“Know Thyself” A Search for One’s - PowerPoint Presentation

Slide1

“Know Thyself”

A Search for One’s

(Biological) IdentitySlide2

Overview

Recent favorable trends in Genetic Testing

How I embarked on The Search

How to find YOUR Biological Parents (or any other relatives) using Genetic

Testing

A

Sample “Case”

A Blow-by-blow Walkthrough of the “Weiss Case”

Parting Thoughts and Observations

Q & A?Slide3

The Miracle

of Cheap Genetic Tests

23andme,

FtDNA

,

AncestryDNA

Cost of sequencing falling

rapidly

Popularity

Millions have taken tests and millions more will as prices fall furtherSlide4

The Plummeting Cost of SequencingSlide5

Lower Cost -> Higher DemandSlide6

A Conversation with Sunil B

Sunil

Bopardikar

: VP of Engineering at my former company

Long informal philosophical chat in 7/2013 after award ceremony

Discussion Topic: Whence

exceptional

talent?

Our conclusion: mostly due to random genetic fluctuations, globally omnipresent, randomly distributed, awaiting accidental discovery

Sunil’s

Q:“Have

you taken the 23andme test?”Slide7

Computer Science

A Happy Coincidence: My choice of Profession

Three Pillars of CS: Data, Data Structures, and Algorithms

CS is all about extracting, gathering and cleaning up data, then figuring out how it all interrelates (pun intended), and finally developing methods to extract inferences

This “mystery” was tailor-made for a CS wonkSlide8

How To Find

YOUR

Biological Parents

A Step-by-step GuideSlide9

Essential Concepts

Consanguinity aka “Match Percentage”

The Most Recent Common Ancestor (MRCA)

Half Relations

Haplogroups

The “

Records Horizon”

The Process of Elimination

The

Numbers

Game (as in “Law of Large”)Slide10

Essential Skills and Resources

Family Tree Construction

Finding (and Dealing with) Living Relatives

Knowing where to find Key Information

Handling RoadblocksSlide11

Consanguinity

Consanguinity: The percentage of VERY slowly changing genetic markers that two people have in common (about 930K SNP’s out of 3.2B base pairs)

International Society of Genetic Genealogy (ISOGG) Chart – the source of all knowledge!

Determines possible / probable blood relations

Used to identify the MRCA

: “Most Recent Common Ancestor”

NB: Adoptees are often

half

-

relations (also caused by subsequent marriages)Slide12

Divide Percentages

(in

red

) by 2 for

HALF RELATIONSSlide13

%

Shared

cM

(

Centi

Morgans

)

Relationship

100%

6800

Identical twins (monozygotic twins)

50%

3400

Parent/child

50%

2550

Full siblings

25%

1700

Grandparent/grandchild, aunt-or-uncle/niece-or-nephew

, half-siblings

23.4375%

1593.75

Double first cousins

12.50%

850

Great-grandparent/great-grandchild, first cousins, great-uncle or aunt/great-nephew or niece

, half-uncle or aunt/half-nephew or niece

6.25%

425

First cousins once removed, half first cousins, great-great-aunt/uncle,

half great-aunt/uncle

6.25%

425

Double second cousins

3.13%

212.5

Second cousins, first cousins twice removed

, half first cousin once removed, half great-great-aunt/uncle

1.56%

106.25

Second cousins once removed, half second cousins, first cousin three times removed

, half first cousin twice removed

0.78%

53.13

Third cousins, second cousins twice removedSlide14

Finding the MRCA

The success of any search for biological parents hinges on the correct identification of the Most Recent Common Ancestor (MRCA) you share with a person whose lineage is already known (or reasonably discoverable).

Basic Concept: Discovering where two family trees converge (in which specific ancestor)

Good news: for most adoptees (

half

relations), MRCA’s are one step CLOSER than they usually are for

full

relations

NB: For

non

-adoptees, the “MRCA” is a

coupleSlide15

Critical Question: How Far Back?

