Human identification of unknown samples can pose challenges to forensic genetic laboratories, especially if they have been severely affected by environmental factors. What plays a role here is not only the fact that Short Tandem Repeat (STR) markers may be too large in fragment size for successful analysis, but also that the investigative work requires other information that cannot be generated by STRs. For example, if the STR profile of an unknown sample does not lead to any hits in databases, other characteristics, such as externally visible traits or the biogeographical ancestry can be derived from the DNA of the sample and provide crucial elements in the identification work. Both phenotypic and ancestry-specific panels require a high number of genetic markers to offer significant information. These markers primarily include Single Nucleotide Polymorphisms (SNPs) as well as insertion and deletion (indels) markers, which can provide successful test results even in the presence of highly degraded DNA. Through the emergence of Massively Parallel Sequencing (MPS) technologies, these markers can be successfully examined in large numbers simultaneously in the same assay. We have evaluated two marker panels known as COMBO and FORCE, which we will discuss in more detail. COMBO includes autosomal SNPs (VISAGE Basic Tool), Y-chromosomal SNPs and the mitogenome in a single PCR-based assay. This allows the investigation of phenotypic characteristics and biogeographical ancestry using three different lines of inheritance. Through specific DNA and mtDNA quantification, the amount of primer can be adapted so that all three components can be represented in a single PCR assay despite of their varying copy numbers in the cell. The FORCE panel is based on capture technology and can therefore be used to examine even higher degraded DNA than COMBO. In addition to the COMBO components, FORCE also contains DNA markers for identification and relationship testing, which can be used to identify more distant relatives. The lecture leads through the evaluation and optimization of the two panels and shows areas of application in forensic genetics.
0:00
[MUSIC PLAYING]
0:03
Hello, everybody.
0:11
This is Walter Parson.
0:13
Welcome to HITS.
0:14
Today I'm going to talk to you about two MPS panels
0:18
with the name Combo and Force.
0:21
And I'm calling here from Austria.
0:25
And in forensic terms, we are connected with the PRIM countries.
0:29
So we can exchange DNA profiles between the PRIM countries
0:33
for our data-basing purposes.
0:35
And as you can see, there is a decent number of matches,
0:38
130,000 between Austria and the PRIM countries,
0:41
particularly with Germany, France, Slovakia,
0:45
Czech Republic, Hungary, UK, Spain, and so forth.
0:49
Now this database consists, of course, of STRs.
0:53
They are very polymorphic.
0:55
They are sensitive technology.
0:57
They are pretty easy and straightforward to pursue.
1:04
The standard method that we use to type the samples
1:08
is capillary electrophoresis.
1:10
And the result that we get from this
1:12
is a peak, or several peaks.
1:15
So we understand that within one peak,
1:18
we see millions of molecules that correspond
1:23
to that part of the DNA.
1:25
It's a summative signal.
1:28
Now with MPS, that's different.
1:30
With MPS, we see every single sequence.
1:32
We can analyze every single sequence
1:35
because the DNA is stabilized to a flow cell or to a chip.
1:40
There is no question that this is more informative.
1:42
However, it is also more difficult to analyze.
1:47
Now I would like to see this in conjunction
1:50
with degraded DNA, which we commonly
1:54
see in forensic genetics.
1:57
The forensic field has been moving
1:59
from standard protocols to midi and mini protocols
2:03
by moving the primers closely, more closer to the repeat region,
2:08
in order to increase the success in typing DNA
2:11
that is suffering environmental stress.
2:13
I have been talking about two cases yesterday in mitochondria
2:19
DNA.
2:19
And mitochondria DNA was the first adopter
2:22
of massively parallel sequencing technologies
2:25
where we can use PCR-based technologies,
2:28
but also capture-based technologies
2:30
to analyze the mitochondria DNA.
2:34
Because the polymorphic markers are SNPs.
2:36
The method that we optimized for doing this in the forensic context
2:44
comes from the field of ancient DNA.
2:46
It's called primary extension capture.
2:48
It was developed to sequence Neanderthal mitochondria DNA.
2:53
The way how this works is pretty straightforward.
2:56
The DNA in the extract is adaptoligated and amplified.
3:01
Then we use human-specific probes
3:05
that are specific to the mitochondria DNA
3:08
that we can use to enrich the human portion
3:12
and taking advantage of the biotene
3:14
streptavidin complex.
3:16
We can pull that out of solution and wash away
3:19
the remaining parts.
3:21
So we can increase the sensitivity of the analysis
3:26
by this method.
3:26
And we have done this with the ion-tourine pipeline
3:29
and the gene studio S5.
3:33
Now, today I would like to talk about different SNPs.
3:35
And I would like to talk about the autosomal SNPs.
