Classification of firing pin impressions using HOG-SVM.

Crimes, such as robbery and murder, often involve firearms. In order to assist with the investigation into the crime, firearm examiners are asked to determine whether cartridge cases found at a crime scene had been fired from a suspect's firearm. This examination is based on a comparison of the marks left on the surfaces of cartridge cases. Firing pin impressions can be one of the most commonly used of these marks. In this study, a total of nine Ruger model 10/22 semiautomatic rifles were used. Fifty cartridges were fired from each rifle. The cartridge cases were collected, and each firing pin impression was then cast and photographed using a comparison microscope. In this paper, we will describe how one may use a computer vision algorithm, the Histogram of Orientated Gradient (HOG), and a machine learning method, Support Vector Machines (SVMs), to classify images of firing pin impressions. Our method achieved a reasonably high accuracy at 93%. This can be used to associate a firearm with a cartridge case recovered from a scene. We also compared our method with other feature extraction algorithms. The comparison results showed that the HOG-SVM method had the highest performance in this classification task.

A comprehensive study into false positive rates for 'other' biological samples using common presumptive testing methods.

The identification of biological fluids or materials in forensic samples is a key requirement in forensic science that relies on chemical and biological based tests, most of which exhibit false positivity. When reporting results from such tests, Forensic Scientists use words such as probable, possible, and likely, without always being able to provide robust support for these conclusions. In collating information about false positive rates for a number of these tests, we found limited research into the cross reactions observed from 'other' biological samples in commonly encountered case sample stains. By 'other' we mean biological fluids or materials that are not the primary target of the presumptive test being used. Here we carry out a specificity study to fill gaps in the literature for a number of the presumptive chemical, biological and immunochromatographic tests used to presumptively screen for blood, semen and saliva. The tests selected for this study are the widely used tests: Luminol, TMB/Combur3 Test® E, Kastle-Meyer (KM), RSID™ - Blood, ABAcard® HemaTrace®, Acid Phosphatase (AP), ABAcard® p30, RSID™ - Semen, Phadebas® 'Tube' Test, Phadebas® 'Press' Test, and RSID™ - Saliva tests. Specificity for each of these was tested in known samples, from volunteers, of blood, semen, saliva, urine, sweat, vaginal material, faeces and breast milk, and then false positive rates were determined.

On the Identification of Body Fluids and Tissues: A Crucial Link in the Investigation and Solution of Crime.

Body fluid and body tissue identification are important in forensic science as they can provide key evidence in a criminal investigation and may assist the court in reaching conclusions. Establishing a link between identifying the fluid or tissue and the DNA profile adds further weight to this evidence. Many forensic laboratories retain techniques for the identification of biological fluids that have been widely used for some time. More recently, many different biomarkers and technologies have been proposed for identification of body fluids and tissues of forensic relevance some of which are now used in forensic casework. Here, we summarize the role of body fluid/ tissue identification in the evaluation of forensic evidence, describe how such evidence is detected at the crime scene and in the laboratory, elaborate different technologies available to do this, and reflect real life experiences. We explain how, by including this information, crucial links can be made to aid in the investigation and solution of crime.

Compatibility of the ForenSeq™ DNA Signature Prep Kit with laser microdissected cells: An exploration of issues that arise with samples containing low cell numbers.

Massively parallel sequencing is rapidly emerging as a valuable tool in forensic DNA analyses. As part of our validation of this technology, we established its compatibility with a laser microdissection cell collection method including a one-tube DNA extraction process. We also used the laser microdissector to explore the number of cells required to generate informative DNA sequence profiles and establish the limitations of the technology. Using the ForenSeq™ DNA Signature Prep Kit (Primer Mix B) and a MiSeq FGx™ sequencer, we successfully demonstrated the compatibility of MPS with a laser microdissection one-tube extraction method, with minor alterations made to the manufacturer's recommended library preparation protocol, including the addition of magnesium chloride to counteract the effect of dithiothreitol on amplification efficacy. This work highlighted several quality issues that may be encountered when preparing sequencing libraries from low quantity DNA samples, particularly that libraries prepared from low cell numbers showed high levels of adapter dimer compared to those prepared from more cells. To remediate this, we replaced the bead normalisation step with a qPCR normalisation method, whereby sequencing libraries are diluted based on their molarity as determined after library purification. The work presented here focuses on the results from the autosomal and Y STR markers as these could be directly compared to results obtained from traditional capillary electrophoresis techniques. Full autosomal STR DNA sequence profiles (27 loci) could be obtained from 50 epithelial cells and 100 spermatozoa (sperm cells). The limit of detection for the ForenSeq™ system was determined to be 25 epithelial and 25 sperm cells for both autosomal and Y STRs. Cells were dissected from both single source samples and mixtures of semen and saliva. There was no apparent difference in sensitivity, presence of contamination or PCR artefacts between libraries prepared from single source samples and libraries prepared from mixed source samples.