Once you’ve identified a match, the critical question you need to answer is: How far back must you go in

their

family tree to find your common ancestor?

This is important because of the “Records Horizon” constraint.

To

answer

this

question:

Estimate

the MRCA’s birth year

As a general rule, it’s very difficult to find records prior to an ancestor’s immigration

Spellings can change, foreign records are usually not digitized (or searchable)Slide16

From Theory to Practice

A Worked ExampleSlide17

Case Background

Cohen, born in 1930, is adopted and wishes to find his biological parents

Liebowitz

is a 3% “genetic match” with him according to 23andMe

Liebowitz

is 24 years younger than Cohen

Questions:

What is their

probable

blood relation?

How far back must the

Liebowitz

family tree be traced to find their Common Ancestor?Slide18

Finding the Relationship

Step 1

: Since Cohen is one generation OLDER than

Liebowitz

, and since they share 3% of their genetic markers, it may

seem

that we need to look for “3%” on the ISOGG chart

Step 2

: Since C & L are almost certainly HALF relations, any matching percentage on the chart must be divided by 2! So we actually need to search for “6%” (or thereabouts)

Step 3

: Assigning Cohen to the “self” bubble on the chart, we scan the next row down (one generation younger) to find

Liebowitz

’ 6% figureSlide19

Finding the Common Ancestor

Step 4

: The only possible explanation

given the match percentage (and relative ages) is

that

Liebowitz

is Cohen’s

half

first

cousin

once removed.

Step 5

: Now that we know how C & L are related, we need to determine how far back to map out the

Liebowitz

family tree.

Preliminary Answer

: The common ancestor is THREE generations older than

Leibowitz

and was one of her great grandparents (and one of Cohen’s biological grandparents)Slide20

The “MRCA” Birth Year

The Key

Question

:

Approximately when

was this common ancestor

born

?

Answer: assuming 20 – 30 years per generation, this person would have been born between

3

x 20 and

3

x

30, or

60 to 90 years before

Liebowitz

(or 40 to 60

years before

Cohen)Slide21

Do Records go back that far?

Since Cohen was born in 1930, the common ancestor would have been born sometime between and 1870 and 1890.

Can you find

searchable

records for them that identify

all

of their children (one of whom is Cohen’s parent)?

If so, identify all of their descendants and see which one(s) are “candidate biological parents” based on ages, cities of residence and

your

birth year and city

If not, the information needed to identify the common ancestor lies beyond the “Records Horizon”Slide22

Review

Based on a match percentage, we determined the likely

relation

using the ISOGG

chart

From this we determined the probable common ancestor

Given how many generations older that common ancestor is than our match, we determine if it’s even possible to find the requisite records identify

all

of our match’s ancestors of that generation (family tree level

)Slide23

Review

We try to rule out as many of these ancestors as possible

(

eg

ones

who died in youth,

etc

)

We then generate family trees

downward

for each of the remaining candidate common ancestors to find our biological

parent

filtering by times and places of residence during adulthood

Observation

: For each generation further back we have to go, the number of candidate common ancestors DOUBLES!

Sneaky Move

: If the match

knows

through which parent we are related (sometimes known), we can eliminate half the tree we need to trace!Slide24

The “General Algorithm”

Take genetic tests with

as many services as possible

Designate your results as “public”

If money is tight, chose services in descending order of database size (23andme,

ancestryDNA

)

Have

adoptee

take the tests (finding matches using their children is more difficult)

Check match results at least monthly.Slide25

General Algorithm (2 of 3)

Considering the “Records Horizon” problem, Select a match percentage

“

action threshold”

Suggestion: Only send “contact invitations” to matches greater than 2.5% or 5%

Be a

S

queaky (but pleasant) Wheel with the invitations

Once contact is established, build rapport and seek to obtain as much of the “Essential Information” as you canSlide26

General Algorithm (3 of 3)

Essential Information:

Names of parents and grandparents

Years of birth and death

City of their residence when they reached sexual maturity

City of birth, marriage and death.