3:38
And I'm thankful to Chris Phillips
3:40
from Santiago de Compostela for lending me his slides here,
3:45
where he did a fantastic job in showing the evolution
3:49
of the SNPs.
3:51
Coming from the human genome projects
3:54
reaching into the forensic scenery,
3:57
we are having three major international collaborative
4:02
consortia that have been dealing with SNPs in the past.
4:06
As you know, these were SNPs for ID,
4:08
Eurofortunate and Visage that were developing
4:13
for optimizing SNP panels for forensic purposes.
4:17
Now we have SNPs for identification.
4:20
We have SNPs for inferring biogeographic ancestry.
4:25
We have like from a somber SNPs.
4:26
We have mitochondrial SNPs.
4:28
And we have SNPs to predict the appearance
4:31
for forensic cases where this is needed.
4:34
Now here I would like to introduce it to a panel
4:38
that is trying to combine the best of those phenotyping
4:42
markers, and that is the combo panel.
4:45
We take advantage of previous work.
4:47
Here is the Visage basic tool that consists of 153 SNPs.
4:52
It is available as an ion-amplicyq community
4:54
panel at Tomofisher.
4:57
And there is another panel which is called Visage Enhanced
4:59
Tool.
5:00
It contains 523 SNPs.
5:03
It's not yet commercially available,
5:05
but both 3% are autosomal/ychromosomal SNPs.
5:11
And we would like to combine them with y-chromosomal SNPs
5:15
and if mitochondrial DNA.
5:17
Because by doing this, we are able to combine all three genomes
5:23
that are relevant for ancestry estimation.
5:26
And we want to combine this in one reaction, which
5:29
is worth up to 1,175 amplicons.
5:34
Now, of course, you can imagine it's
5:35
important to look at the primer pool quantities,
5:39
because we are looking at different template amounts
5:41
in the cell.
5:42
There is much more mitochondrial DNA
5:44
than nuclear DNA, particularly than y-chromosomal DNA.
5:48
As a standard configuration, we are recommending using 5 times
5:55
Visage Basic Tool or Enhanced Tool, 5 times the y-chromosome
6:00
panel, and 1 time mitochondrial DNA panel.
6:03
We need 2 pools because we have 2 mitochondria pools,
6:06
and that is how we combine them.
6:09
However, we strongly recommend to do quantitative PCR
6:13
to estimate both nuclear and mitochondrial DNA quants
6:17
so that we get an understanding of the amount of target
6:23
in our very sample.
6:25
That can be shifted depending on the sample type
6:27
that we are investigating.
6:30
So I'd like to show you some of the results
6:32
that we did in this study.
6:34
Let's look at sensitivity first.
6:36
This is the combo sensitivity from 1 nanogram down
6:39
to 16 picogram.
6:41
And in this plot, you have green colors
6:43
for the Visage Basic Tool, the red color for the y-chromosome
6:48
panel, and the blue color for the mitochondrial DNA panel.
6:54
And as you can see here, the signal
6:56
is decreasing as expected.
6:59
We have an average read depth for nuclear snips for 5,000,
7:02
and mitochondrial snips for 3,000 for 1 nanogram,
7:05
and that is getting lower down to 500 reads
7:09
for the nuclear snips and less than 300 reads
7:12
for the mitochondrial snips, and that is expected.
7:16
We get 3 mitochondria gnomes down
7:18
to 250 picogram of DNA.
7:21
In that complex, we see only small gaps.
7:25
So these are almost 3 mitochondria gnomes down to 62 picogram.
7:30
And we get some partial mitochondria gnomes down to 16 picogram.
7:34
So that is slightly less successful than we would see
7:38
if we had amplified the mitochondrial DNA panel alone.
7:42
But it's still fairly good.
7:43
At the lowest concentration of 16 picogram,
7:49
there is only 58 of those 1,000 snips
7:52
that come with a read depth below 100.
7:56
And if we look at concordance, and please note
7:59
that the white chromosome axis here runs from 95% to 100%,
8:04
we have more than 99% correct genotypes,
8:08
even at the lowest concentration of 16 picogram of input DNA.
8:12
There are very few little dropouts and new cores,
8:16
as expected with this small amount of DNA.
8:20
However, we need to appreciate we have low performing markers.
8:23
They are even low performing in the original sets
8:28
without multiplying in the panel.
8:30
And that can be due to fade amplification,
8:33
that can be due to strand bias or base missing cooperation.
8:37
Let's look at case work.
8:41
We have mock case work sample here.
8:43
These are the get-knap case work samples.
8:47
And you can see the females don't contain the Y portion.
8:50
And the males have both the autosomal and the Y portion.
8:54
This is only showing the nuclear snips of the combo here.
8:59
We were looking at sex specificity,
9:02
so we were using female samples to see
9:05
if they show any Y background, which is not the case.