Forensic STR allele extraction using a machine learning paradigm.

We present a machine learning approach to short tandem repeat (STR) sequence detection and extraction from massively parallel sequencing data called Fragsifier. Using this approach, STRs are detected on each read by first locating the longest repeat stretches followed by locus prediction using k-mers in a machine learning sequence model. This is followed by reference flanking sequence alignment to determine precise STR boundaries. We show that Fragsifier produces genotypes that are concordant with profiles obtained using capillary electrophoresis (CE), and also compared the results with that of STRait Razor and the ForenSeq UAS. The data pre-processing and training of the sequence classifier is readily scripted, allowing the analyst to experiment with different thresholds, datasets and loci of interest, and different machine learning models.

Evaluation of massively parallel sequencing for forensic DNA methylation profiling.

Epigenetics is an emerging area of interest in forensic science. DNA methylation, a type of epigenetic modification, can be applied to chronological age estimation, identical twin differentiation and body fluid identification. However, there is not yet an agreed, established methodology for targeted detection and analysis of DNA methylation markers in forensic research. Recently, a massively parallel sequencing-based approach has been suggested. The use of massively parallel sequencing is well established in clinical epigenetics and is emerging as a new technology in the forensic field. This review investigates the potential benefits, limitations, and considerations of this technique for the analysis of DNA methylation in a forensic context. The importance of a robust protocol, regardless of the methodology used, which minimizes potential sources of bias is highlighted.

A review of bioinformatic methods for forensic DNA analyses.

Short tandem repeats, single nucleotide polymorphisms, and whole mitochondrial analyses are three classes of markers which will play an important role in the future of forensic DNA typing. The arrival of massively parallel sequencing platforms in forensic science reveals new information such as insights into the complexity and variability of the markers that were previously unseen, along with amounts of data too immense for analyses by manual means. Along with the sequencing chemistries employed, bioinformatic methods are required to process and interpret this new and extensive data. As more is learnt about the use of these new technologies for forensic applications, development and standardization of efficient, favourable tools for each stage of data processing is being carried out, and faster, more accurate methods that improve on the original approaches have been developed. As forensic laboratories search for the optimal pipeline of tools, sequencer manufacturers have incorporated pipelines into sequencer software to make analyses convenient. This review explores the current state of bioinformatic methods and tools used for the analyses of forensic markers sequenced on the massively parallel sequencing (MPS) platforms currently most widely used.

The development of a method of suspension RNA-FISH for forensically relevant epithelial cells using LNA probes.

Messenger RNA profiling is becoming a common method for body fluid identification in forensic science but there are disadvantages when cell mixtures are present from more than one individual. A method that could identify and separate such cell mixtures would simplify downstream analysis. To do this, we have developed a novel method of RNA suspension-fluorescent in situ hybridization (RNA S-FISH) using a locked nucleic acid (LNA) probe for the keratin 10 (KRT10) mRNA that is suitable as a potential marker for epithelial cells. As sample size may be restricted in forensic samples, this method has focused on minimizing cell loss whilst maintaining signal strength. Furthermore, we have shown that it is possible to obtain full DNA profiles from 150 RNA S-FISH labeled cells isolated using laser microdissection.

A time-course analysis of mRNA expression during injury healing in human dermal injuries.

The determination of dermal injury age is important in forensic practice. It helps answer questions that are important to an investigation such as the timing of the injury and incident, the order of infliction (where there is more than one injury), the survival time after injury (post-infliction interval) and the relation of the injury to the incident. Despite the importance of injury age determination, there currently exists no reliable method to estimate dermal injury age. In this study, the expression of the following 14 mRNAs was studied in human dermal injuries and their usefulness in the estimation of human dermal injury age was evaluated: dual specificity phosphate 1 (DUSP1), interleukin 7 (IL7), vascular cell adhesion molecule 1, tenascin C, cluster of differentiation 14, interleukin 6, tumour necrosis factor alpha (TNFα), interleukin 1beta (IL1ß), chymase 1 (CMA1), collagen type III alpha I, interleukin 2, collagen type I alpha I, collagen type I alpha II and vascular endothelial growth factor A (VEGFA). DUSP1, IL7, TNFα and VEGFA showed an initial decrease in expression during the early stages followed by an increase in expression towards the middle and late phases. IL1ß and CMA1 expression was limited to specific time points. The remaining markers either showed inconsistent expression or were undetected in our samples. The expression patterns of the detected markers suggest they have potential to predict injury age, especially during the initial stages of injury healing, if used in combination with one another.