Use the steps outlined in the “Worked Example” to determine the common ancestor (

eg

grandparent)

(

Adoptees

) Map

all of

their

descendants, ruling out as many as possible. One of the remaining descendants is your biological

parent

(Adoptees) Repeat

with as many matches as needed until you have identified both biological parents

Slide27

Tools, Pitfalls, and ResourcesSlide28

Haplogroups (Adoptees)

An essential tool for determining

which

parent you share with somebody else

Maternal and Paternal “letter groups” designating gene clusters shared by people originating from a specific geographical region

MHG (Maternal

Haplogroup

) passed on the X

chromosome from mother to child

PHG passed on the Y (sorry ladies

) from father to son onlySlide29

Haplogroups (cont’d)

All women along

a matrilineal

line share the

same MHG. They pass their MHG to offspring.

All men along

a

patrilineal

line share the same

PHG. They pass their PHG only to

male

offspring.

Women have no PHG, only an MHG

Two males sharing the same father (or paternal grandfather) have the same PHG

Any two people sharing the same mother (or maternal grandmother) have the same MHG

Negative cases can

exclude

possible relationshipsSlide30

Potential Roadblocks

Difficult

to find records prior to

emigration due to:

Mangled

spellings in passenger manifests

“Americanized” names (

Calogero

-> Charlie

)

Records are in foreign languages

Records abroad not likely to be digitized or online

Families whose members have been estranged for many years

(Adoptees) Relatives

who simply refuse to talkSlide31

Census Problems

Census takers often misspelled “tricky” foreign names, so searches fail

Immigrants lied to census takers about about when they immigrated, their naturalization status or their age

Immigrants may have

avoided

census takers altogether

Pre-1900 US

Census Records

are

very

sketchySlide32

Statistical Ambiguity

Relation

AVERAGE Match

Possible Range

1st Cousin

12.50%

7.31% - 13.8%

1st Cousin once removed

6.25%

3.30%

- 8.51%

2nd Cousin

3.13%

2.85% - 5.04%

2nd Cousin once removed

1.50%

0.57% - 2.54%

3rd Cousin

0.78%

0.30%

-

2.00%

4th Cousin

0.20%

0.07% -

0.50%

Problem: What relationship does a 4% match have with you? There are

TWO

answers! Things get fuzzy when ranges overlap.

Source:

https://customercare.23andme.com/hc/en-us/articles/202907170-Average-percent-DNA-shared-between-relativesSlide33

Differing Generational Cycles

The average years between generations can vary by

branch

of the family tree.

As the number of generations back to the common ancestor grows, the more ambiguous people’s

“

generation count” becomes.

Example: Two

people of the

exact

same age today can

actually be

one generation

apart

if one person’s branch averages 20 years per generation and the other’s averages 30 years

(over a 60 year period)Slide34

Pointers

Tree building: Accept the fact that you must subscribe to

ancestry.com

and become highly proficient in its use (or hire someone who is)

To build

a Family

Tree,

derive relationships

from Census

data

and other people’s public trees

!

(Adoptees) Find

“Search Angels” –

eg

Priscilla Sharp

Extract family relations from newspaper birth and marriage announcements, and obituaries (the hands-down BEST source of all)Slide35

Still More Pointers

Obtain Death Certificates to learn

names of parents (

eg

maiden names) and birth dates

Resist

the urge to spam distant

(low match) relations indiscriminately

out of

desperation

Prepare

to endure

Endless Discouragements:

Contact invitation response

rates are

very

low.

Responses

are often long

-delayed

Respondents may

question your motives

People

will say your

odds

of finding a biological relative are

minuscule

(and they are correct

!)Slide36

Avoid “Combinatorial Explosion”

The only way to combat these problems (for non-adoptees) is to trace both back family trees until you find the MRCA(s)

For each

generation further back you must go to find the

MRCA, the number of people you need to consider (trace) grows

exponentially

Conclusion:

Avoid

longshots

.

