Biology Reference
In-Depth Information
The three logical principles for the interpretation of DNA evidence and its quan-
tification that have been suggested [8] are as follows:
(1) when evaluating DNA evidence two assumptions should be considered;
(2) the probability of occurrence of the evidence under each of the two assump-
tions should be quantified; and
(3) the ratio of the probabilities under two assumptions should be quantified and
considered.
By considering the two propositions where one is the alternative of the other (the
prosecution proposition compared with the defence proposition), then the probabil-
ity of the evidence if the prosecution case is true can be determined. If the DNA
profile of a crime scene sample matches a suspect's DNA profile then there can be
two explanations.
(1)
Prosecution hypothesis (
H
p
)
: the DNA profile originated from the suspect.
(2)
Defence hypothesis (
H
d
)
: the DNA profile did not originate from the defen-
dant but originated from another person.
The likelihood ratio is described in Equation 9.2. It is the probability (Pr) of the
DNA evidence (
E
) given the hypothesis put forward by either the prosecution (
H
p
)
or the defence (
H
d
):
Pr
(
E
|
H
p
)
(9.2)
Pr
(
E
|
H
d
)
The prosecution hypothesis is that the defendant left the biological material at
the crime scene. The DNA profiles of the defendant and the evidence from the
crime scene match, and therefore it is certain under the prosecution's hypothesis that
the defendant left the material, hence Pr(
E
|
H
p
)
=
1. The probability of occurrence
under the defence hypothesis is equal to the probability of observing the profile if its
source was somebody other than the suspect in the population to which the defendant
belongs. The ratio of the two probabilities is given by:
1
profile frequency
Likelihood Ratio
=
(9.3)
Let us assume that a certain crime scene stain yielded a DNA profile that has
matched that of a suspect who is being prosecuted. The frequency of the matching
DNA profile was 1 in a million. The likelihood ratio in this case would be:
1
0
.
000001
=
Likelihood Ratio
=
1 000 000
