Generating Investigative Leads with Alternative Database Search Approaches
Crime scene DNA samples are typically analyzed using an autosomal STR methodology and searched in local, state or national databases using a direct approach, which has a hit rate of ~30-50%. But what about those samples that don’t generate leads using direct database comparisons? There are now multiple alternative procedures that can effectively generative investigative leads from these samples. In this webinar, Jianye Ge, from the Center for Human Identification at the University of North Texas Health Science Center, will compare three methods of searching for relatives in DNA databases. Ray Wickenheiser, from New York State Police crime lab system, will present cases using Y- STR analysis and genetic genealogy, as well as a status update on the state's familial search program. In this webinar you’ll hear about: • Indirect search approaches using familial searching, Y-STR database searching, and genetic genealogy • Case examples using various search methods from a forensic DNA laboratory
You Might Also Like
Hello, and welcome to generating investigative leads with alternative database
search approaches
brought to you by forensic and sponsored by Thermo Fisher Scientific.
This is the first webinar in the five-part 2021 Future Trends in Forensic DNA
Technology
Series.
My name is Michelle Taylor, Editor-in-Chief of Forensic, and I will be your
moderator
throughout.
For today's webinar, you can earn one hour of continuing education credit.
Following the conclusion of the webinar, you will receive an email with
information
on how to obtain CE Credit Documentation.
We have a great lineup scheduled to present to you today, but before we begin,
I'd like
to take just a moment to cover a few logistics.
At the end of the presentations, we will hold a question and answer session.
To ask a question, click on the "Ask a Question" tab in the upper right corner
of
your screen.
Please also note that the right side of the screen features an overview of
today's webinar,
as well as more information about our speakers.
If you have a technical question during today's event, click on the "Test your
Connection"
button at the bottom of your screen.
From there, you can access additional webinar support.
We also invite you to use the social media widgets beneath the webinar to share
with
your friends and colleagues.
First, you will hear from Eugenia Jueux, the Associate Director of the Center
for Human
Identification at the University of North Texas Health Science Center.
He has over 15 years of research and development experience in bioinformatics
and forensics
with more than 70 publications.
His research relates primarily to computational analysis and interpretation of
DNA forensic
data.
The software programs he developed have been used by the federal and state
government agencies
to assist in solving criminal cases.
Next, you will hear from Ray Wickenhiser, who is the Director for the New York
State
Police Crime Lab System.
He has over 36 years of forensic science experience, with 19 of those as a
crime lab director in
local and state crime laboratories.
His areas of expertise include crime lab administration, investigative genetic
genealogy, quality
management, forensic DNA, serology, hair and fiber trace analysis, physical
matching and
comparison, glass fracture analysis, and forensic grain comparison.
Thanks, Michel, for that introduction and thanks, I'm officially for inviting
you today.
I'm going to talk about one of my recent papers published in the Journal of
Forensic Sciences.
In that paper, I compare three different approaches of searching in relatives
in DNA
databases to assist a forensic investigation.
Before I get into the details, I would like to share some numbers reasonably
published
by an NIJ.
According to a recent NIJ blog, there are 250,000 unsolved homicide cases in
the US, and more
than 100,000 cases were accumulated in past 20 years long.
So there are so many unsolved co-cases, and we are desperately looking for
tools that
can help us to solve co-cases.
DNA database is one of the most powerful tools that can assist in co-case
investigation.
What typical case you collect sample from crime scene, you send it to the lab
and generate
the DNA profile, typically the Osoma STI profile, and then you search the STI
profile against
the coded database and hope there is a direct match.
If yes, you develop a lead for forensic investigation, but in many cases, you
don't have a direct
match, then what you can do is to search for indirect match, which means the
potential
relatives of the forensic profile in DNA databases.
There are three different approaches to search relatives, depending on which
database you
search.
From new search, use the Osoma STI profile search against the traditional Osoma
STI database
like codis to look for first-degree relatives.
YSTI database search YSTI database to look for any male relatives.
And inviscated genetic genealogy or IgG use high-density SNP profiles to search
against
SNP database to look for any relatives, close or distant male or female.
As a matter of which one you use, all these approaches need certain degree of
family or
genealogy investigation.
Either in a small family, it just contains some first-degree relatives or very
distant
relatives in a big family.
And after you generate some leads through the database search, eventually you
still need
to get the sample from the suspect and generate Osoma STI profile to compare
with the crime
scene profile to confirm the identity.
Next, I'm going to talk about the details of these approaches.
Fermiive searching is to detect potential relatives of a forensic profile in
DNA database.
This is the workflow of a forensic processing process.
You have a forensic profile generally from a crime scene sample and you search
against
DNA database, measure the similarity between the forensic profile and the
offender profile,
and then you rank the offender profile by similarities.
And the similarity can be kinship index given a specific relationship or number
of shared
alleles.
Because they are very limited number of STI low-sized including coders, we may
only detect
the first-degree relatives with high confidence.
So first-degree relative means their parent child or four-simile relationships.
Now after you rank the offender profiles, the top candidates can be selected
for further
investigation.
The top candidates are the most likely profile being the relatives of the
forensic profile.
And in the first investigation, you may use non-genetic information like age,
place of
residency to fill out the unrelated candidates.
With the forensic profile's male, so you may use YSTAP profile to fill out the
unrelated
candidates.
And in some cases, if the forensic profile is female, you may use mitochondrial
DNA to
fill out the unrelated.
And the KELT-D-O-J did use mitochondrial DNA in familial search.
And after this task and filtering, you have very good chance to detect
relatives.
And if they are really in the database.
One of the reasons we believe male search should work is based on this BJS
report in
1996, a question of population in the United States.
According to this report, at least 42.8% of the dual-emase have close
biological relatives
who are incarcerated.
Biological relatives means the father, mother, brother, sister, and child.
These are usually not biological relatives.
If we further break down into details, there were about 36.5% inmates had four-
sib incarcerated
and 22.8% had parents or children incarcerated.
Which means there are lots of first-degree relatives in the DNA database if all
the inmates
DNA profile were entered into KELT-D-O-J.
And the clinical system contains about 20 STL-O-Sci, which is good enough to
detect the first-degree
relatives with pretty high confidence.
And for more distant relatives, probably not so accurate.
These are the reasons why from the research in short work, and it did work in
the help
in many cases.
One of the most famous familial search cases was the Grim Sleeper case.
