Voting has opened in the ISHI Threesis Academic Challenge! This Threesis challenge is an opportunity for entrants to refine and hone their communication skills to articulate the impact their research has on the field of DNA forensics. View all of the entrants on this playlist on our YouTube channel. View the entrants below and vote for your favorite by clicking the “like” button on each video.
Consider the following criteria while voting:
Did the presenter introduce the topic well?
Need and relevance of the research to the field of DNA forensics
Did the presenter explain the methods and results concisely and clearly?
Did the presenter make good use of slide design/images/etc.?
Did the presenter speak at an appropriate pace and volume?
Was the presentation captivating?
UNIVERSITY OF TORONTO MISSISSAUGA | FORENSIC GENETICS
Secondary Transfer of Semen in Laundry Baskets Prior to Laundering
Secondary transfer refers to DNA which has arrived on an object via one intermediate object or person. Forensic biologists are increasingly being asked to consider how DNA was deposited on an object in addition to the classically reported source attribution match statistics. The weight of forensic DNA evidence can be represented by source-level or activity-level likelihood ratios to quantifiably compare the probabilities of two mutually exclusive hypotheses. To quantitatively evaluate activities, published research on how factors such as objects involved, source of DNA (touch or fluid), and contact time must be available for the scenario in question. This study empirically evaluates the secondary transfer of DNA from semen deposited on four types of fabric bedsheets onto six types of fabric garments through contact in a laundry basket. Experiments investigate transfer via wet semen stains, dry semen stains, and dry semen stains with damp garments. Preliminary results demonstrate no identifiable transfer of dry semen stains to dry or damp garments, but a proportion of trials (3.36%) presented identifiable and significant transfer of wet semen stains to dry garments. Completion of the study will involve DNA typing of extracted samples, from which the quality of transferred DNA can be determined. Study results will allow for the calculation of probabilities, and subsequently likelihood ratio determination, for activity-level scenarios involving the passive transfer of DNA from semen on bedsheets onto fabric garments.
UNIVERSITY FOR DEVELOPMENT STUDIES, TAMALE – GHANA | FORENSIC SCIENCE
Identification of Unknown and Nameless Dead Bodies in Ghana: Making a Case for Policy Action
Globally, fundamental human rights are said to also extend to the dead. The sheer volume of unidentified dead body cases poses a significant challenge to every nation that believes in the rudiments of human rights. It is estimated that on any given day, families, medical examiners, humanitarian and disaster management agencies and law enforcement in Ghana face issues concerning missing and unidentified bodies. In Ghana, there seems to be an ununified approach to reporting or identifying nameless bodies but sooner or later we will have to address this issue from a philosophical viewpoint as a nation. It also appears that the science involved in tackling this “silent mass disaster” seems to have been ignored and this paper serves as an appeal for action-driven policies to be implemented toward human identification in Ghana.
INSTITUT FÜR FORENSISCHE GENETIK | FORENSIC SCIENCE
Optimisation and Validation of a STR-Based Multiplex PCR for the Analysis of the Family Felidae
In my bachelor thesis, cat DNA from six different cat breeds and from Eurasian Lynx and a European wild cat was collected with buccal swabs. The DNA extraction is based on a magnetic beads method and the DNA was amplified with the MeowPlex Assay described by John M. Butler. The assay contains 11 STR markers and a sex-marker. After amplification, the DNA fragments were analysed with a capillary electrophoresis and electropherograms with full DNA profiles of cats were obtained. Right now, this system is optimised and validated, and it also works for wild cats like lnyx and not only for domestic cats.
The electropherograms were compared in order to find alleles, specific for a certain cat breed so that if this allele appears in an electropherogram, it can be linked to a cat breed with a certain probability.
INSTITUT FÜR FORENSISCHE GENETIK UND RECHTSMEDIZIN HAMBURG | FORENSIC SCIENCE
Biostatistical Analysis for Specific Parentage Assessment of Refugees from Eritrea, Ghana and Somalia
The use of population-specific allele frequencies for the biostatistical calculation of parentage is becoming more and more important in modern times, as extremely precise results can be obtained in this way.
On the occasion of this bachelor thesis, population statistics were performed to determine specified allele frequencies for the countries Eritrea, Ghana and Somalia. These three countries were chosen due to their high refugee rate, which usually requires family reunification in the country of refuge, Germany, by means of parentage testing. For this purpose, human samples from routine laboratory work were taken, analysed and consequently a genetic profile was established. From the genetic profiles of the individuals as well as from the information on nationality, ethnicity-specific data on the gene loci could be collected. The data collection was then used to calculate new allele frequencies specifically for these three populations. In order to make a decision on the quality of the newly calculated allele frequencies, the previously used Ghanaian and European frequencies for all three countries in the Forensische Genetik und Rechtsmedizin Hamburg were compared with the specified allele frequencies.
