Every time a new chemical enters the EU market around 2500 animals undergo toxicological testing. This is to provide information on how the new substance can influence human health. With the thousands of new chemicals entering our market annually, we also urgently need in silico prioritization methods based on non-animal models.
In certain scenarios, computational methods can surrogate responses of animals and the quantitative adverse outcome pathway (qAOP) concept is one of such examples. qAOP has a huge potential for regulatory applications in chemical risk assessment in the near future. These methods can predict if you develop cancer or another health outcome.
While there is an increasing number of qAOP models being proposed, their curation remains poor, and automated extraction methods are needed. In addition, the computational predictive tools lack frameworks to guide their development and failed to be user-friendly. This is important as the future users will be risk assessors without computational skills.
The main objectives of this computational master thesis are following:
- Assess the existing published qAOP models (probabilistic and mechanistic) and propose future computational data collection for qAOP (for instance cancer or neurodegeneration)
- Develop further our existing Python based user friendly software tool for modelling qAOPs (i.e.).
- Build an example of qAOPs. (i.e. effect of pesticides on neurodegeneration)
- Integrate existing transcriptomic data (dose response) to the tool.
With this thesis you will get an overview of how the qAOP concept has advanced and you will push the boundaries of toxicity risk assessment in the near future. You will be working with publicly available data and using a high-performance computer (HPC) system.
Skills required: No prior knowledge of biology or chemistry is needed. However, Python, R or another programming languages are needed to progress with the task of this thesis. Natural language processing, machine learning, graph theory, and SQL is advantageous but not necessary. Knowledge about the the Resource Description Framework (RDF) framework for representing interconnected data on the web is advantageous but not necessary. RDF is used to integrate data from multiple sources. Creativity and goal-oriented person. You will use GitHub to follow your progress and goals. Care for the environment, animals, and other humans is a must.
Working environment: You will be working partly at the Norwegian Institute of Public Health (NIPH, Folkehelseinstituttet) that contributed to risk assessment during SARS-Cov-2 pandemic and advise to the Ministry of Health. We do care about a nice working atmosphere and student's well-being and progress.
Supervisors: Main supervisor: Marcin Wojewodzic, Researcher at Norwegian Institute of Public Health (https://www.fhi.no/) and Norwegian Cancer Registry. Email for interview: maww@fhi.no Internal supervisor at IFI: Professor Torbjørn Rognes.
Literature:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7261727/
https://www.pymc.io/projects/docs/en/stable/learn.html