This case was a serial killer case happened in Los Angeles more than 30 years
ago.
In this case, STL-P-FAL were generated from the evidence, but no direct match
was found
in the DNA database search.
So in 2010, about 10 years ago, a familial search was conducted using certain
Kingsford
measure 200 candidate profile were determined as potential relatives of this
perpetrator
for first-degree investigation.
And the YSTF profile of these 200 candidates were also tested, and only one
profile should
exact same Y-hab type with the evidence profile, which belonged to Chris
Franklin.
So very likely the perpetrator of this case may be a close relative of Chris.
Then please also examine the first record of his family and exclude Chris's
brother
like his father Lonnie Franklin.
It could be the source of the DNA evidence.
Then ALAPD sent an undercut team to follow Lonnie Franklin and then retrieve
his DNA sample
from a slice of pizza.
And it turned out to be a match, so it solved the case.
This is the first successful familial searching to solve a homicide case in the
US.
YSTF database searching is very similar to familial searching.
It follows almost the same workflow.
But there's a few things that different.
First, you use a forensic YSTF profile to search against YSTF database, so
which means
you have to have a YSTF database to search.
Second, the similarity measure is different.
For YSTF, it compares the number of mismatch loci or stabs between a pair of Y
profile.
Third, the first investigation types of Osonma YSTF profiles of the candidates
and the candidates
male relatives.
And depending on how far you go, you may investigate the whole male lineage of
the candidates.
So why do we want to use Y?
First of all, most violent crimes were committed by men.
Here are some statistics summarized by DOJ.
For example, 99% of the first four rapes were committed by men.
And white mausom is only inherited between males from father to sons.
And so we can use white mausom to trace the male lineage.
And this feature is very useful in forensic investigation.
It can be used in kinship analysis missing person cases.
Familiar search can use Y to exclude unrelated candidates.
And white can also be used in matrix cases in for ancestry, link multiple
different cases.
And through these many cases, you slowly accumulate Y profile to build a Y
database.
And then you can use the Y database to search for male relatives and help solve
some cases.
There have been many successful Y database search cases.
I would like to talk about one good example that check the replicates happening
in China.
This was also a serial killer case similar to the green sleeper case.
The Osama S.T.A.P. files generated from the crime scene sample.
But there was no direct match in the DNA database.
Around 2013 or 2014, many crime labs in China started building local Y
databases.
And the city where this case occurred also started building its own Y database.
In 2016, the local police department relaunched an investigation of this code
case with the
local Y database.
And the perpetrator, Gao Cheng Yong, was tracked down after his uncle was
arrested for a mining
crime.
So the process was that a man was arrested for a mining crime.
And his Y profile entered into the Y database and matched with the evidence
profile of this
case.
And then the police department investigated his family and the Y.S.A.P.
S.T.A.P. profile to confirm that the Gao Cheng Yong was the perpetrator of this
case.
So this case used Y profile to detect uncle nephew relationship.
And the familial search would not be able to help in this case.
Since Y.S.A.P. database searching is not very common in the United States.
So I'd like to talk about a few other countries that have built their national
Y database.
China is the pioneer in Y.S.A.P.
database.
And there have been at least 20 million profiles, more profiles in the national
database and
they're still rapidly increasing.
The structure of the Y.S.A.P. database is very similar to Osama's T.A.P.
database. It contains Fender, VST, Franci profiles.
And it can be used for any case.
Many provinces in China require typing both Osomo and the Y chromosome S.T.A.P.
profiles
for all the crime scenes sample they collected.
And then they will search both Osomo and Y.S.A.P. profiles in the database
search.
And there has been at least tens of thousands of successful Y.S.A.P.
database in search cases in China.
And there are more cases with confession after rest.
And because the database size is so big and to increase the destination power,
China Y.S.A.P.
database require at least 35 Y.S.A.P.
Singapore also have built their own national database.
They plan to retype all 170 samples they stored by the end of last year.
But probably it's because of COVID-19.
But anyway, that's the plan.
And they mostly use the Y.S.A.P. database for sexual assault cases.
New Zealand and Australia also start their national Y.S.A.P. database in the
past few
years.
And the sample size may not be that big, but it's growing very fast.
The database of the New Zealand and Australia is very similar to Singapore.
They mostly use for sexual assault cases.
But the technical detail may be different for different countries.
One good thing about Y.S.A.P.
database is that you don't need to pass new legislation specifically for Y.S.A.
P.
because the current DNA law do not specify which DNA market to include in the
database.
So it should cover those ensemble in Y.S.A.P.
in this table, I only list four countries as examples.
And we know there are several European countries have built their own Y.S.A.P.
One successful Y.S.A.P.
I mentioned in my paper was from New Zealand.
The third approach is the investigative genetic genealogy or IGG.
It was very hard in the past few years in the United States and it did help
solve loss
of cold cases.
This approach is similar to Y.S.A.P.
database searching in terms of detecting distant relatives.
But the foundation is completely different.
IGG searches the high density SNP database, not STIA database.
In IGG can detect both male and female relatives.
This is the typical workflow of IGG.
Usually the law enforcement agencies or crime labs will also source the whole
workflow to
the external service providers.
And the service providers will use microarray or whole genome sequencing
technology to generate
high density SNP profiles from the crime case, for example.
And to obtain a meaningful searching results, usually the subs lab will require
at least
one nanogram of DNA from the sample.
And the SNP profile will be searched against the SNP database like Jup Match to
look for
potential relatives.
And Jup Match used an algorithm to measure the IBD segments between a pair of
SNP profiles
and determine their relationships.
And if we with enough number of SNPs, it can detect very distant relationships
like the
fourth degree or fifth, sixth, very distant relationships.
This first three steps probably won't take too long to finish.
More efforts will happen in the last step, the genealogy search in the PAPD
records.
It relies on the individual genealogists to dig into the PAPD record and find
or confirm
the relationships.
So it may take weeks or months to get it down.
And there's no automatic software available to do this.
So it's very lab intensive.
The most famous IGG case was the Golden State Killer Case.
And I believe all of you heard of this case.
This was also a zero-kill case.
In this case, SNP profile was generated from the promising sample and the
search against
the Jup Match database.
And based on the search results, 25 family trees would generate as potential
relatives
of the PAPD.
And then the other information was used to rule out the unrelated candidates
like the
age, gender, the place of residency, etc.
And eventually only the angel was slapped and he was found through third cousin
relationships
in a huge family tree with about a thousand people.
