Why a Framework to Interpret Evidence is Required: Illustrated by the Miscarriage of Justice of Amanda Knox
The value of DNA evidence to support a proposition of the presence of a person of interest is present, is typically provided as a likelihood ratio (LR)– often provided as a high number of many millions/billions. However, if this is presented to a court, without a description of the limitations, then the evidence may be misunderstood. The mere presence of a DNA profile that is consistent with a given donor says nothing about 'how', 'why or 'when' the actual transfer occurred. The hierarchy of propositions framework is used to evaluate evidence at different levels. A different likelihood ratio (LR) is calculated at each level; a LR calculated at a lower level cannot be carried over to a higher level. The sub-source level directly addresses the donor of a DNA profile, whereas the source level addresses the type of body fluid that is present in the crime-stain. Activity level propositions are required to address the case circumstances and require a formal assessment of indirect and direct transfer, persistence, and recovery of the DNA profile and mRNA from known individuals and background from unknown individuals. Bayesian networks are used to evaluate the evidence using source and activity level propositions. The dangers of not using a coherent framework to interpret DNA evidence are amply demonstrated by the miscarriage of justice of Amanda Knox and Raffaelle Sollecito. The pair were accused and convicted of the murder of Meredith Kercher in Perugia, Italy, in 2009. A series of appeals and retrials then ensued, culminating in their exoneration in 2015. The case highlighted a series of catastrophic errors that were pervasive at all levels of the criminal justice system.
You Might Also Like
[Music]
Hello everybody, my name is Peter Gill.
I'm Professor Emeritus of Forensic Genetics at Oslo University Hospital.
I'm going to talk to you today about the limitations of DNA profiling evidence
and why it is important to have a framework to interpret evidence.
More recently in the courts, they are becoming much more interested in what we
call the activity level.
Now what the activity level is is not the donor of the DNA profile,
but it's more to do with the how or why or when that the evidence came to be
where it was.
So I have an example on the slide from a court appeal hearing where the judges
said,
it is we think becoming increasingly common for there to be little disputes in
the majority of cases
as to whether the DNA is the DNA of a particular person.
And I think that this summarises it quite nicely.
Our DNA profiling evidence is incredibly powerful and we may give statistics
such as one in a billion,
but what does that mean about the actual deposition of the DNA,
the method whereby the DNA was deposited.
And this was brought to the forefront fairly recently in an appeal case,
the Queen versus William Francis Jones, where a DNA profile was actually found
on the pin of a hand grenade.
Now the statistic for this was one in a billion and at trial it was sufficient
for the individual to be convicted.
But on the appeal, the appeal, the conviction was quashed because the activity
level propositions
had not been considered properly. So what are the alternative methods of moving
DNA from a person's hand
to the pin of a hand grenade? Well it might be from secondary transfer example,
for example,
where another individual has shaken hands with the suspect and then touched the
pin of the hand grenade
or it might be because of a contamination incident in the laboratory for
example.
So all of these issues need to be properly considered when we go to court.
So what I would like to do is to introduce a framework to help us to interpret
evidence.
So let's suppose a DNA profile has been recovered from a crime scene.
We need to think about what the prosecution of the defense alternatives are to
describe how the DNA became evidential.
And typically these would follow the scheme which is shown on the slide.
So the alternative propositions would typically be the DNA came from Mr. Smith
who is the suspect
or the DNA came from an unknown person. And what we do is we calculate a
likelihood ratio
which is the probability of the evidence if the prosecution proposition is true
divided by the
probability of the evidence if the defense proposition is true. So what does a
DNA profile do and what it does not do?
Well the presence of a DNA profile tells us nothing about the how, why, or when
that it became evidence.
So it would be completely wrong to infer that the one in one billion statistic
has any relevance to the activity.
So how can the scientists help the record in this respect? Well to do this we
will now turn to the interpretation method
which employs the hierarchy of propositions framework.
And the hierarchy of propositions is layered into four different sub-levels or
levels.