A comparison of the previously used and the newly calculated specified allele frequencies for the ethnic groups Eritrea, Ghana and Somalia showed a clear optimization of the biostatistical results in a parentage calculation when using the new specified allele frequencies. This confirms that the generation and application of ethnicity-specific allele frequencies should be preferred in everyday calculations and assessment of parentages (using Statistefix).
Trapping Sperm to End the Sexual Assault Kit Backlog
Intimate swabs collected in sexual assault kits contain mainly female epithelial cells and few sperm cells. The established method of separating these cell types is a labor-intensive and time-consuming process. This research examined the use of targeting sperm cells with sperm-specific antibodies or “traps”. By conjugating these “traps” to magnetic beads, sperm cells can be specifically targeted and isolated from mixed cell environments. This project shows the potential of this magnetic-based isolation strategy to capture sperm cells in a reduced amount of time, which could enable greater efficiency in sexual assault kit evidence processing.
Eye Colour in rs12913832:GG Individuals – Exploring Genotype-Phenotype Discordances
Eye colour can currently be predicted from DNA at the categorical level thanks to the discovery of several strongly associated single-nucleotide polymorphisms (SNPs). Many forensic DNA phenotyping models have harnessed these associated SNPs to generate statistical predictions of pigmentation traits. One important marker for eye colour prediction within these models is the locus rs12913832, located within the HERC2/OCA2 gene complex. An A allele at this position has a positive effect on melanin production, while a G allele has a negative effect. Thus, individuals with the GG genotype are expected to have blue eyes, while individuals with the AA or AG genotypes are expected to have brown eyes. However, not all individuals with the GG genotype possess blue eyes, therefore, this research aims to answer the question: What other common markers are influencing iris pigmentation among GG individuals? For this study, 1,811 GG individuals from the Canadian Partnership for Tomorrow’s Health (CanPath) dataset were analyzed. Of this cohort, 904 individuals did not self-report blue eyes. A genome-wide association study was conducted (assigning blue, green, hazel, and brown eyes to a linear regression) and significant SNPs were found in four well documented pigmentation genes. Uncovering influencing variants in cases of genotype-phenotype discordances will not only help to advance current understandings of the genetic determination of iris pigmentation in humans but expanding SNP prediction panels to include novel variants has the potential to improve prediction accuracy for forensic DNA phenotyping approaches.
Fluorescent Nuclear DNA Staining Serves as a Screening Tool for Hair Selection to be Used for DNA Profiling
Hair contains little to no nuclear DNA and typically goes through mitochondrial testing for analysis. STR testing has been unsuccessful since there have been few reliable methods to screen for possible nuclear DNA. Recent studies have provided different incubation periods for staining using a formulated DAPI solution. Based on previous research, 4 separate methods have been produced. For this experiment, 18 participants donated their hair either plucked or naturally shed. The hairs were then categorized into either anagen, catagen, or telogen. Hairs were stained using each of the 4 methods which had different incubation times: 5 minutes, overnight, 1 hour, and immediate. Staining the hairs for 5 minutes provided the clearest image and the most visible nuclei. The most promising hairs were selected for DNA profiling. Hairs categorized as anagen had the highest yield of DNA after quantification. A minimum of 500 pg of DNA was found necessary to generate a full profile, and DNA concentrations as low as approximately 230 pg can generate a partial profile. The purpose of this research is to identify an effective screening method for determining if hair recovered as evidence could be suitable for nuclear DNA profiling.
INSTITUTE OF FORENSIC MEDICINE, UNIVERSITY HOSPITAL FRANKFURT, GOETHE UNIVERSITY FRANKFURT
Forensic DNA Phenotyping – Adaption and Validation of the HIrisPlex-S System for Molecular Biological Analysis of Eye, Hair and Skin Colour
For traces that cannot initially be assigned via an STR profile, other DNA test systems are required, such as Forensic DNA Phenotyping (FDP). FDP is a collective term for alternative molecular genetic analyses, including DNA-based prediction of externally visible characteristics. Thus, DNA can serve as a “biological witness” in unexplained cases where there are no witnesses. In recent decades, several single nucleotide polymorphisms (SNP) have been identified for this purpose. For example, the published HIrisPlex-S system (Walsh et al. 2013, Chaitanya et al. 2018) allows the estimation of eye, hair and skin colour based on 41 previously recorded SNPs. This study focused on optimising and validating the well-known HIrisPlex-S model. The validation was based on conducting a subject study involving 100 individuals with different phenotypes. The optimisation of the HIrisPlex-S system presented in this study allowed the analysis of all predictive 41 SNPs. The optimisations included adjustments of primer concentrations, redesign of primer sequences and the design of an additional assay for seven markers whose representation did not always work reliably with the published methodology. Finally, 90% of eye, hair and skin colour phenotypes could be estimated with the matching phenotype
UNIVERSITY OF TORONTO | MASTER OF SCIENCE (FORENSIC GENETICS, DEPARTMENT OF ANTHROPOLOGY)
Characterization of Seven Major Canadian Population Groups Using Massive Parallel Sequencing
A critical component for forensic cases involving biological samples is the generation of a DNA profile and its corresponding statistics, which estimate the rarity of a particular DNA profile within a given population. Such statistics are generated through allele frequency data for various populations produced from extensive population studies. Currently, such large-scale studies relevant to the field of forensics involving Canadian populations are scarce. Previous literature mainly provides allele frequency data for Canadian Indigenous populations and individuals of European descent, which may be non-reproducible when investigating other population groups within Canada.