So you can see there was lots of work involved in the genealogy search and much
more intensive
than just generating DNA profile.
And I talked about three different interactive DNA-based searching approaches.
Next I would like to compare them from multiple different perspectives.
Both from low search and the wide-base searching are conducted in the
accredited forensic DNA
laboratories.
And use this validated standard forensic technology to generate either a soma-
stia profile or
Y-stia profile by the DNA analyst.
But IgG usually was also to the external service provider.
And those service provider usually will use either microarray or whole genome
sequencing
technology to generate high-density SNP profiles.
But those technologies are not validated by forensic labs.
The forensic technology are more sensitive.
It can generate four STF profiles with about 125 people in DNA or even lower.
And the IgG service provider usually requires at least one nanogram of DNA to
get meaningful
SNP profile, but the more they better.
Social searching searches against the soma-stia database like codis, why STF
database searching
searches, why STF database?
Both STF databases are managed by the government usually in the profiles in
those database
are verified and the database are secured.
But the IgG searches against the public-sasible database like GIMMATCH database
and the profiles
in those databases are usually uploaded by the average people like us.
So the profiles are usually not verified.
So the databases are less secured and more vulnerable to attacks.
And the data leakage can happen by hacking or special design queries.
There have been at least two papers to talk about how to use special design
queries to
get the SNP profiles out of GIMMATCH database.
All these three approaches use different data.
So the way to measure the relationships are different for different approaches.
The familiar search used like a racial or number of shared alleles for IBS to
measure
the relationship between a pair of soma-stia profiles.
And it can detect the first degree relatives.
The YSTF database measures the number of mismatch locile steps for a pair of Y
STF profile.
And theoretically it can detect all males in the same lineage.
But if it's too far away from each other, then the mutation accumulates so it
will make
more difficult to detect the very distant male relatives.
IGG measures the total length of IBS, IBD segments between a pair of high-
density SNP
profiles.
And sometimes MIMMATCH also uses IBS to determine the relationship.
And depends on the number of SNPs in the profile, it can detect very distant
relationship like
the search to be or more distant relationships.
There's not official summary on how many successful cases on each approach.
But based on the the researchers and the information we know for familiar
search, they probably
have several hundred successful cases mostly in US and Europe.
And several hundred successful IGG cases mostly in US.
And there are probably at least tens of thousands successful cases for YDV
search, but mostly
in China.
In terms of the cost, the most familiar search in YDV search happens within the
crime labs.
So you can use existing instrument.
The cost of DNA test will depend on how many samples you want to process.
It may cost 700 to $7,000.
If you don't have a YDV and you want to build a database by re-typing all the
reference
samples, then it may cost a lot of money.
But if you just accumulate all the white profiles generated from the record
cases, then it won't
cause anything.
The IGG, if you want to do the IGG within a crime lab, you need to purchase a
very expensive
instrument and maintain instrument, very expensive.
And so that's why most of labs will also allow the IGG service to an external
service
provider, which may cost several thousand dollars or more, depends on the case.
In terms of the general time, familiar search and YDV search usually only takes
days to
weeks since it happened within the crime lab.
Then IGG may take longer weeks to months because of the extensive genealogy
searching.
This one thing I want to point out here is that the YDV search may also need
genealogy
search for some cases.
Now the identity of the profiles in YDV is usually unknown.
So if you get ahead, then you know who to investigate.
But if the relationship is too distant, you are not so sure if the relationship
is true,
then you need to have a genealogy search to confirm the relationship.
Also explore the other lab tips around the hit.
So in terms of legislation, familiar search allowed by at least 12 states in
the US and
are many other countries, but it's prohibited by Maryland, Washington, D.C and
Canada.
Why is that database searching covered by the contentless reservation for Osama
-Sti
database since the content DNA law do not specify which market to include the
database.
In IGG there's no legislation as of last year, no regulation for IGG, but has
been introduced
and proposed by multiple states in the US, either banning or regulating IGG,
depending
on the states.
The advantage of the familiar search and YDV search are that they use existing
technology
in the legislation.
The technology are validated in the proficient test data verified and secured.
Fermill search can only detect first degree relatives and it may generally
large can
they list off searching, so it will still lie on YSTR to fill out the on-
related.
YSTR, YSTR database have higher searching efficiency for male lineage and
sometimes
we can even associate Y2 surnames.
But it can only detect male lineage and if you don't have YDVs and you want to
retype
all reference profiles then it may cause a lot, but if you just accumulate the
work
profiles from regular cases it costs nothing.
In IGG it's able to find distant relatives and female relatives as well, so it
's better
than Y, but the technology is not frantically evaluated, but probably okay
since it's just
for investigative purpose.
IGG is heavily aligned on the public snap database, but the snap data is may
not be
secured and the people have lots of concerns on the privacy inclusion.
And they also have a little line on the genealogist to do the published record
digging, so it
will take a long time and there's a lot of effort involved.
In a short summary, all these three approaches are viable and they can increase
the value
of DNA day spacing.
YSTR searching and IGG are more advantageous things they can detect more
distant relatives
and it should be able to assist in more challenging cases.
But familial searching can be performed more readily within the forensic lab
settings.
The database is the key, without database there's nothing you can search.
In the US, Osama is tier database and the genealogy database are ready for
search, but
many other countries may not have the genealogy database ready.
In the US, probably not too many labs have their own local Y database, but if
you start
accumulating your Y profiles through regular cases and in the future, I believe
you will
have some good surprises.
Just like many years ago, people upload their SNP profiles to gene match their
base.
And each country or state may have this own unique situation in economics,
technology
availability, culture, legal considerations.
So each state or each country should decide the best approach to fit those
circumstances.
In our center, we have touched all three approaches.
I developed a familiar search software and Peking Fs edition many years ago.
It has been used in many lab labs.
And if you are interested in this software, you may contact me or the tech
transfer office
or university, or believe or tech transfer office have decided to provide this
software
for free or really low cost.
In the currently, we have an NIJ funded research project to use high density
SNP panel for missing
person cases.
The challenges are how to get good SNP calls from the bone samples, relative to
great
examples, and also how to develop a good algorithm to deal with those low
quality SNP
calls.
And I am also working on a pilot local Y-steady base project.
And hopefully in the next few years, I will have some good stories to share.