SNP model development for the prediction of eye colour in New Zealand.

The ability to predict externally visible characteristics (EVCs) from DNA has appeal for use in forensic science, particularly where a forensic database match is not made and an eye witness account is unavailable. This technology has yet to be implemented in casework in New Zealand. The broad cultural diversity and likely population stratification within New Zealand dictates that any EVC predictions made using anonymous DNA must perform accurately in the absence of knowledge of the donor's ancestral background. Here we construct classification tree models with SNPs of known association with eye colour phenotypes in three categories, blue vs. non-blue, brown vs. non-brown and intermediate vs. non-intermediate. A set of nineteen SNPs from ten different known or suspected pigmentation genes were selected from the literature. A training dataset of 101 unrelated individuals from the New Zealand population and representing different ancestral backgrounds were used. We constructed four alternate models capable of predicting eye colour from the DNA genotypes of SNPs located within the HERC2, OCA2, TYR and SLC24A4 genes using probability calculation and classification trees. The final model selected for eye colour prediction exhibited high levels of accuracy for both blue (89%) and brown eye colour (94%). Models were further assessed with a test set of 25 'blind' samples where phenotype was unknown, with blue and brown eye colour predicted correctly where model thresholds were met. Classification trees offer an aesthetically simple and comprehendible model to predict blue and brown eye colour.

Detection of target-probe oligonucleotide hybridization using synthetic nanopore resistive pulse sensing.

Despite the plethora of DNA sensor platforms available, a portable, sensitive, selective and economic sensor able to rival current fluorescence-based techniques would find use in many applications. In this research, probe oligonucleotide-grafted particles are used to detect target DNA in solution through a resistive pulse nanopore detection technique. Using carbodiimide chemistry, functionalized probe DNA strands are attached to carboxylated dextran-based magnetic particles. Subsequent incubation with complementary target DNA yields a change in surface properties as the two DNA strands hybridize. Particle-by-particle analysis with resistive pulse sensing is performed to detect these changes. A variable pressure method allows identification of changes in the surface charge of particles. As proof-of-principle, we demonstrate that target hybridization is selectively detected at micromolar concentrations (nanomoles of target) using resistive pulse sensing, confirmed by fluorescence and phase analysis light scattering as complementary techniques. The advantages, feasibility and limitations of using resistive pulse sensing for sample analysis are discussed.

Characterising stutter in forensic STR multiplexes.

Stutter is an artefact seen when amplifying short tandem repeats and typically occurs at one repeat unit shorter in length than the parent allele. In forensic analysis, stutter complicates the analysis of DNA profiles from multiple contributors, known as mixed profiles, a common profile type. Consequently it is important to both understand and predict stutter behaviour in order to improve our understanding of the resolution and interpretation of these profiles. Whilst stutter is well recognised and documented, little information is available that identifies and quantifies what influences the formation of stutter. In this work we use a novel approach to examine this. We have used synthetic oligonucleotides comprising multiple repeat units to test; the influence of repeat number, the influence of repeat sequence and the impact of interruptions to the repeat sequence length. Using multiple replicates allows detailed statistical analysis. We have confirmed a linear relationship between stutter ratio and repeat number. We have shown that increased A-T content increases stutter ratio and that interruptions in repeating sequences decreased stutter ratios to levels similar to the longest uninterrupted repeat stretch. We also found that there was no relationship between stutter ratio and repeat number for a repeat unit with an A-T content of 1/4 and that half of the interrupted repeat sequences stuttered significantly less than their longest uninterrupted repeat stretches. We have applied the knowledge gained to examine specific features of the loci present in the AmpFlSTR(®) SGM Plus(®) multiplex kit used in our laboratory.

A comparison of stochastic variation in mixed and unmixed casework and synthetic samples.

Understanding the behaviour of mixed DNA profiles is of paramount importance in forensic DNA analysis. Key parameters are those of heterozygote balance and mixture proportion and its variability. These parameters have been previously explored as a function of the average peak height of the active alleles in single source and mixed samples derived from pristine DNA. Here we report a comparison of this data with data obtained from casework samples. This allows an assessment of the difference in the distribution of heterozygote balance between mixed and single source stains and between casework mixtures and synthetic mixtures constructed from pristine DNA.