Trying to track down very low match percentages can drive you nuts.Slide37

Key Sources of Information

Genetic Testing

Sites

AncestryDNA.com

(

Mandatory

)

23andme.com (

Mandatory

)

ftDNA.com

?

Gedmatch.com

? (Optional)

Federal and State Censuses

Genealogy

Websites

Ancestry.com

(

Mandatory

)

Familysearch.org

(Free!)Slide38

Key Sources of Information

Newspapers (digitized

archives searchable online)

Obituaries

 The

Motherlode

Marriage Announcements

Birth

Announcements

Digitized

newspaper Archives (

eg

fultonhistory.com

)

Search

Angels

Priscilla Sharp

Local Genealogical Researchers (library gnomes)

City, County and State Vital

Records

Birth, Marriage and Death records

Adoption AgenciesSlide39

Key Sources of Information

“People Search” Sites

Intellius.com

Radaris.com

Obituary Search Sites / Google obit searches

Adoption

GroupsSlide40

A “Real Life” Case:

Walkthrough

of the Weiss Case

(2014 – 2015)Slide41

Mother’s Side

Based on my conversation with Sunil, I take the 23andme test and buy another for my father in 12/2013

I introduce my father to Priscilla Sharp, an “Adoption Search Angel”, and she and I then urged him to request a copy of his (heavily redacted

)

adoption record from the State of NY

First Clue: In

2/

2014 I learn on 23andMe.com that Melanie

Ott

shares 5.81% (

ie

~6.25%) of her genetic markers with me and is of my

generationSlide42

The Match that Started EverythingSlide43

Pieces Fall into Place

I

send

Melanie

Ott

an

invitation to connect

on 2/14/14

She

does not answer

until

12/17/14

(

!)

I immediately tell my father about my discovery

He notes he has a 12.4% match with Melanie suggesting we are on the right trail, so he contacts her

By

process of elimination, per the ISOGG table, the only possible explanation that fits the facts is that

she is

my

half

1

st

cousin (not 2

nd

cousin as the site claimed)

Therefore we share ONE grandparent who is also my father’s

parent

.Slide44

23andMe - OverviewSlide45

Parallel Efforts

On 12/19/14 Melanie

provides the names and descriptions of her grandparents

which my

father

then forwards to

Priscilla Sharp, a

“Search

Angel

” (

truly

worthy of the title)

Priscilla notes that

Nathalia

Baker (nee

Gratz

) is

Hungarian

matching the “Hungarian Jew”

described in

the

redacted (“non-identifying”) adoption report from NYSlide46

Seeking Certainty

Although I respected the strong circumstantial case Priscilla made, I wanted

definitive

proof before claiming anyone was a blood relative, so I resolved to find at least one of Natalie’s surviving children to take the 23andme test

Problem: Due

to

“family issues”, Melanie has

no idea where any of them lives (!!)

Her father

refuses to

tell her on

religious

grounds (!!!

)Slide47

AFW, Private Investigator pro tem

Based on census records (

ancestry.com

), I learn that Melanie’s father (Richard Baker) has three siblings: Howard, David and Carol

Huge Problem: “Baker” is a

terrible

name to look for in a nationwide search!

I conclude that it’s impossible to find them based on name alone.

Nobody knows Carol’s

married

nameSlide48

The Rest of the Story

Priscilla

Sharp SA, also

told me about a massive online

searchable archive

of Buffalo newspapers

I

spent days searching

it and

finally found

the

married

name of Carol Baker (Roeder)

in her wedding announcement from the early 60’s

Using this

very

uncommon

name, I

located her

in

Florida

using

intellius.com

Melanie contacted

her aunt Carol to

get the current addresses of all of the other

Baker

children

(Melanie’s uncles

)Slide49

Mystery Solved

I decided that getting one of her

sons

to take the test would pin things down beyond any doubt (since my father and he should

share

a

MHG but

not

their PHGs)

Carol told us that Howard had died, but that David and Richard were also living in FL.