Finally, I would like to thank the invitation for my friends in Sumo Fisher,
Andrew, Sherry,
Christina, and Jonathan, and I would like to thank the support for my colleague
in UNTCHI,
Bruce, John, Christina.
Also I would like to thank Mark, Brian, Colin, Chris, Seton, and Carol for
providing their
data in this study.
Thank you.
And now I will pass it over to my colleague, Wayne Wickenhaiser.
Thank you, Mike, and Dr. G.
And I will start with my presentation now on the forensic applications of YSTRs
, cases
and databases.
So my name is Ray Wickenhaiser and I am Director of the New York State Police
Crime
Lab System.
Just first for a disclaimer, any of the opinions that I am going to express are
those of myself,
not the New York State Police, Azklat, or any other entity.
So just to provide you with a brief outline of what I am going to talk about
with you
today.
First of all, talk about investigative leads and what we do in forensic labs
and what
our process is.
I am going to talk about YSTR case applications specifically and give some
concrete examples
of some great case successes we have had.
Starting with Anna Schof, which is a great YSTR case, followed by a familiar
search in
cases, one of her big successes in solving decades long on Solve case of Wendy
Jerome,
homicide in Rochester, New York.
I am going to talk about familiar searching in New York, what we do and some of
why we
do it and then also touch upon investigative genetic genealogy.
And then to summarize, provide a path forward and that is how do we provide
data-driven
investigative leads.
So to jump into my talk, evolution of forensic science, things have come a long
way in the
last number of decades.
So what is forensic science, really the application of science to the law and
the mission we
have in crime labs is, as we are applying that science to law and our cases,
our goal
is to maximize the value of evidence.
So we are in a privileged position where we provide independent scientific
analysis,
its objective, it is based on data, it is reproducible and it is something we
can demonstrate.
And so that adds a lot of value to investigators and investigating their cases
and to the justice
system as a whole.
So the services that we provide, we identify items, so you think about drugs,
toxicology,
is it blood, is the semen, fingerprints, so forth.
Comparing it to known sources adds a lot of value.
Now where did that blood or that semen come from?
And providing those associations.
So we absolutely can eliminate sources, potential sources with certainty.
So that is great and very valuable, especially those people are being
eliminated.
And when we include somebody, we can give an idea of how much weight is and how
valuable
or how probative is that inclusion.
So that adds a lot to investigators and the judicial system.
And what I really want to emphasize today is the investigative information and
helping
investigators develop and eliminate suspects is a really a great service, is
something
we provided for years, but has really come to the fore with some of the tools
we have
now which we're going to discuss.
So when you look at some of the new investigative tools that I knew, I may be
saying well,
why stairs aren't that new, but relatively speaking they are and some of their
applications
are.
So I'm going to be really searching and talking about that in the case.
Investigative genetic genealogy has really come to the fore and being able to
solve decades
long cases.
So again, a great technique and talk about that a little bit.
So what you're really seeing is a shift from direct one to one matches where
you have to
have a suspect.
Now using databases, we're able to provide those investigative leads in terms
of here
is a person of interest where there was not one previously before using that
database
before using a database.
Now we're expanding comparisons, particularly with familial search to include
kinship information
and that's a bit of a mental leap.
So we're not doing a direct comparison, but in fact an indirect comparison, but
using
DNA and our knowledge of how it is inherited can really add value in terms of
expanding
the size of that database.
And something again, the privacy advocates have a point, something we want to
take very,
very carefully and sensitively, but it does add value to investigations and
really again
back to our mission and that is maximizing evidentiary value.
So when do we do this?
So right now we certainly are reacting to investigator requests.
They're asking what cases and demanding cases be done, but certainly we have
potential crime
labs because we're an aggregator of data, people make a lot of requests from
different
jurisdictions and those go into our backlog and eventually into unsolved cases.
We can review our cases for candidates in terms of which cases are potential
candidates
for this new technology and then reach back to our investigator.
So it goes both ways.
They ask us, now we're in a position where we can look at what we have and say,
hey,
have you thought about this for this case?
So speaking specifically about YSTRs, as you may know, the Y chromosome is for
males only
or in males only.
So only certain kind of cases where there is a male, of course that's a lot of
our cases
as males are suspects, the vast majority of the time.
But because all the short tandem repeats are STRs we look at are linked, we can
't use the
product rule, which means if they were independent, we'd multiply them together
, that's not the
case.
So hence we have to apply much more conservative statistics, counting method
and so forth.
So as a result, having the same Y chromosome match with an STR profile is going
to be a
lot more common than autosomal inclusion.
So it's really not individualizing.
But what we really can use them for when we have a Y as a male who's in a
mixture with
a female, which is quite often the case, particularly when we have sexual
assault cases, there's
a bunch of female which may drown out effectively that minor male component.
That's where the YSTRs is very, very valuable.
And it's also when you have a relative who's a male, so we're searching for
male candidates,
such as what we do with familial searching, using that Y where you know the
autosomal is
different, but using the Y as a tool to eliminate really lets the DNA do the
heavy lifting for
familial searching.
So we can much more quickly eliminate candidates, whereas we just dumped that
whole list on investigators
that would really be overwhelming.
So that's where the lab can really provide a tremendous service, let the DNA do
the work.
So the first case I wanted to discuss was great YSTR case success at a homicide
case,
August 10th, 2014, the body of 17 year old Anna Rose Shove was pulled out of
Lake Ontario
at Brennan Beach.
She was vacationing at an RV resort near the beach with her family.
There was some initial suspicion that Anna's death was more of an accidental
drowning because
she was a pretty good swimmer, but because there was other factors including
she had
some bruises cut to her head and around her eyes and her mouth, there was some
suspicion.
So certainly any unattended death is this you would look into it and there
certainly was
some clues that indicated perhaps it was not an accidental grounding.
So a number of items were submitted to the crime lab.
There's a sexual offense kit that was collected.
That's her autopsy was submitted.
The date that's indicated there on my PowerPoint slides, there was additional
items, including
her bathing suit purse.
There were fine in the lake and notably were weighted down with rocks and that
was submitted.
So again, evidence there that perhaps it was not an accidental drowning.
Other people on the beach that day indicated that she was seen with a 32 year
old man,
a name Stephen Sotanic.
Hope I'm not mispronouncing that too badly.
And a buckle sample was collected from him and submitted on the date also
indicated in
2014.
He had a visible scratch on his neck, one questioned by state.