9:10
And we were looking at species specificity,
9:12
and it's basically the primates that expect to give signals
9:18
with this panel.
9:20
If we look at matter-converity DNA from the same reactions,
9:22
just different plots, we do see nice amplifications
9:27
and sequencing results for the mock case work samples
9:30
and for the sex specificity.
9:33
And we do see some background signal in species specificity,
9:36
for example, in dog and horse, which has been established before,
9:41
and which is the background of this reaction regarding the species.
9:47
Here are some HID applications.
9:50
We were looking at bones.
9:52
We got bones from Sarador.
9:55
I was talking about this collaboration yesterday in our cases.
9:59
She's from the University of Heidelberg, the Institute of Anatomy.
10:03
She was providing two juvenile skeleton remains
10:06
about 150 years old.
10:08
They were of interest for particular reasons.
10:11
And we have three other bones from Marta Dippenberg.
10:14
One of her main projects is identification of World War II
10:17
soldiers from a particular town called Classin.
10:21
She was providing samples for that respect from her study.
10:26
And if you look at the plots here,
10:27
we see very nice performance of the combo bandle also
10:30
for this particularly highly degraded DNA.
10:35
We do see quite balanced amplification over the--
10:39
and sequence read all over the panels.
10:44
And here in the middle, you see some of the results.
10:47
This is a PCA analysis in terms of biographic ancestry.
10:52
Down here, you see a metroconomer hypergroups
10:54
and vichromosomal hypergroups for estimating
10:58
the phylogenetic background of the maternal and paternal
11:02
lineages.
11:02
So this was a very brief overview of the work that we did.
11:11
This is going to be submitted for publication fairly soon.
11:15
Apparently, the combination of autosomal
11:16
vichromosomal hypercom
11:21
we need two amplifications.
11:23
But it works despite the different copy numbers
11:26
of autosomal vyan mites in the cell.
11:29
That is important because it saves
11:30
limited DNA from evidentiary samples.
11:34
We recommend DNA quantitation because it's
11:37
supporting the choice of the balanced primer concentrations.
11:41
The assay is pretty stable down to DNA
11:44
with the few cell equivalents.
11:47
And it also yields successful results for case work samples
11:50
and for old and degraded samples.
11:53
We have some low performing markers
11:55
that we need to take into consideration.
11:57
They can either reduce the discrimination power
12:01
or they can reduce the prediction estimates.
12:05
The lab workflow is pretty straightforward.
12:08
It's MPC technology, so that is relatively easy to perform.
12:12
However, the analytical start, the part is still a bit complex.
12:17
So we need to have third party software in order
12:19
to analyze all the different genomes.
12:21
We need Harris-Plex-S to analyze
12:24
the phenotypic markers of appearance,
12:27
widely for the vichromosome, MPC for--
12:29
or converge for the metacombra DNA and structural PCA
12:34
for the biogeographic ancestry of the autosomal markers.
12:39
The costs are still a bit high,
12:40
but Homba opens the door to do BGA on limited samples.
12:46
And there may be cases where this is relevant and needed.
12:52
Now, changing gears here, I would like to introduce you
12:54
shortly to another panel that we are working with,
12:58
and that's the force panel.
13:00
It has been developed by Andreas Tilma
13:02
and the group of Shala Marshall at Aftil.
13:06
It's an all-in-one SNP marker set,
13:08
as they say, for investigative genetic genealogy leads
13:13
and for forensic applications here on the left side,
13:15
you see the SNP markers and the purpose.
13:19
They are not using any disease markers.
13:22
They use IgG-compatible markers, and they
13:25
are able to identify fourth degree relatives with this set.
13:30
It has been developed on an Illumina platform
13:32
using 20,000 base.
13:34
The method is very similar to what we saw
13:36
with the primary extension capture.
13:38
It uses capture technology to increase the specificity.
13:43
They were basically using it on control DNA reference quality
13:47
samples and on bones.
13:51
They found a high degree of concordance,
13:52
which was appealing to publish data.
13:57
There was a collaborative study going on in Intellaboratory
14:00
study that the authors took care of,
14:04
and they presented the results of this study,
14:06
which were preliminary results at the 12th Hepid marker
14:09
meeting last year.
14:11
Here is a YouTube video that you've
14:14
very invited to look at to show the results
14:18
and to see more details about this panel.
14:20
The interesting part is that many laboratories that took place
14:23
here used different technologies to see how the force panel
14:27
would work in the different laboratories.
14:31
Now, we wanted to adapt the force panel
14:34
to the Antoine workflow.
14:36
And the way how we did this is we were preparing our libraries
14:39
with the IN+ fragment library kit.
14:43
We were using different cycle numbers for PCR.
14:46
I'm coming back to this in a minute.
