CAG Clinical Bioinformatics (CBIO)

The analysis of patient data is an integral part of clinical decision making. As data becomes more and more abundant, the need for optimized bioinformatics data analysis algorithms increases. These algorithms are developed, implemented, and applied by bioinformaticians, scientists trained in biology, computer science, and mathematics. Applications of analysis algorithms, such as machine learning (often referred to as artificial intelligence), to biological data fall under the umbrella of bioinformatics.

The objective of CBIO is to develop and implement bioinformatics algorithms in the clinic, in order to optimize clinical decision making in terms of precision and time expenditure of practicing clinicians. CBIO brings together academic bioinformatics researchers working on innovative algorithms to extract clinically relevant knowledge from patient omics data, clinical bioinformaticians working to optimize the workflows and reporting systems for analyzing and communicating the information to clinical decision makers, and the clinicians using the data for decision making. Researchers in CBIO will develop and implement a host of clinical decision support tools, for example algorithms for predicting and preventing surgical relapse of colorectal cancer patients, algorithms for dramatically reducing the time spent on disease monitoring of leukemia patients undergoing treatment, and advanced learning frameworks for summarizing large scale genomics data for precision treatment of a number of different cancers.

Since the dawn of the medical profession, doctors have based their decisions on data collected from patients. The difference between the practices of the last century and the current one, is the amount of available data points doctors can use to base decisions on. As an example, a patient genome sequence represents 3.2 billion data points, and while highly informative, the data is not useful to decision making doctors, until distilled to the translational biomarkers encoded in the DNA. The transformation of these data to actionable knowledge is handled by bioinformaticians, who are data scientists trained with a strong domain knowledge in the field of biology.

Bioinformaticians develop tools for handling, storing, processing, analyzing,and summarizing huge amounts of biological data, translating them into clinically actionable knowledge about patients. We do this using advanced statistical and mathematical models, for example machine learning (abbreviated ML throughout, but also often referred to as AI). As such bioinformatics is a highly interdisciplinary field, which can be further subdivided into two main categories: “bottom up” bioinformatics (programming new algorithms to solve specific tasks) and “top down” bioinformatics (implementing multiple algorithms to solve general tasks).
For simplicity’s sake, we can say that academic bioinformaticians work bottom up to innovate algorithms, and clinical bioinformatics work top down to implement bioinformatics into clinical practice. As technologies for generating ever deeper profiles of patients evolve, the demand for both academic and clinical bioinformaticians continues to increase.

As evident from a 2020 report by the Novo Nordisk Foundation, the need for bioinformaticians in the health care sector is growing by approximately 30% every 5 years. While this is being partially addressed by educating bioinformaticians to work with the data generated at the hospitals, there is a gap between innovation at the universities and implementation at the hospitals This is, in part due to clinical bioinformaticians lacking a community structure, where “applied health care bioinformaticians” can interact with “academic research bioinformaticians”.

The idea for this CAG comes from a small-scale version of this type of “innovation meets production” community, carried out at Center for Genomic Medicine in the mid 2010s. Academic researchers from University of Copenhagen (including chair Lars Rønn Olsen) spent a day a week at the Center for Genomic Medicine with clinical bioinformaticians (including co-chair Frederik Otzen Bagger). This resulted in the identification of challenges and opportunities for bioinformatics in clinical decision support, as well as the development and implementation of the corresponding solutions. These innovations provided, for example, cost saving technologies by enabling assay consolidation, as well as significant reduction in the time from sampling to reply.