And again, I just want to note all kinds of cases, particularly DNA and
especially
why it's part of the entirety of the case.
So somebody's not going to be convicted on just the YSTR alone.
But nor should they probably on DNA alone.
You're really looking at the composite of the case.
So you can what you see is a picture that's starting to become clear here.
So the DNA report was issued also in 2014 on samples from the autopsy and swabs
from
her bathing suit and very notably a YSTR DNA profile from the fingernail clipp
ings and
scrapings from the right hands of Anna Rose show match the profile of the
suspect.
Really interesting to me is she was found in the water.
There was not an autosomal profile but a YSTR.
So now we'll take what we can get to me.
It was just astounding that we're able to get a result given the situation that
her body
was found in.
So remarkable work there and obviously great tools help you do that.
So less than four hours after deliberation, the jury found Mr. Satanic guilty
of second
degree murder in Anna Rose showed death and he was sentenced to 25 years in
prison.
During the trial he testified in his own defense.
He admitted he had spent some time with her in the beach and hunter.
He stated that he'd hit her purse after finding what he thought was a suicide
note.
He also claimed he took her other belongings including the notes to a different
beach where
he built a memorial tour.
And obviously if you're a juror, you have to decide whether that's something
you want
to go with or not and certainly helps to have objective evidence of the DNA
that's helping
you to make that decision.
So the next case I wanted to focus on was Wendy Jerome, a familial searching
case which we
very recently had success on.
And so this was a decades old homicide case committed in 2014 in Rochester, New
York.
So then 14 year old Wendy was last seen alive leaving her home to deliver a
birthday card
to a friend around 7pm on Thanksgiving Day and she had an 8pm curfew.
When she didn't return at 8pm her parents were frantic, reported her missing
and very
sadly her body was found less than 3 hours later behind a school very nearby
her home.
She had been raped, brutally beaten and her throat had been slit.
So here's just a summary of what was called in the paper, what is familial
searching in
the familial search.
We use an algorithm to compare people in the data bank so we're talking about
the CODIS
data bank to this crime scene sample and look for relatives.
So we're looking for an indirect match.
There was a full CODIS profile developed with no hits for a number of years.
So we're really trying to expand that search, develop investigative leads that
we can provide
to investigators.
So a suspect was arrested and I just want to note for everyone, this is just an
arrest.
Individuals are innocent so proven guilty so this is preliminary but just your
background,
they were 20 years old at the time.
They lived near the homicide scene in Rochester at the time.
The two individuals did not know each other.
The suspect abruptly moved away from Rochester.
They were arrested in Florida and a trial is pending.
So I do want to, after highlighting these two cases and the success we achieved
and talk
about just how we do familial searching in New York.
So currently we have kind of a unique process in which we have a department of
criminal justice
services, DCGS where the application is actually made to that department and we
are the lab
that would essentially be the wet lab for that administrative body.
So the request is submitted to DCGS.
It must be a homicide or sexual assault or with exigent circumstances, so major
public
safety threat.
Where significant resources have been expended to try to solve that case.
There must be at this time a full DNA and full YSTR profile in CODIS with no
hits.
There must be assigned MOU from the police chief in the DA and the New York
state police
crime lab then conducts a familial search once those cases have been vetted and
approved
through a process.
We use YSTRs once we have generated a list and with essentially we take the
entire database,
state database.
We generate a likelihood ratio comparing it to essentially how likely is each
individual
related to this crime scene profile.
And then we assess essentially the top candidates and any of those that have a
threshold and
again depending on how many loci we have in there whether it's 15, the
threshold is 10,000
likelihood ratio and 5,000 when we have 15.
So once we've met that, we then let the YSTRs do the work as I had mentioned
previously
and sort through those samples so we can then ideally provide someone who has
meeting the
likelihood threshold and also has the same YSTR profile.
So we had an extensive validation process that was reviewed and approved by our
DNA subcommittee
because we have a commission and DNA subcommittee that oversees all of the
crime labs within
the state of New York.
So that was approved as I mentioned by the DNA subcommittee as well as the New
York Commission
on Forensic Science.
So as I mentioned, likelihood ratios, we have a threshold that they must meet
and then
as well as the Y testing.
We currently have a validation expansion in the approval process which we're
moving up
to the 20 loci and we have also included a much higher threshold that will
remove that
requirement for the YSTR profile.
So that's really for cases where we have a female suspect or we don't have a Y
STR profile
available because we may have those kinds of cases where we just don't have
enough DNA.
So what it's do is really expanding the candidates that they don't have to have
that full profile.
And we have roughly 15% female in the database so we want to be able to look at
those kinds
of cases as well.
You want to touch upon IGG or investigative genetic genealogy and what that is
is the
generation of a SNP profile which is not the STRs.
We typically deal with a crime lab.
No crime labs to my knowledge are doing them so this is an outsourced where we
still have
some DNA that is at the crime scene from that perpetrator.
We want to put it into a database as a SNP profile and those databases are
direct to
consumer databases not law enforcement databases and somewhat in the same way
as we do with
familial searching essentially get a relative who's a suspect and then work
with that potential
candidate if you will to see if there's any potential there or develop family
trees and
whatnot.
So just to give you a little bit of an idea of brokenness into three phases how
IGG has
done.
So phase one exactly what we do now go to the crime scene get that DNA develop
an STR
profile put it into co-host.
Phase two which I'm going to focus on a little bit more is really that gene
alogy component
which is new to us new to crime labs and can be done totally externally to a
crime lab
and then back to phase three which now the investigation has a suspect gets a
known sample
bring it to the crime lab we do an STR profile and do the direct comparison.
So phase is one in three entirely what we're used to done in a crime lab what's
new is
the stage two but it's the key of how we get a suspect and I do want to kind of
touch
upon that as well.
So we're in the process of working on cases as many of you are within the state
police
so we would look through our cases to come up with candidates as well as talk
to our investigators
and we've been able to submit cases you talk to us about cases that way as well
So talk about phase two just how that is done.
So once we do have sufficient sample that's both quality and quantity so we're
looking
at least on an anagram at this point and working with whatever outsourced labs
in order to
make sure that there's a sufficient quantity there they will develop the SNP
profile from
the crime scene sample.
The next phase is for that to be put into those couple of databases notably jet
match
and family tree that do have individuals in the databases who are permitting
their samples
to be used to be searched against and candidates are then surface so they're
known as matches
but essentially they're potential relatives and from those potential relatives
a genealogist
can then essentially build a tree to look for a most recent common ancestor.
