t born in Barcelona.

John Costello was someone else entirely on the day he was born.  Will I know him when I see him in a record?


Last Thoughts


Now that you’ve had some time to mull, what do you think?  Did someone lie?  Or is there another explanation that I just can’t see, yet?

But if someone did lie, who was it?

And, why?



*All names of living people have been changed to protect their privacy.  Although, I have always called my Mom, Mom, so I decided no code name was necessary.  😉


Please note that the DNA results of myself and Sarah are unnecessary when comparing to individuals who can be compared to our mothers.  We got everything that we have (in this scenario) from our mothers.  However, Sarah is my contact person in the family and for that reason, I included her in the table out of respect for her constant and very gracious work with me.   I included myself because I have tested at 23andMe and my mother has not.  Rather than leave myself blank compared to individuals who can be compared to my Mom, I filled it in.  I’m glad I did.  If ever I questioned the value of testing more family members, I definitely do not question it now!  Seeing the crazy swings in numbers for me compared to my matches, and the more stable differences in numbers when looking at my Mom’s matches, I am 1 million percent sold on testing everyone.  I thought I got it before.  After 39 pages of tables, I REALLY get it now.  Test ’em all people, test ’em all!

Seriously, TEST THEM ALL!



{I’m done shouting for the day.  xoxo}