At the bottom we have the sub-source level and this typically refers to the DNA
profile itself.
So we may have a low-quant stain, there's no background information and it's
impossible to make an
association between the DNA and the body fluid from where it came from. And we
call this the sub-source level and it's the statistic that we give at this
level
which is typically provided to the court. The next level is called the source
level.
Here we ask the question well where what body fluid has this DNA come from? Has
it come from blood? Has it come from semen?
And to do this analysis is necessary to have additional information. We need a
test to test the body fluid for example.
And we go up to the level, the main level of interest is called the activity
level and this asks the question did the defendant take part in the connected
activity?
And it's possible for scientists to report at this level if they have
sufficient background information.
And finally the third level is the offense level where it may be possible to
combine different types of evidence.
When we go to court DNA won't be the only evidence, it will also be fibres,
glass, etc.
And the offense level asks whether we can combine all of these different types
of evidence into a single statistic.
Now why is this important? Well if a scientist goes to court and gives a figure
of one in one billion
the danger is that statistic may be thought to mean something in relation to
the activity
of the crime for example stabbing an individual.
So it's very important that the scientist puts this statistic into his
perspective by saying that this has got nothing to do with the activity, how,
the, why, or the when.
This mistake is often known as the CSI effect.
So I use this diagram to show how DNA may become evidential.
Essentially when you have a DNA profile or when you see a DNA profile you don't
know how it's been transferred.
But what we can do is to make a list of the possible methods and this diagram
shows the possible methods of transfer.
So this is a timeline you can see at the beginning there's opportunity for
a adventitious transfer that's accidental transfer by an individual.
And then a bit later on we have the crime events where we hope there's
opportunity for transfer from the actual perpetrator.
Now after the crime event there is usually a period of time before its
discovery
and then maybe transfer in this intervening period as well.
Once the investigators arrive they set up a crime scene and they detect and
recover material.
And from that point onwards it is also possible that they may accidentally
contaminate the crime scene.
They may move DNA from one area to another for example.
And of course when it gets to the laboratory then these dangers are also
apparent and we have to take great care to make sure that our laboratories are
completely clean and contamination free.
Also we need to be aware of a psychological effect which is called confirmation
bias.
And this is the tendency to search for interpret, favor, or recall information
in a way that reinforces one's existing beliefs or hypotheses
while giving disproportionately less consideration to alternative possibilities
And I will give you some examples of this later in my talk.
So what are the different modes of transfer of DNA profiles?
Well there are effectively three different routes to transfer a DNA profile.
There is the direct method for example if I hold a knife then my DNA may be
directly transferred to the handle.
But indirect or secondary transfer can also occur that is if I shake my hands
with another individual then that individual
holds the knife it's possible that he or she may transfer my DNA to that knife
by that means.
And the third route is contamination and this is accidental transfer in the
laboratory or by crime scenes and investigators.
It's also worth bearing in mind that the mixtures are quite common in casework
and each component of the mixture may itself be transferred by a different
method.
So here's an example of secondary transfer individual A,
clasps the hands of individual B and then B holds a knife and the question is
is DNA from A
transferred to the handle of the knife.
So let's go through an example.
Let's suppose a woman is found dead in an apartment.
The pathologist confirms death by an assault during which the victim fell she
hit her head on a stone fireplace which led to a massive hemorrhage.
Suspicion falls on the ex-partner who cohabited with the victim up to four
weeks previously
and is known to have previously threatened the victim.
A mixed DNA profile is found from the body of the victim underneath her fingern
ails and that matches the ex-partner.
So we now need to work through the hierarchy of propositions.
So I'll show you how this works now.
We start with the sub-source propositions.
So this typical example we would state the DNA came from Mr. Smith and Miss
Jones.
The defense proposition would be that the DNA came from Miss Jones and an
individual who is unrelated to Mr. Smith.
And then we would report in the usual way the evidence is a billion times more
likely if the first proposition is true rather than if the second proposition
is true.
