12.11.2012 Views

Examination of Firearms Review: 2007 to 2010 - Interpol

Examination of Firearms Review: 2007 to 2010 - Interpol

Examination of Firearms Review: 2007 to 2010 - Interpol

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

quantifiable measures <strong>of</strong> reliability and accuracy, quantifiable measures <strong>of</strong><br />

uncertainty in the conclusions, and the lack <strong>of</strong> research on observer bias and sources<br />

<strong>of</strong> human error. DNA analysis however has proven <strong>to</strong> give consistent results, with a<br />

high degree <strong>of</strong> certainty, sufficient <strong>to</strong> demonstrate the connection between evidence<br />

and a specific individual or source. In this respect the s<strong>to</strong>chastic effects observed<br />

during PCR and especially Low Copy Number typing (LCN) can complicate the<br />

procedure. To make no mistakes, or have a zero error rate is practically impossible.<br />

Therefore it should be a good scientific practice <strong>to</strong> show the limits <strong>of</strong> a specific<br />

method and <strong>to</strong> address for example the impact <strong>of</strong> potential bias. Thompson 41 gives<br />

an example <strong>of</strong> a blind procedure, preventing observer bias, in forensic DNA pr<strong>of</strong>ile<br />

interpretations. He first argues that “the widespread failure <strong>to</strong> adopt blind procedures<br />

for making DNA interpretations sets the stage for target shifting”. In which target<br />

shifting means that examiners change their interpretation on the basis <strong>of</strong> the<br />

available reference pr<strong>of</strong>iles from suspects which they know are likely <strong>to</strong> match.<br />

Thompson 41 proposes <strong>to</strong> prevent this by sequential unmasking, in which the features<br />

<strong>of</strong> the contested material are observed first. After this the examiner has <strong>to</strong> point out<br />

which DNA pr<strong>of</strong>iles could contribute <strong>to</strong> this material. Only after that he can continue<br />

his investigation, comparing the reference samples with the contested material.<br />

As the DNA technology is advancing, the statistical part becomes more and more<br />

important. Hedman et al 42 proposed a statistical model for unbiased quality control <strong>of</strong><br />

forensic DNA pr<strong>of</strong>iles (e.g. capillary electrophoresis electropherograms). The<br />

measure proposed is the forensic DNA pr<strong>of</strong>ile index (FI). Where the <strong>to</strong>tal peak height<br />

<strong>of</strong> the capillary electrophoresis electropherograms (e.g. the sum <strong>of</strong> the heights <strong>of</strong> the<br />

observed STR peaks given in relative fluorescent units (rfu)) <strong>to</strong>gether with the mean<br />

local balance (e.g. the mean <strong>of</strong> intra-locus balances or discrepancies between peak<br />

heights within a heterozygous STR marker), and the Shannon entropy (e.g.<br />

discrepancies, or inter-locus balance, between the sum <strong>of</strong> the peak heights between<br />

the markers) 42 are combined. The FI can be used <strong>to</strong> ensure the overall quality <strong>of</strong><br />

DNA pr<strong>of</strong>iling within a labora<strong>to</strong>ry.<br />

5.2 Low Copy Number<br />


<br />

In order <strong>to</strong> analyze samples that contain less than 200 pg <strong>of</strong> template DNA 28,43 (in<br />

literature referred <strong>to</strong> as low template DNA (LTDNA) or <strong>to</strong>uch DNA) LCN can be used.<br />

There are several variants <strong>of</strong> the LCN technique, the most basic (the variant with 34<br />

PCR cycles, a replication strategy with two or more replications, and a validated<br />

interpretation methodology) is used in the Netherlands, the UK and in New<br />

Zealand 44 . LCN uses STRs <strong>to</strong> produce a DNA pr<strong>of</strong>ile. In comparison <strong>to</strong> the<br />

conventional STR methods applied in normal casework LCN uses methodologies <strong>to</strong><br />

increase DNA detection sensitivity. For example modifications <strong>to</strong> the PCR (e.g. other<br />

primers) and/or post-PCR manipulations 34,28,43,45 .<br />

The increase in detection sensitivity gives a concomitant increased risk <strong>of</strong><br />

contamination. For example more sensitivity <strong>of</strong> detecting DNA, enhances the chance<br />

<strong>to</strong> find not only the DNA pr<strong>of</strong>ile <strong>of</strong> the perpetra<strong>to</strong>r, but also the DNA pr<strong>of</strong>ile <strong>of</strong> other<br />

persons 46 (whose cells were present on e.g. the victim due <strong>to</strong> casual contact 28 ).<br />

Furthermore s<strong>to</strong>chastic effects are also more likely <strong>to</strong> occur. S<strong>to</strong>chastic effects are<br />

heterozygote peak imbalance, allele or locus drop-out, and stutter (for more<br />


 337


Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!