So what that is really doing is going to if you have somebody who says the
first cousin
to the crime scene that would mean that you'd have to go up to grandparents the
second
cousin you'd go up to great grandparents and so forth so you're looking at
somebody who
is the most recent common ancestor of the candidate that you've got a match to
and the
actual crime scene.
Then once you've had that most recent common ancestors you've got somebody who
's an ancestor
in common you build the trees back down and then look for descendants and once
you have
those descendants you're looking for where the family trees may actually cross
because
ideally you'll have multiple individuals ultimately you're looking for step six
which is now you've
got those family trees you've got somebody who has both sides or relatives on
both sides
that come together in one family so you've got somebody in the maternal and
paternal side
and the descendants that fit the profile so they lived in that same location
they've
been the right age they are the right sex typically male other attributes
perhaps that
you might be able to include hair color eye color if those are the kinds of
things you're
able to develop in the SNP profile and hence develop a suspect so a lot of work
there but
again the same goal as familiar searching is a number of the other things we do
developing
something elite that we can turn over to our investigators and again back to
what they're
you know known to be doing get that profile and then get it back to the lab so
we can do
the direct STR comparison.
So one of the things we really again my opinion really emphasizes is always
having the code
of cell eligible crime scene profile before you proceed you want to have a kind
of constant
flow and treat the cases the same inequitably and how you're going to do that
and use the
kind of most reliable techniques that are going to solve cases first and then
build upon
those in the cases that you have not solved so what we're doing here and what I
would
really subscribe that we would all do is apply both all of your techniques to
whatever policies
out there department health policy on conducting investigative genetic gene
alogy and it's very
very similar to what we're doing with familial searching so we're restricting
it to major
cases I talked about having sufficient sample and by having a codus profile
first and not
getting a hit what you ensure the following one is that it sounds so simple to
say it
it's a codus qualifying profile but what that does is guarantee a whole lot of
other
things one is the profile is quality it's been generated in a credited lab that
is audited
to the FBI TOS it's got all of those requirements that we all know for a crime
lab your audited
your professional testing of qualifications all those things have listed and
more it's
independent objective as you've developed that profile from a sample you know
is that
the crime scene reasonably from a suspect and you're going to do that well in
advance
of having any suspects to compare it to so it's going to be independently done
it's
objective it's going to be probative ideally you've got that smoking gun so to
speak of
the any that's found in a really telling evidential position and sexual
assaults with semen that
kind of thing would obviously be that kind of sample equitable and as I
mentioned using
the codus first and then progressively layering the more aggressive approaches
to solving
your crime ensures that you treat cases equally you're doing it a very
organized and a good
flow kind of good manner that's very consistent and then you wrap up the
efforts for those
cases where you would have unsolved or have installed them by earlier efforts
so it really
helps to have that organized flow of treating cases the same and equitable so
the key components
I guess in summary maximizing evidential value the crime scene evidence is to
use that kind
of a layered progressive approach and really using technology that we have to
save investigator
time I mean they're fantastic but they can only do what they're given in terms
of tools
so using the YSDRs as I mentioned for familial searching and limiting a number
of candidates
so once they're really once you know you can say our strong candidates but
including training
as well as that knowing that these are family members we really have to treat
them carefully
they are not the suspect but they're getting close using databases responsibly
inviting
by terms of service so I mentioned letting the DNA save and directing
investigative work
and of course the great example Golden State kill and it really started at all
unsolved
in four decades with traditional techniques along comes investigative genealogy
and some
fantastic work solved in 63 days from the data that database entry speaks for
itself but something
we really have to use carefully and judiciously as I mentioned and demonstrated
how we use
our familial search screenings and those YSDRs to screen the candidates again
protecting
their rights so we really want to give leads that are you know ones that we
know are as
good as we can make them having MOUs with law enforcement and prosecuting
attorneys to make
sure that if we're going to do the work which is considerable that they're
ready to move
ahead with that case and having that training to ensure that the leads are
known to be related
individuals not that individual and so that it can be something that really is
done properly
so want to make sure that we use the tools judiciously and we make sure that
they're
there for the next case and the next case and next case and something that
everybody can
get behind given that there are privacy issues and we really want to be
respectful of that
so providing investigative leads to investigators has always been one of the
most important jobs
in the crime labs and so these tools are providing more of the same.
With that I'd like to acknowledge some fantastic efforts of our biological
science case work
data bank and CODA staff who worked on these cases and on our familial
searching and investigative
genetic genealogy initiatives so thank you for your kind attention and with
that I'd
like to open us up for some questions.
Fantastic thank you Ray for that great information and insight.
It is almost time for the Q&A portion of our webinar if you have not ready
please take
a minute to ask our presenters a question using the Q&A dialog box on your
screen.
While you think about your questions we're going to throw up three polling
questions for
you quickly the first one.
Does your agency utilize Y chromosome analysis?
Yes YSTRs for sexual assault YSTRs for homicide cases YSTRs for mixture
analysis Yscreen for
sample screening or no?
Let's see what everyone says.
Looks like the majority of you say no followed by closely by yes you use YSTRs
for sexual
assault cases.
Have your agency used familial search or genetic genealogy?
Yes we have used familial search.
Yes we have used genetic genealogy.
No but we plan to use familial search in the future or no but we plan to use
genetic genealogy
also in the future.
See what you guys said?
Yes you have used genetic genealogy which makes sense since it is one of these
fastest
growing disciplines within forensic science.
Alright last question.
Would you like to receive more information on today's webinar topics?
Yes please contact me about YSTR analysis.
Yes please contact me about Yscreen.
Yes please contact me about autosomal STR analysis.
Yes please contact me about other ways to generate investigative leads or no
not at this time.
Let's jump into our Q&A then and answer any questions that have come through
the console.
Alright presenters.
Our first question is can YSTR analysis discern the differences/similarities
between father
son and then onto more distant relationships such as uncles and cousins?
I can answer that question.
So in general you see the more distant relatives will have more mutations each
other.
But if you compare the father son pairs and the cousins pairs the difference is
not that
significant.
You may use the...
There's no clear cut you can differentiate the father son pair in distant
relatives.
But to differentiate the males in the same lineage versus the unrelated you
probably can
use with WIFI for example may use four or five mutation steps to differentiate
the relationship
or not.
But hard to tell if it's what relationship.