But this tells us nothing about the activity level.
So next I want to move forward to the activity level and describe this.
And of course to move forward to the activity level is important that the court
is alerted to the limitations of the sub-source DNA evidence.
And this is done by writing a caveat into the statement to the effect that it
is accepted if it is accepted by the court that the DNA came from Mr. Smith.
I am able to address the question of the activity.
So we're asking the court to accept the sub-source propositions to begin with.
So now I want to address the activity level propositions.
First of all it's important to stress that the scientist is not evaluating the
propositions only the evidence if the proposition is true.
And this is a very important distinction.
Also the scientist goes to court with circumstances of the case which are which
are be given to him or her.
There are sometimes at court these circumstances may be changed and if this
happens then you would need to re-evaluate your evidence.
So it's really important that we are really clear about the propositions and
the case circumstances before we go to court.
Again the scientist can put a caveat in his or her statement to the effect that
if this case circumstances are changed or if there is new information then they
will need to re-evaluate my evidence.
So in the case example that I'm giving you the activity level propositions
would follow for the prosecution.
Mr. Miss Jones scratched Mr. Smith on November the 21st but the defense
proposition would be that Miss Jones did not scratch Mr. Smith on November the
21st
and has had no social interaction with her since October the 21st.
He did not assault her. So you can see that these propositions are entirely
different to the sub-source propositions but they do address the court relevant
question.
The propositions are formulated as you can see to describe the points of
opinion or the points of difference between the prosecution and the defense.
Also these propositions can be formed or they should be formed independently of
the knowledge of our results.
And I hope you can see that it's entirely possible to produce these
propositions without knowing what the results are.
This is important because I mentioned confirmation bias earlier on so we should
try to develop our propositions avoiding confirmation bias and the best way to
do this is to produce them
before we know the actual results.
It's no problem having multiple sets of propositions. This is not a barrier and
this often happens if there is some uncertainty about which propositions to use
So if the first proposition is true then the DNA recovered is a result of
direct transfer because it must have been transferred at the time of the
offense.
So what the scientist must ask is what is the probability of transfer
persistence and recovery of DNA if Miss Jones scratched Mr. Smith?
The alternative, the defense alternative is what is the probability of transfer
persistence and recovery of DNA if Miss Jones had not scratched Mr. Smith?
And it follows on from that the question is how did Mr. Smith's DNA transfer to
the fingernails if he is innocent?
Well I've already told you that he was cohabiting with the victim in the
apartment four weeks beforehand.
So the question is will this DNA from Mr. Smith, will it persist in the
apartment for four weeks?
Can it be transferred to the victim by the second-year transfer and will it be
recovered upon analysis?
So that affects me as the defense question.
So how do we actually evaluate evidence like this? Well we do this by
experiments.
It's really important to do experiments and I had this slide up because I want
to stress that experiments are much much more important than experience.
Scientists are often asked to give a basis for their opinion and it may be on
experience but the experience can be very very subjective.
So I would always stress that the importance of doing experiments.
So what sort of experiments can we do to evaluate this kind of evidence?
Well they're quite easy to set up.
Obviously you can't do anything which might be considered to be violent.
Towards another individual.
But the sorts of experiments that we can do is to divide individuals into pairs
and then individual A likely scratches individual B.
And then we can take samples from individual A fingernails over periods of time
And this helps us to answer the question of how long does DNA persist for?
We could of course consider many other factors like washing the shedor status
and we need suitable controls etc.
And then we need to find out how long does DNA persist in the environments
itself?
And to answer these kinds of questions we need experiments to evaluate the the
presence of background
DNA. So for example if
if someone leaves an apartment how long will their DNA remain in that apartment
Will it be a few hours, a few days, a few weeks etc.
Other factors to consider are whether the premises have been cleaned or not.
Now I want to talk very briefly about Bayes' nets. I'm not going to go into any
detail about this at all.