14:49
There is the force MIBADE's kit from GASIL albovirosciences
14:52
that can be used for this purpose.
14:56
The DNA is then amplified with the Platinium PCR supermix
14:59
hypodelitic kit, and the libraries are quantified
15:02
with the quantitation kit and pooled to 30p comolar,
15:05
or what we have in this particular sample.
15:09
Template preparation and chip loading
15:11
was done using the ANSHAP system.
15:12
And this was run on the Niantine Studio S5 system.
15:17
Now, we were trying to take it a bit further.
15:20
And in addition to reference samples and bones,
15:22
we were trying to look at hair shafts, which
15:25
is probably the most demanding sample
15:27
that we can look at in forensic purposes.
15:31
For the reference sample 2800M, we only used one nanogram
15:36
while the authors were recommending 10 nanograms.
15:39
So we do see an 88% partial profile with high concordance.
15:44
We do see some a little drop out and a little bit
15:47
of a little drop in.
15:50
For the buckless samples of two volunteers who provided hairs,
15:55
we see about a 65% partial profile.
15:58
Again, it's only one nanogram.
16:00
So there is room for improvement here.
16:04
With the standard hybridization protocol in hair shafts,
16:08
so we tried to capture nuclear DNA in hair shafts.
16:12
With the standard protocol, we didn't get above 1%.
16:16
But we tried to improve this by either increasing
16:20
the cycle number and/or putting a second round of hybridization.
16:26
And then we were able to achieve at least a 6% partial
16:28
and a 17% partial profile in these hair samples.
16:33
It's still very low.
16:34
But remember, it's nuclear DNA of a hair shaft.
16:38
And finally, we were looking at a bone sample.
16:40
This bone sample is 800 years old.
16:43
So it's very degraded.
16:45
It contained about 0.1 nanogram of bone DNA.
16:50
And we were using about less than 2 nanogram full amount
16:55
for the experiments that we did.
16:56
And it resulted in a 71% partial profile.
16:59
Now let's look at the concordance here.
17:06
And as you can see in the hair samples,
17:09
this is very, very able to check for concordance.
17:13
We do see a little dropout.
17:16
Of the concordance rate is about 60% to 70%.
17:20
The improve or the gain in success rate is coming with the price
17:28
of a little dropout and discordant profiles, which
17:31
is a result of the stoichiometric effects
17:34
that we see here with these very low amounts of DNA.
17:40
So the preliminary conclusions that we have
17:42
is that in general, the transition
17:44
from an Illumina-based protocol to ion torrent
17:47
was successful.
17:49
We used optimization steps to increase PCR cycles
17:53
in two hybridizations.
17:55
They improved the result significantly in terms of gain.
17:58
We still do see dropout and low read depth for one nanogram
18:06
samples.
18:06
We need to fix this.
18:07
And I think this could be fixed.
18:11
We see low read depth and that leads
18:15
to loss of discrimination power or limited prediction.
18:17
So we need to take this into consideration
18:19
when booking it such technology.
18:22
Hair shafts contain very limited DNA, nuclear DNA.
18:26
With the standard protocol, we got only very few cores.
18:30
But the increase of second number and hybridizations
18:33
resulted in 300 to 900 markers, which
18:37
could be helpful information in a forensic case.
18:40
It's probably difficult to get predictions.
18:43
But we could have important information
18:45
that could help in a case.
18:46
But we need to be aware of the possibility of this coding
18:50
calls.
18:52
So alternative technologies, probably shotgun,
18:55
would be more promising for hair shafts.
18:58
It seems to be the case that the majority of hair shafts
19:01
just don't include enough nuclear DNA in order
19:04
to run a panel with this technology.
19:07
The bone samples performed pretty well.
19:10
Even they were old and degraded.
19:12
They performed as well as reference buffer samples.
19:15
So we certainly need to get a bit more experience
19:17
and plan further optimization.
19:20
But it could be a promising tool,
19:22
because it's an all-in-one essay that
19:25
takes snips to a level where we can use it for multiple purposes.
19:30
And that is exactly what our toolbox
19:33
in forensic genetics requires.
19:37
I would like to thank Tina and Kat,
19:40
they were and are brilliant scientists, both in the laboratory
19:43
and in analytics.
19:45
They have been working on these two panels for the past years.
19:49
I would also like to thank my colleagues
19:51
at the Institute of Human Medicine and NSPROC.
19:54
Thanks for listening to this presentation.
19:57
Enjoy the rest of hits.
19:59
And if you have any questions, I'm here trying to answer them.
20:03
Thank you very much.
20:04
[MUSIC PLAYING]
20:08
[MUSIC PLAYING]
20:11
[MUSIC PLAYING]
20:15
(gentle music)
20:17
(soft music)