Okay great thank you for that.
Next one you know both of you talked about genetic genealogy.
Are you familiar with many cases from outside of the US where agencies are
using genetic
genealogy to generate leads?
I haven't seen some reports.
The out of US there's been some cases used high density petals of some cases.
For example there's a report from a group of research from Sweden.
They use the whole genome sequencing data search against a jet match database
to determine
the identity of a missing person and they were able to narrow down the region
to a particular
country.
The Croatia.
So I will say that's a success by providing an impact of leads.
I'm not sure if they eventually determine identity but at least they provide
some impact
of these and also there's other countries like China and other countries.
They're also trying genealogy technology.
Okay great.
Next question.
Generally what are the criteria used to determine if a case would qualify for
familial
search?
Okay so I can take that question Johnny.
So the criteria that we typically use is that you have a full profile and code
is with no
hit and that automatically includes all those quality things that I had
mentioned previously.
Also that you leave and I don't like to term exhaust at all investigative leads
You've certainly put a lot of energy into that case and it remains unsolved.
So I guess the question is are all the leads ever exhausted?
So we tend to just stick with a lot of investigative energy has not been able
to solve that case.
We have a full profile.
It's a major case so we reserve these really for sexual assaults and homicides
and that
is really a probative sample.
We do have MOU that we put into place that ensures that investigative agencies
are willing
to proceed with the case.
We really want to spend all this effort on a case that is going to more likely
yield some
results.
So those are the things we have in place in New York State.
Okay perfect.
We've been to our next question for you Ray.
What kind of information is included in the MOU between the lab and the police
slash DA's
office?
And just in case some of not familiar, can you give us the acronym for MOU?
What does that stand for?
Sure.
Thank you very much.
So MOU stands for Memorandum of Understanding and what we really want is some
kind of a
commitment because these are cases that we're going to go through a lot of
additional efforts
to do not only the statistics on it.
We're going to do a bunch of Y testing and sort out with additional testing
some investigative
leads.
So we want that memorandum of understanding on both sides of the investigators
to say
this is a case that this will really help them.
That's not solved.
That it meets the criteria.
But also on the prosecutors side, the lawyer, is this something that with the
positive results
and they find somebody, they're going to pursue it.
So we just really want to make sure that all the parties that are at the table,
that it
meets the criteria.
It's an unsolved case.
The lead is going to really help the case and they are going to move ahead.
And what we have found is there have been a number of cases that at the lab, we
think
they're still in pursuit, but the case may have been solved some other means.
This is just the way to make sure you're going to put in this additional effort
Everyone is committed to move forward.
And it's actually on paper so we can get that commitment.
So that's what we do.
So the extra mile just to make sure everyone is.
All right, Johnny, we have a few questions for you.
So let's grab you back in here quickly.
In Chinese laboratories, which Y STR type and kits are typically used?
So they mostly use the Y follow-up plan and developed by a similar fisher.
And then there are a few other kits developed by Chinese local companies.
But mostly they use Y follow-up plan by similar fisher.
Okay, let's stick with China for a second.
With less than 35 Y STR loci in the Chinese database, how many hits do they get
per Y
STR search typically?
So it depends on the criteria you set.
You have many mismatches you allow.
Let's say you have 35 markets to search and you allow up to like three or four
mismatches.
Then you will get more hit than just four match.
But typically they will get about 10 to 20 on average hit.
It's a four match in the database.
Okay, let's turn to Y STR databases in the US.
How many Y STR databases with reference profiles exist in the United States?
I'm not sure how many.
I don't know the answer.
The labs allow to upload their Y profile to codists in this no statistics on
how many
Y profiles in codists.
And then there's an order of upload the codists.
So some labs are probably already cumulating some Y profiles in their local
name system
or something.
But there's no such kind of statistic.
Okay.
And now you can both probably answer.
But let's start with you.
What are your thoughts on Y STR database searches when you don't have an autos
omal STR
profile?
That's fine.
So it really depends if you have a Y database to search, yes, then you just
search it.
You don't lie on the Osama's database.
When you get hit in a Y database search, then you can just in that way hit the
family.
So then you type the Osama's STR profile from those family members.
Then you don't lie on the Osama's STR database.
Gotcha.
Okay.
Right.
Anything to add there?
Yeah.
And I would add, you know, anything they provide a lead can be of value.
I guess the challenge with Ys is because it's not going to be specific to that
mail but
include anyone in that whole patrilineal line or at the whole mail side of the
family
line, the challenge there is that you're going to look at folks that are not
the perpetrators.
So as well as long as the investigators are well advised and hence the training
part,
just to know that you have to be extremely careful with folks that they may
help you
get to the suspect, but they're not the suspect themselves.
So it's just that taking the care that you know you're going to get other folks
that
are included that are not the perpetrator.
Whereas with autosomal, we're used to when you get an inclusion, you know, it's
that
person as opposed to this indirect information.
Gotcha.
Okay.
All right.
Another one for you, right.
What is there a main reason why most NDIS, oh, participating labs haven't
incorporated
the IGG into their toolkit?
Is it cost?
Is it something else?
Yeah.
So it's actually several.
Certainly it's costly and it takes not only the cost part, it's time and any
new technology,
like this is really, you know, we think of it as old news that happened in 2018
in terms
of forensic timing, that's very quick.
And so because it's technology that typically we don't have in the lab to get
the instruments
in, to do the full validation, that's a big lift for many crime labs where we
're already
in a backlog situation.
So it's a kind of combination of the time, the instrumentation, but it does
come down
to cost and resources as well.
But to come up with that kind of overnight is a big challenge.
So right now most of, most everyone is outsourcing or sending to labs that are
not kind of traditionally
forensic or crime labs.
Over time, I think it's something that a number of us will be looking carefully
at.
This is something we can bring in house, you know, apply of our forensic
expertise.
Absolutely.
Okay.
Johnny, let's turn to you quickly.
The familial searching software that you created, is that used by the Texas
Department of Public
Safety Code Islam?
Yes.
The STELCIS, yes, and the Texas DPS lab is what's the first lab using the
software?
Yes.
Oh, okay.
That's great.
We have a question about the X chromosome.
What do you think of the X chromosome?
I mean, is it helpful for forensic investigations or is it not really of great
use?
I would say the X chromosome may be useful for certain tinsh cases, but
probably not
a best tool for searching relatives in the database.