But a Bayes' net can be thought of as rather like a flow chart.
And this flow chart has all of the possibilities
enshrined within the nodes.
So the DNA results can be explained because individuals X and Y are kind of
habitated for example
or because Y scratched X
or Y scratched another unknown individual etc.
So all of the possibilities are enshrined in the nodes and behind each node
there is a probabilistic calculation.
These probabilistic calculations are used to calculate a likelihood ratio at
the activity level.
I'm not going to go into any details about this during this talk.
But this is the favoured way of carrying out these kinds of calculations.
Now I want to talk about a case which I was involved with some time ago.
It's a famous case which involves the miscarriage of justice of Amanda Knox and
Raphael Isolechitel.
This case went on for many many years. The original conviction was in the year
2009.
But there was a final acquittal was in 2015. And in the interim period there
were various
appeals and retrials where the pair were found to be innocent and then reconv
icted and then
finally exonerated. So I want to go through what happened with this case.
First of all I think when cases like this arise it's really important to
analyze them to find out why the mistakes occurred. And this is because when
mistakes occur it is
rarely isolated to a single case. These kinds of mistakes typify what may be
current thinking
in a laboratory or laboratories. And so it's really important to investigate
them to prevent
other miscarriages of justice or to identify cases where similar mistakes may
have occurred.
So I'll just go through the case circumstances. In November 2007 Meredith Kurtz
ier was stabbed
in the neck. She was stabbed in her flat which she shared with Amanda Knox and
some other students.
The prime suspect called Rudy Guerdi was quickly identified. And his DNA was
discovered at numerous
locations in the crime scene. The important feature here is that he never had
any
legitimate access to the crime scene. He didn't live there, he didn't visit the
crime scene. So
there was no real reason for his DNA to be there. However he implicated Knox's
a selectuto
as collaborators and that's the reason why they became suspects. And the reason
prosecutors went along with this is because they insisted that Guerdi could not
have acted alone.
So where was the DNA evidence found? Well first thing we have a knife which was
recovered from selectuto's flat. Secondly there was a bar class found at the
crime scene which had
been forcibly removed. This was used to implicate selectuto. And there were
blood spots in the
bathroom which were mixtures of Knox and Kurtzier, the suspects and the victim.
So this is a picture of the knife. And you can see here it's quite a large
knife and the areas of typing are shown by the letters. And when the profile on
the knife angle
was analyzed it's attributed to Amanda Knox. It was a good good profile so
there's no dispute
about that whatsoever. But there was also a poor quality profile on the knife
blade which was
attributed to the victim Meredith Kurtzier and the likelihood ratio was very
low however
of the region of about 100. Right so it's important to describe how this knife
actually became
evidential in the prosecution's eyes. In fact this knife had been found in
selectuto's flat.
He was the boyfriend of Amanda Knox and they found it in the cutlery's jaw. The
knife was
observed to be extremely clean but the investigators noted that there was a
strong smell of bleach.
So they hypothesized that the knife had been used in the crime and it had been
cleaned with bleach.
Knox's DNA profile was detected on the handle of the knife and there was a very
weak profile
attributed to Kurtzier the victim. There was no evidence of blood and this is
important I think
all the tests for blood were negative. So my comment here in relation to the
hierarchy of
propositions is that the interpretation is currently at the sub-source level.