Development of an electrochemical polypyrrole-based DNA sensor and subsequent studies on the effects of probe and target length on performance.

DNA sensors have a wide scope of applications in the present and emerging medical and scientific fields, such as medical diagnostics and forensic investigations. However, much research-to-date on DNA sensor development has focused on short target DNA strands as model genes. In this communication we study the effect of the length of oligonucleotide probe and target strands as a significant step towards real world applications for DNA detection. The sensor technology described uses the conducting polymer polypyrrole as both a sensing element and transducer of sensing events - namely the hybridization of complementary target oligonucleotide to probe oligonucleotide. Detection is performed using electrical impedance spectroscopy. Initially sensor development is performed, wherein we demonstrate an improvement in stability and sensitivity as well as show a reduction in non-specific DNA binding for fabricated sensors, through use of a specific dopant and post-growth treatment. Subsequently, we show that longer target DNA strands display increased response, as do sensors containing longer probe DNA strands. It is suggested that these results are a feature of the increase in negative charges associated with the longer DNA strands. The results of this comparative study are aimed to guide future design of analogous sensors.

The effect of cleaning agents on the ability to obtain DNA profiles using the Identifiler™ and PowerPlex® Y multiplex kits.

A year after the introduction of Identifiler™ into the forensic DNA laboratories of the Institute of Environmental Science and Research Limited (ESR), increasing occurrences of dropout of the three loci, D7S820, D18S51, and FGA, were observed in samples where the DNA was not degraded and sufficient DNA was present that full DNA profiles were to be expected. The dropout was either partial or complete at these loci. Full profiles could sometimes be obtained by reamplification of samples using the same input amount of DNA. After a thorough investigation of the methods and procedures used in the laboratory, the cause of this inhibition was identified as the cleaning agent TriGene™ ADVANCE. This was determined after the deliberate addition of varying amounts of different cleaning reagents into the DNA amplification reactions. At concentrations of 0.004% TriGene™ ADVANCE caused inhibition resulting in tri-loci dropout. At concentrations of 0.04% and higher, complete inhibition was observed. An effect was also seen on the amplification of samples using the Y STR profiling system PowerPlex(®) Y. This work highlights the importance of checking all reagents and chemicals prior to use, even those with no apparent direct influence on the DNA profiling process.

A method for DNA and RNA co-extraction for use on forensic samples using the Promega DNA IQ™ system.

The use of messenger RNA profiling to identify the origin of biological samples (e.g. blood, semen and saliva) from crime scenes is now at the stage of being implemented into routine forensic casework. We report on the successful modification of the Promega DNA IQ™ system to enable co-extraction of DNA and RNA from the same sample without compromising the potential DNA profile. Using the protocol in our laboratory for extracting DNA using the DNA IQ™ system combined with the Zymo Research Mini RNA Isolation Kit™ II we demonstrate the simultaneous co-extraction of DNA and RNA from the same sample for routine DNA and mRNA profiling for the identification of both the individual and the biological stain.

The development of a mRNA multiplex RT-PCR assay for the definitive identification of body fluids.

With current methodology, DNA profiling can identify an individual from a sample of biological material but it does not reveal what body fluid or tissue source the DNA profile originated from. We have developed a multiplex PCR system using messenger RNA (mRNA) that can identify blood, saliva, semen and menstrual blood in individual stains or in mixtures of body fluids. Messenger RNA transcripts specific to each type of body fluid have been identified and a multiplex reverse transcriptase-polymerase chain reaction (RT-PCR) system developed to identify these body fluids along with three housekeeping genes. This multiplex can detect semen and seminal fluid (semen without spermatozoa present). Furthermore, we have targeted the co-isolation of RNA and DNA from the same sample and, with the RT-PCR multiplex, we can determine the type of body fluid present as well as generate a DNA profile(s) from the same stain.

The use of bacteria for the identification of vaginal secretions.

We have used the 16S-23S rRNA intergenic spacer region for identifying vaginal specific bacteria. Lactobacillus crispatus and Lactobacillus gasseri were detected in vaginal secretions but not in semen, blood or saliva. Our data indicated that both L. crispatus and L. gasseri were detected in vaginal secretions from women with different levels of expression of hormonal genes including pregnant, pre- and post-menopausal women, and a woman who has had a hysterectomy. Therefore, we have demonstrated that these Lactobacilli are promising new markers for the forensic identification of vaginal secretions. We have incorporated the Lactobacilli markers into a mRNA multiplex system to produce an 11-plex assay that can identify circulatory blood, menstrual blood, saliva, semen (in the presence and absence of spermatozoa) and vaginal secretions.