Especially, you don't have, usually you don't have XTS in the database.
And also, even if you search, it's very, very complicated.
It's not as simple as Y or some also.
I would say X chromosome has a very limited value in the database age.
Okay.
Now, this is kind of a broad question, but maybe you can at least give us some
insight
into how are databases validating, certified, accredited, how are these things
regulated
and you know that we can use them?
So I can probably jump in and take that one.
Perfect.
For any of the databases, and I'll call the NDS databases and National DA Index
System,
those are regulated by, I believe, the DNA Identification Act of 1994.
So it's, there is a requirement that anybody who participates in those NDS, we
call them
kind of CODIS as an acronym, but databases have to be accredited.
They have to adhere to the FBI quality assurance standards.
So you've got all of that mandatory requirement.
You've got to be audited externally every two years internally, every year kind
of offsetting
that, profession's suggesting a lot of requirements.
So accreditation and that mandating just makes sure that there's a lot of
oversight.
Then a lot of it is security aspects and so forth.
So I feel that databases that are with the NDS system are extremely well
regulated.
Now the challenge when you go to the direct consumer is while those companies
may very
well look after them because it's their business, so you would expect they are.
But because it's not under the law or those kind of mandates, there you've got
terms of
service and all of those things that are used for genealogy.
And that's because it's outside of the crime lab regime that's really kind of
up to those
companies.
But I'll just say I have a tremendous amount of confidence in the quality and
integrity
of the NDS system and the CODIS profiles that are put into that system.
Absolutely, it's important.
Hi, Jean-G, another one for you.
Are you using microarrays to generate a high density S&P profile or are you
using whole
genome sequencing to then create the S&P profile to upload into jet match?
So in all sense, we are testing both technology for missing person cases.
Most microarray and whole genome sequencing, we're still working on that.
And the initial data shows that the microarray has pretty good sensitivity
better than what
we expected.
In the future, we'll probably have more data to ensure.
Okay, perfect.
You earlier in your slideshow, you showed an indirect database searching
comparison chart.
Is there a way to access that after this webinar is concluded?
Yes, it's on what the table is actually part of the paper that's published and
you can
send whatever you want to have the paper if you don't have access to the paper,
just send
an email to me.
I can share with you.
Perfect.
And I'll note here, Audience, that this webinar will also be available on the M
AND on the
forensic website.
So in addition to using the paper, you can get the chart from the paper.
You can come back and watch this anytime you want and you'll be able to see the
chart.
I'm going to have time for a few more questions.
Let's run through them.
Maybe just a couple more.
All right.
John G., what about a mitochondrial DNA database?
Can you kind of talk about that and what that would be like?
So I believe codis allow you to upload mitochondrial DNA profile to codis
mostly for the missing
person cases.
But probably the number of samples are probably not that big in codis.
And we do know that the KLDOJ did use mitochondrial DNA screening in one of
their familiar searching
case.
So it shows that mitochondrial DNA that they can be used in familiar search.
Okay.
For familial searching.
Have either of you considered searching anything less than a full or a similar
profile?
So I can speak to that one.
So the quick answer is yes.
When we started in New York because we were given a very quick mandate to get
familial
searching up and running, the original policy was for full profiles only.
However, as we gained experience, we've now done a validation with less than
full profiles.
We've had to increase our likelihood ratio just to make sure that we're not
getting
adventitious hits, we want to kind of narrow those candidates, the ones that
are more likely
to be actual relatives.
So we do have a proposal that before DNA subcommittee and commission that
allows us to work with
profiles that are not full profile.
So certainly somewhere there's always a line, there's always a bar, but we
really want to
try to provide good investigative leads.
But because we're using YSTRs to help us doing a sort there, it really does.
We're letting the DNA do the work in terms of doing extra lab work for
investigators
and really have much more targetedly.
So we are hoping to get that approved and in their future so we can expand the
types
of cases that we look at.
Okay, great.
Let's wrap up here with two more IGG questions.
When you're working with IGG, how do you separate the mixtures for the IGG case
work?
So I can probably speak to that too.
Our challenge there, we're really using technology that we're not, it's brand
new to us in the
crime lab, we're new to SNPs.
So because that's the profile we're currently looking at, we're relying on
those vendor
outside labs.
And at this point, you really need or you have to have a profile that you can
subtract.
So you can deal with a mixture and it's typical sexual assault where you're
going to have
the male assailant, female victim.
And if you have her profile, so it'll be about a 50/50 mixture, they can
subtract that out.
So you can deal with straightforward two person mixtures in rare circumstances
where you have
a very minor component of like a 10 to one ratio.
They can just subtract that out.
But it's really limited to that much more simple mixtures where you have the
full known
profile of one of those two people, which typically is going to be your victim.
So there could be some mixtures done.
But I'll just say it's early and hopefully as time goes, we'll be able to work
more and
more with mixtures because we see those in the majority of our forensic case
work.
That makes sense.
All right, last question and it's a good one.
It's an interesting one.
Do you think that the general public S&P profile added to jet matched, let's
say, will
decrease in the future and be limiting for IGG?
So I actually think that it will go the other way personally.
And of course, I'm very sold on what we do.
And I think we're going to take care to do it correctly.
I feel that as the public becomes more and more educated in what great work we
're doing,
not only to solve these cases, but to identify previously unidentified human
remains.
I feel they'll get more confidence to allow their profiles to be used by law
enforcement
or medical folks to identify these unfortunately unsolved cases where someone's
gone missing.
So over time, I think that will just grow as people become more knowledgeable.
But again, we'll see how it plays out.
But so far, when you look at the cases that have been solved, I think the
product speaks
for itself.
And again, we just have to be very judicious and it's used to make sure that no
one makes
any missteps in terms of privacy and so forth.
Absolutely.
Okay.
All right.
Well, audience, that's about wrapped up the time we have for the Q&A.
In 24 hours, this webinar will be available on demand if you would like to
watch it again
or share it with friends and colleagues.
Additionally, you will receive an email with information on how to obtain CE
credit documentation
for your participation today.
Thank you for attending the first webinar in the 2021 Future Trends in Forensic
DNA Technology
series.
The five part series continues on April 1st with the Rapid DNA Booking Station,
Implementation
Considerations and Best Practices.
You can register now on www.forensicsmag.com.
Please be sure to check your email for more information on how you can obtain
CE credit
documentation for your participation in today's webinar.
And we look forward to seeing you next time.