There's no direct evidence
to determine the body fluid origin and there's certainly no evidence to suggest
the activity of
transfer. And I would stress here I talked about confirmation bias earlier on
and this is an example
of that because the bleach cleaning hypothesis is an example of confirmation
bias. It is something
which was dreamt up by the prosecution because it's obviously it is extremely
inconvenient to
have this lack of evidence. So the bleach cleaning hypothesis was provided as a
way of explaining
away the lack of evidence in this particular case. There was actually quite a
good defense team
I hastened to add. The defense observed starch grains on the knife as you might
expect if you
cut a loaf of bread for example. So if we now start to think about the activity
level propositions
then two alternatives come to mind. First of all either the knife was used in
the murder
or the alternative is the knife was only used by Knox to prepare food. Now
interestingly
the this wasn't really stated to the court in this way. I'm stating this out in
an idealized
way the preferred way. It wasn't presented in this way to the actual court but
this was
obviously discussed and the judges who convicted the pair first trial made the
statement that it
appears more likely to have been derived from having held the knife to strike
rather than having
used it to cut some food. Okay so that was the judges conclusion but of course
scientifically
there is no justification for this. If you find DNA on the knife handle there's
there's no way I
can tell you from the DNA pattern whether it was deposited whether I was
stabbing somebody or
whether I was cutting a loaf of bread the profiles will appear absolutely
identical. Therefore this
was nothing but pure speculation. On the other hand as I mentioned earlier
there were multiple
stains from Greedy. He had no legitimate reasons being the apartment. His DNA
was found in multiple
locations in the murder room and in the bathroom. Even though Knox lived in the
apartment her DNA
was not found in the murder room. Well which is quite interesting. The only DNA
attributed to
Sileciuto was the minor profile found on the bra class and of course Sileciuto
had
legitimate access to the flat because he was the boyfriend of Amanda Knox.
So mixtures of DNA and Knox of Kurtia were also found in the bathroom. Now this
was interpreted by
prosecutors to indicate that Knox had washed her hands after the murder. Okay
so let's just think
about this one for a minute. You have a test for blood and you have a mixture
of DNA but this does
not mean to say that the DNA from the two individuals was deposited at the
first time. They are two
completely different tests. Add to that that we have a prior expectation that
if we go to anybody's
bathroom you're going to find their DNA in there from sweat or skin cells or
whatever. It's not
been necessarily come from blood. So we can conclude that there must have been
an existing
background or DNA from Knox and Kurtia from skin cells in the bathroom. Hence
we would expect to
find mixtures. So the presence of blood does not mean to say that both
DNAs came from blood. Probably the DNA from Kurtia came from blood but equally
the blood, sorry,
the DNA from Amanda Knox could have come from skin cells. Okay so here is the
shared bathroom
and as you can see there by the plug you can see the blood stain and you can
see the blood
stain was sampled by the investigators and this is what led to the mixed
profile. But as I said in
the last slide you've got a test for blood but it doesn't mean to see that all
the DNA that you
see there has come from blood. It could equally come from skin cells. So again
I have to draw
attention to confirmation bias. There were many stains which matched Goerdi not
just from the room
but from the victim's body. None was obtained from Knox and there was just one
on the bra class
which matched the selectitire. But again obviously this is quite inconvenient
to the prosecutors.
There is no Amanda Knox DNA in the murder room. So what they then bizarrely
stated was that the
after the murder they postulated that Knox and selectitire had selectively
cleaned the room. Now
what do I mean by that? What I mean is that they actually went into the bedroom
and then they
removed their own DNA deliberately leaving Goerdi's DNA in place. And of course
this is just a
sheer nonsense. There is no way to actually selectively clean a bedroom. You
can't see
someone's DNA. So how can you carry out selective cleaning? But nevertheless
this kind of evidence
was presented in court and it was sufficient in the first instance to lead to a
conviction.
But then well I'm just going to fast forward now. As I said there were a number
of appeals and
retrials where all of these issues were examined and re-examined again. But in
the final conclusion
the Maraskar Bruno conclusion which was the final judgment which exonerated the
pair.
The judges stated that the confiscated knife could have been moved by Knox to a
deal of
de la pagola for domestic use on the occasion of parties of other events and
thus also used by
Kerger. So these judges then accepted the issues of secondary transfer. The
knife probably wasn't
the murder weapon and that there had been no consideration of the activity
level propositions
at the previous trials. The method of collection was criticized. It was not put
into a particularly
clean box. It was put into a common cardboard box of the sought to package
Christmas gadgets.
