menu

Jessica Xin Hjaltelin

This project aims to improve early detection of pancreatic cancer (PC) by exploring its link with diabetes, especially new-onset diabetes (NOD). PC is one of the most difficult cancers to diagnose early, resulting in one of the lowest 5-year relative survival rates at only 12%. Using artificial intelligence (AI) and data from Danish health registries, the project will develop models to predict PC by analysing patient histories, including hospital diagnoses, laboratory measurements, GP and hospital medications, and primary care visits. Early detection of PC is crucial since it significantly increases survival rates. By focusing on patients with NOD, who are at a higher risk of developing PC, the project seeks to identify subtle signs of the disease early on. This will involve creating AI models that can accurately pinpoint individuals at high risk. The project aims to make significant strides in early cancer detection through timely diagnosis and intervention.

Sarah Rennie

Beyond serving as simple intermediary molecules in the transfer of information from DNA to proteins, RNAs must be tightly regulated, and their correct processing is essential for maintaining cellular function. mRNA fate is strongly influenced by a diverse group of covalent modifications and interactions with RNA binding proteins, which regulate all stages of its life cycle. Recent technological advances are generating large quantities of data in this field, yet there is a need for new computational methods to fully utilise these datasets. To this end, we will develop powerful new frameworks to 1) investigate how RNA modifications are dynamically regulated at the single-cell level and 2) identify the molecular determinants underlying modification patterns. In addition, further research is required to elucidate emerging links between RNA modifications and neurological disorders. Our newly developed approaches will be applied in the context of Parkinson’s disease, aiming to pinpoint the regulatory roles and timings of RNA modification in dopaminergic neuron development.

Michael Westbury

Preserving biodiversity is not only an urgent matter for the natural world, but also for long-term human health. As the world undergoes dramatic human-induced change, we need to identify species most at risk of extinction and formulate mitigation tactics. Current risk identification approaches are very time and resource intensive, making them unfeasible for many species. The genome of an organism provides valuable information about its past, present, and future and could be leveraged to rapidly identify whether a species is at risk. Despite this, we currently lack the tools to do this for a wide range of species. Our research program will bring the world of genomics to species without access to high quality samples or funding by developing new approaches to reliably assemble and analyse their genomes. Furthermore, we will take advantage of recent advances in machine learning to provide a rapid and efficient method to predict extinction risk in living species based on their genomes.

Jonas Lybker Juul

The COVID-19 pandemic has left a devastating impact on societies across the globe, with immediate and indirect costs in terms of human lives, life quality, and economics. Any improvement in how society deals with future pandemics promises a considerable impact economically and health-wise. This proposal will strengthen epidemic preparedness in three ways:

By developing new statistical methods to leverage large-scale genomic sequencing to understand how diseases spread through populations.

By studying how disease mitigation strategies can be both effective and efficient.

By improving key algorithms used when producing forecasts of case numbers in epidemics.

By improving society’s ability to forecast case numbers, infer how diseases spread, and mitigate outbreaks, InForM aims to save human lives and economic costs in future pandemics.

Jonas Meisner

Genomics-driven precision medicine has the potential to revolutionize global health by predicting disease risk and uncovering underlying causes of complex diseases. To ensure that predictions are accessible, accurate and effective for people from diverse backgrounds, we need to develop innovative computational methods for genomics. The goal is to create tools that can analyze large amounts of genetic data to identify new patterns and connections to help us predict and prevent diseases. By focusing on the most prevalent diseases today, such as cardiometabolic diseases, this project intends to make a meaningful difference in public health. This work combines state-of-the-art techniques to make sense of complex genetic data using haplotype information. This involves developing methods to accurately identify the ancestral backgrounds of individuals, identify genetic risk factors for diseases, and predict disease outcomes based on combined large-scale genetic and health data.

Patrick Munk

Bacteria have waged chemical warfare for billions of years. Their metaphorical armory of weapons and shields are antibiotics and antibiotic resistance respectively. As humans have appropriated antibiotics, resistance proliferates globally and now kills 1.3 million people annually.

To slow this silent pandemic we need to both find new weapons and understand which bacteria are shielded from what weapon. Growing public microbiome datasets can help us do that, but there a number of issues. They contain a large layer of metagenomic dark matter we don’t understand and our current analyses are done sample-by-sample and don’t exploit how different datasets synergistically can inform each other.

There are a number of obstacles to solve before we can recover both the useful and harmful microbial armory hidden in global microbiomes. With this project, I seek to develop new computational approaches, make computer programs and harvest the hidden armory for humanity to exploit and slow the AMR pandemic.

Picture credits: Lars Svankjær / The Young Academy of Denmark

Stefan Sommer

The project aims to develop data science and AI methodology needed for analysis of the shape and form or animals, organs, and plants – their morphology. Morphology has traditionally been important features in studies in fields ranging from biology over health science to plant science, but it is hard to quantify morphological differences, to do well-defined statistical analysis, and to integrate morphological data in data science and AI analysis pipelines. The project aims at solving this by developing the necessary statistical, machine learning and mathematical foundation for such analysis together with software implementations that are directly useful for researchers in the above fields. By doing this, we will make large classes of high-resolution data available for research to make new discoveries, to improve our understanding of animal evolution, to develop more resilient crops, and to increase our understanding of the relation between human organ shape and evolving diseases.

Jakob Skou Pedersen

Solid tumors release circulating tumor DNA (ctDNA) to the blood, where it can be recovered from the plasma. Detection and analysis of ctDNA may transform cancer care. However, in many clinically relevant settings, it comprises only a minute fraction of the circulating free DNA (cfDNA), with most coming from healthy cells. The cfDNA can be cataloged in vast data sets using DNA sequencing techniques. We will use generative AI statistical modeling techniques to detect subtle ctDNA signals and characterize the underlying cancer biology. Predictive methods will be trained and evaluated on comprehensive public cancer genomics and local cfDNA data sets. The goal is to contribute cancer biology insights and help advance cancer care with methods for early diagnosis, disease surveillance, and cancer characterization.

Aasa Feragen

The upcoming AI act will likely lead to widespread use of trustworthy AI methods such as explainability and algorithmic fairness. Such methods are crucial to ensure that AI is implemented responsibly and safely into increasingly critical societal functions, such as medical imaging. Nevertheless, the reliability of explanations and algorithmic fairness models is rarely addressed in state-of-the-art responsible AI research. In this project, we will showcase how trustworthy AI algorithms can fail, and develop theoretical and practical links between uncertainty in AI models, and failure modes of their trustworthy counterparts.

This has several advantages: First, it gives us potential tools to assess whether trustworthy AI algorithms are likely to fail, so that we can safely use them when they don’t. Second, these methods come with a straightforward generalization to the modern generative AI models, for whom trustworthy AI tools are currently largely unavailable.

Veronika Cheplygina

Machine learning (ML) competitions aim to improve healthcare algorithms, like detecting lung cancer in chest x-rays. Teams worldwide compete to develop such algorithms, driven by prizes or prestige. Competitions encourage innovation but often produce similar algorithms that excel in one accuracy metric but fail with diverse real-world data.

Relying on one accuracy metric is insufficient for measuring algorithm quality, for example for rare patient cases. It leads to many similar algorithms with high environmental cost. Moreover, competition can deter women and minorities from entering or staying in data science.

I propose competitions which look at multiple metrics across different patient subgroups, with applications in chest and cardiovascular disease. Using the latest ML techniques we will develop methods to improve evaluation of algorithm robustness while reducing the carbon footprint. We will also study how competitions affect women and other groups in data science.