Richard wanted nothing to do with genetic

tests

David took the test, and his results solved the case: he shares 25% of his markers with my father thereby establishing their “half siblinghood”

David’s MHG

matched

my fathers, but his PHG did not

proving

that their common parent was his mother.Slide50

Confirmation Of Shared MaternitySlide51

Father’s Side

In 12/15, I suddenly learned that Nancy R shares

12.5%

of her genetic markers with

me (!

)

and is about one generation older.

Only

plausible explanation of

this high a percentage (

per ISOGG): she is my half-

aunt and therefore my father’s half-sister

Since we know who my father’s mother is (and she is not Nancy’s mother), the shared parent must be her

father

On the site, Nancy posted a

tiny

3-person public family tree

naming

her father

 Slide52

Father’s Side

Census data, along with a slew of other records, put him (Vincent

C)

in Buffalo from 1907 - 1968

I undertook

a massive

research

project on

and on Christmas Eve

emailed

a complete dossier on him (including pictures) to my father and siblings announcing that I had finally cracked the case.

I asked my father to take a

AncestryDNA

test to prove this

theory by measuring his match with Nancy

He did,

and the results corroborated

the contention that they were half siblings.Slide53

AncestryDNA - OverviewSlide54

AncestryDNA - DetailSlide55

EpilogueSlide56

“Estimated Relationships”

The “Estimated relationships” shown on genetic matching sites

overlook

the

possibility of half relations, yet just about

all

adoptee relations are half relations!

The “Estimated relationships” these sites suggest are therefore

highly

misleading for adoptees

Example: My father and Melanie

Ott

share 12.5%, the exact percentage of “full” first cousins. Obviously impossible given their age difference.

Example: Nancy R and I share 12.5%, usually implying first

cousinhood

: WRONG! Nancy is a generation older than me and is my half-aunt.Slide57

Adoptees: Walking on Eggshells

Suggesting to a complete stranger that you may be an unknown close relative can be off-putting

It may provoke suspicion, defensiveness, religious concerns,

etc

Consider what you are doing: Asking a stranger to accept (and face) the fact that one of their erstwhile revered ancestors “messed around”

Many do NOT want to “speak ill” of the dead

Others will “See no point” in including you in the family (or feel it does more harm than good)Slide58

Why there is Hope

People, especially those of a scientific bent, are curious about their genetic lineage, hereditary proclivities

Others, who are genealogically inclined, want to find distant relatives

Others are curious about, or want to contribute to gene mapping research

As time passes, an increasingly large percentage of the population will have taken these tests

Social Mores have changed: Stigma formerly attached to “illegitimacy” is almost completely goneSlide59

Why there is Hope (Continued)

The 1950 Census data will be released in 2017

More and more genealogical data is being digitized every

day around the world

All it takes is ONE (good match)

In spite of the sheer

improbability

of it all, we pulled it off, so you can too! It’s just a “numbers game”Slide60

Resources for Further Study

ISOGG, Interpreting match percentages:

http://www.isogg.org/wiki/Autosomal_DNA_statistics

https://dl.dropboxusercontent.com/u/21841126/DNA.pptx

http://www.genealogyjunkie.net/miscellaneous-

notes.html

Search Angels:

http://reunion.adoption.com/adoption-records/search-angels.html

http://priscillasharp.blogspot.com/p/search-angel.html

From

Reeva

:

Steve Morse’s Ellis

Island

Record Search Site:

http

://stevemorse.org

/

JRI-Poland database

:

http

://jri-poland.org/

jriplweb.htm

Polish

Newspaper Searching:

http

://jri-poland.org/

monitor.htm

Polish Books

of Residents:

http://jri-poland.org/

bor.htmSlide61

Q & A ?