And the judges dismissed the bleach cleaning hypothesis as illogical which is
very good news.
So then we come to the bra class but this is worth mentioning because the
evidence was used to
convict Sir Sollettito. It was found on the floor of the apartment. It had been
forcibly
removed during the assault. But crucially the bra class was not collected as
evidence until
40 days after the crime. I'm not only that but the bra class wasn't actually in
the original
position. And in addition to that there's actually a video of the investigators
passing the bra
class to each other using latex gloves and they're not changing these latex
gloves.
So here there is ample opportunity for DNA to be transferred from one item to
another.
So again the poor collection of the evidence here has really discredited
the analysis of the result of this particular item. And of course once your
crime scene has
been compromised there's nothing you can do to rescue it. And we've done quite
a lot of research
on latex gloves and they are a really excellent method to transfer a DNA from
one item to another.
So we really do need to make sure that we change our gloves every time we
handle an item.
The bra class as I mentioned it wasn't collected for 40 days so there were
numerous possibilities
for cross contamination. And just to conclude as I said earlier once the crime
scene is compromised
this is irreversible. You cannot rescue it. And this was recognized in the
final judgment
where they said the use of logic and intuition cannot in any way compensate for
the lack of
evidence or the inefficiencies of the investigations faced with missing
insufficient or contradictory
evidence the judge should simply accept it and issue a verdict of acquittal.
And then there's this
strong plea for objectivity which I think we would all agree with where the
judge is
plea for following the scientific method starting with Galileo on the
application of the scientific
method. So what are the reasons for miscarriages of justice? Maybe not only
this miscarriage of
justice but others which are maybe undetected. Well first of all it's quite
clear that prosecutors
are working backwards they're actually fitting the evidence to the crime. The
inconvenient
results like the bleach cleaning hypothesis was explained away by for bizarre
reasons.
But most notably I think there was a failure of judges to take adequate
accounts of defense
criticisms because these problems were highlighted during the actual trial but
the judges tended
to favor the prosecution propositions rather than the defense criticisms. And
finally there's a
failure of basic quality standards in collecting the evidence and that resulted
in
contamination of the crime scene. Well to summarize the case that I've just
discussed with you
well first of all the right conclusion was achieved in the end. Rudy Guerreidi
remains
the sole convicted individual for this this crime and Amanda Knox and Raphaeli
's
Sileshto were exonerated. But the problem is of course is that the pair went to
prison
for several years and they were completely innocent of the crime. So it's
important for us to understand
the principles which I've been explaining to you the hierarchy of propositions,
the consideration
of the activity level and the limitations of the DNA profiling evidence.
Because if we can explain
this to the court then miscarriages of justice like this are much less likely
to occur.
Okay to summarize my talk I've only really scratched the the surface. There is
currently a huge amount
of research going on an activity level and the development of software to help
us to carry this
out. Although it has to be said that I think that we are still at quite an
early stage
of development because not many laboratories report activity level at present
but I think
in the future this will change. First of all educational programs are needed.
Scientists,
judges and lawyers need to be educated in this field. We need to carry out
coordinated
research programs so that we can inform the models with the suitable
probabilities. What is the
probability of secondary transfer? What is the probability of direct transfer
for given situations?
There's a need to agree on the measurements that we will use. Currently we're
working
with average peak height for example but there may be other methods which can
be used. So there
needs to be a lot of collaboration and discussion within the international
community to help us with
this. There's a need to produce a knowledge base so a knowledge base is a
collection of data
that scientists can refer to to help them make decisions. We need to move
towards continuous
models. We already have this with mixture software and I see that this will go
along a parallel
route. So we will end up with sophisticated software solutions to help us to
solve these
problems. Now the engine for this is known as Bayesian networks or Bayes nets.
So we need to
generate libraries or Bayes nets that scientists can use and these need to be
programmed of course.