Furthermore, currently established testing methods provide only a superficial way of comparing allele data using DNA length. Using a newer method of DNA analysis known as massively parallel sequencing (MPS), higher resolution allele data can be produced, allowing for a more thorough generation of accurate DNA profiles. Furthermore, MPS can generate additional ancestry and phenotype data, which can be utilized as critical investigative information for cases such as missing persons or disaster victim identification.
The objective of this research is to generate allele frequencies, ancestry, and phenotype data using MPS for European, East Asian, Oceanic, American, African, South Asian, and Southwest Asian Canadian populations. The generation of novel population statistics from this research has a significant impact for forensic cases as it increases the ability to better estimate the rarity of a DNA profile within any major population in Canada. Additionally, data generated from this project will explore the unique genetic diversity and make-up of Canada’s populations.
Classification of Ancestry by Supervised Machine Learning Model and Forensically Relevant SNPs
Current Forensic DNA analysis relies on producing a DNA profile from short tandem repeats (STRs) on a Capillary Electrophoresis (CE). Although this method is well established and reliable, recent developments of Massively Parallel Sequencing (MPS) on Single nucleotide polymorphisms (SNPs) have the potential to overcome some limitations of CE such as missing reference profiles and database hots to provide further insight into human identification. The study aims to generate population genetic data of ancestry informative SNPs from MPS for whole blood samples of Canadian individuals self-identifying as European, East Asian, Oceanic, American, African, South Asian, and Southwest Asian. Multiple marker types will be sequenced using MiSeq Forensic Genomics System (Illumina), including 56 ancestry-informative SNPs. The proposed methods are to use multivariate modelling of machine learning algorithms, including Random Forest (RF), Principal Components Analysis (PCA) and Gradient Boosting (XGB), to provide a robust classification of samples to estimate ancestry information. Data will be transformed into one-hot encoding (0 and 1) based on SNP haplotypes. PCA will be used as a data visualization and reduction technique. RF and XGB will be used for supervised modelling. Model accuracy will be reported in terms of specificity and accuracy. Confusion matrices and Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) will be calculated to determine model accuracy. By machine learning algorithms to analyze MPS data, we will be able to predict the biogeographical ancestry of samples advancing the current methods used in human ancestry identification.
No More Jane Does: Expanding the Utility of mtDNA Profiles in Forensic Human Identification
Current forensic mitochondrial DNA (mtDNA) profiling captures the control region, a 1.1 kbp segment that contains two hypervariable regions harbouring approximately 30% of all variation present in the mitochondrial genome (mtGenome). Forensic mitochondrial profiles report single nucleotide polymorphisms (SNPs) that are detected in the control region of a sample after comparison against the revised Cambridge Reference Sequence, and samples are assigned to a haplogroup based on the SNPs captured. The use of an incomplete mtDNA profile (i.e., containing only the control region) limits the specificity with which haplogroups can be assigned, in turn limiting the power of discrimination of mtDNA profiles. Whole mtGenome sequencing, however, captures 100% of the genetic variation present in a sample and allows for the assignment of samples to highly specific subclades, reducing the number of individuals that could be the source of the mtDNA profile. This research aims to optimize and validate mtGenome sequencing of challenging sample types such as post-mortem tissue, as it is faster and cheaper to process soft tissue compared to hard tissues, which are commonly used for mtDNA profile generation. Further, the use of mtGenome sequencing expands the utility of mtDNA profiles by increasing a profile’s power of discrimination. Providing a validated whole mtGenome sequencing approach for challenging soft tissue samples will allow laboratories to conduct whole genome sequencing on a wider variety of samples and facilitate a move towards forensic mtDNA profiles that comprise the full mtGenome.