Transfer, persistence and recovery of forensic DNA traces and the application of Bayes Nets to interpret evidence that is relevant at the activity level (ReAct)
There are many studies on activity level, but either the data are not available, or the data are difficult to access and use in calculations. This is because there is no standardisation of data or the analytical method. ReAct is a project that includes cooperation between 33 forensic laboratories from 20 countries and is funded by the European Network of Forensic Science Institutes (ENFSI) monopoly program funding 2020. The purpose of the project is to execute an experimental design that provides data to estimate probabilities of direct and secondary transfer to be incorporated into Bayesian Networks to calculate likelihood ratios at activity level. The data and programs will be made publicly available.
Novel methods for forensic air and dust sampling
The Forensic Genetics research group at Oslo University Hospital is at the forefront of research on air and dust sampling for forensic applications. In a recent study (Fantinato et al. 2023) we developed novel methods to collect and isolate human DNA from the air. We showed as a proof of concept that human DNA can be collected from the air in sufficient amounts to yield STR genotypes of the recent occupants of a room. It was further demonstrated that dust samples may be useful to detect occupants of a room over longer time periods. This project will utilize experiments performed in households to advance research on bioareosol for forensic applications. The results will provide new knowledge and tools for the use of human bioaerosol to aid investigation of criminal cases.
Microbiome analysis for forensic applications
This project focuses on investigating how one can use microbiomics as a method to determine the original location of a sample, and if this is usable in the forensic toolkit. More specifically can microbial composition in air and soil be used to tie objects, people and other samples to a recent location? Additionally the project will investigate how one can optimise methods for this inference and make them robust enough to have value from a forensic standpoint.
Forensic body fluid identification:
The overall aim of the project is to develop a test for detection and quantification of mRNA body fluid-specific markers in blood, saliva and vaginal mucosa – three relevant cell types in forensic application. The commercial RNA quantification systems today, measures, typically, the total RNA in a sample. They are neither human nor body fluid/marker specific. In mRNA body fluid analysis, the cDNA input amount for RNA profiling on CE analysis is crucial, as too low or too high cDNA may fail the analysis. The test may also act as a presumptive test for blood, saliva and vaginal mucosa.
Methods to detect highly degraded materials and to improve discriminating power
The aim is to create custom MPS panels with a great number of SNP markers (collaboration with the supplier) with applications both in relationship inference, ancestry inference and mixture evaluations.
Kinship analysis with additional STRs
In some instances, kinship analysis would benefit from inclusion of additional STR markers to increase statistical power in our calculations. We are therefore testing the quality of commercial kits that include additional STRs. Our aim is to implement this to real case work and to share our findings with the forensic community.
Disaster victim identification
Disaster victim identification (DVI) is a branch of forensic kinship analysis in which victims of disasters are matched against relatives of missing persons by means of DNA. This project addresses computational challenges of DVI that are a major limitation of current methods. We have recently published promising results using joint calculations in DVI cases. Our ongoing work aims to further develop these methods, and implement them in practical software tools.
UiO:Life Science: Medical, legal and lay understandings of physical evidence in rape cases (Evidently Rape)
Medical, legal and lay understandings of physical evidence in rape cases (EVIDENTLY RAPE) studies how physical evidence matters and can be a factor in how medical and criminal justice institutions approach the crime of rape. EVIDENTLY RAPE is a multidisciplinary research project that explores how physical evidence is harvested, selected, tested, communicated and applied throughout the criminal justice process in cases of rape. The communication of medical knowledge and its margin of error to the police, lawyers and legal and lay judges creates numerous challenges. Medical, legal and lay understandings of physical evidence in rape cases (Evidently Rape) - UiO:Life Science
New methods to estimate the time since deposition of forensic traces
A method to accurately determine the time since deposition (TSD) of a crime-stain would be of great importance in criminal investigations. This method would make it possible to separate crime relevant stains from those not related to the incident in question and link time-dependent evidence directly to the crime event. In this project we are investigating time dependent changes in the observed ratio between different mRNA markers and how this can be used to estimate the time since deposition of a blood stain. We have shown how degradation a selection of markers show correlation with time and how this data could be used with an elastic net approach to model changes (Hänggi et al. 2023). Further, we are working on implementing new markers for more robust analysis and analysis of longer intervals of time.
MPSproto: A PGS for utilizing MPS-STR sequence information.
We have developed MPSproto as an extension of EuroForMix to improve handling of stutter artefacts and other typing errors that commonly occur in MPS-STR data. MPSproto implements two models for read depth: gamma and negative binomial.
EFMex: Using EuroForMix for exhaustive calculations.
We have developed EFMex, a shinyApp, a wrapper of EuroForMix, which makes it easy for the user to perform LR calculations based on exhaustive evaluations of hypotheses. This approach is very promising when multiple related contributors are typed and needed to be evaluated.
Extension of kinship model in EuroForMix
RelMix is a software which implements the possibility to specify a pedigree for one or several of the unknown contributors in a mixture and perform mixture interpretation based on qualitative model (Hernandis et al 2019). We aim to integrate this possibility into a EuroForMix extension such that a quantitative model can be used.