Currently I'm involved with a project which is known as the REACT project
organized by the European Network of Forensic Science Institute and in this
project we have
more than 30 laboratories where we are carrying out collaborative studies and
experimentation
which will help us to produce knowledge bases and also show aspects like
reproducibility etc.
And importantly we think this project will be will help
and give the laboratory's first-hand experience in generating the data that
they can use in Bayesian
networks so they can report the results to courts.
Thank you for listening to me. It's been a great pleasure to be able to talk to
you today and if
you have any questions then please email me or contact me after the session.
[silence]
Hello everyone and I hope you're enjoying the hits conference as much as I am.
My name is
Jonathan Tehback with Thermo Fisher Scientific and we're fortunate today to be
joined live
by Professor Peter Gill who is joining us from the UK and we were just
fortunate enough to hear
that wonderful talk from Peter and he's here now to answer your questions so
please feel free
to enter any questions you have in the dialogue box at the bottom of the screen
But we're going to get rolling with the first question which is that well in
adversarial
court systems like in the United States Peter do you think that the defense
should shoulder
more responsibility for questioning the weight of evidence at the sub-source
level?
Yes hello everyone thank you for listening to my talk. Yes it is really
important actually
the defense does get involved with this framework. I realize that it is
actually quite difficult
within the adversarial system to do this because there may not be sufficient
communication
with the defense but it's in under those circumstances the forensic scientist
has no option
other than to pose the defense questions him or herself but of course when you
get to court
the problem there is that the defense may not not agree with these propositions
and under those circumstances you may have to really evaluate the evidence. So
we always put a caveat
in our statement saying this is my understanding of the case circumstances this
is my understanding
of what the defense is saying this is my understanding of what the prosecution
is saying
if you are telling me something different then I will need to go away to re
evaluate
the evidence. So these kinds of caveats are really important because it tells
the court
exactly what assumptions that you are working under.
Yeah that's thank you and in light of that do you have any specific guidance or
recommendations
for reporting language or testimony language that could be used by forensic
scientists to
better clarify the limitations of a DNA inclusion at the sub-source level?
Yes I point you to a couple of documents that we prepared under the auspices of
the
International Society of Forensic Genetics where we go into quite a lot of
detail
on these issues as written by the DNA Commission. The DNA Commission has
scientists from all over the world including from the US of course
and you will find the guidance in there.
All right thank you Peter. So another question here do you think that the
scientist should be
fully aware of all the background info in a case and isn't there a chance that
this could
bias the scientist doing the DNA analysis work? Well during the investigative
phase
the scientist will be gathering information about the case. If the scientist
doesn't necessarily
need to have all of the information but certainly needs the information that is
relevant
to the aspects of the case which he or she is reporting on.
So long as we are treating this in the framework that I talked about I don't
think that there will
be any bias because you're always evaluating the evidence given a certain
proposition.
You are never saying what the propositions are. You're always evaluating the
evidence
given a proposition. This is a very nice way I think to avoid accusations of
bias.
Right so maybe a last question here with the time we have. Can you tell us just
a little bit
more about the output from the React project that you mentioned and how this
could help
DNA laboratories deal with the issues you described in your talk?
Yes of course one of the big problems that we have is lack of data, lack of
education
and actually lack of software to actually do the calculations. So the React
project is funded by
the EU Commission and it's running for two years and we have 30 laboratories
involved and we're
evaluating a particular kind of case where DNA is transferred to a tool which
is used in burglary
so it's quite a simple scenario. We're evaluating the evidence when people
directly transfer their
DNA to the handle of the tool and we are evaluating the evidence when there are
secondary
traps going on as well. So this is quite a big project. We'll produce a lot of
data which I hope
will be useful. Thank you Peter and I'm afraid we're out of time. I wish we had
a little more time
to talk with you but I want to thank you for your participation in the
conference and for your
excellent talk and for your time today. So I wish you a great evening there in
the UK.
Okay thank you Jonathan. Thank you to everyone for listening. Thanks everyone.
Goodbye.
[ Silence ]