Methods for Digital Phenotyping

On 1 June 2018, bioinformatician Professor Dr. Sebastian Köhler took up his position as junior professor jointly appointed by Charité – Universitätsmedizin Berlin and the Berlin Institute of Health. His professorship for “Methods for Digital Phenotyping” is associated with the ECDF.

The 36-year-old studied bioinformatics at Freie Universität Berlin, where he also completed his doctorate, before becoming a postdoctoral researcher at Charité. His research deals with the topical issue of “digital phenotyping”, which is to do with making symptoms and clinical anomalies (phenotypic characteristics) comparable by sorting and arranging them in such a way that they are usable for a computer database or certain automated applications.

In his research he has set himself three priorities: Firstly, he is pushing forward with the development and supplementation of a so-called ontology of various diseases, that is, a comprehensive, controlled and standardized vocabulary to describe phenotypic characteristics of individual diseases or patients. “Until now, we have mainly used English-language data for this purpose. In the future, I want to concentrate more on evaluating German medical reports,” says Professor Köhler, whose work focuses primarily on so-called “rare diseases” that occur in fewer than 5 people in 10,000.

Only by means of such an ontology will this data become suitable for databases, and for this reason it is also an important prerequisite for the second research focus: Together with European and international partners, and on the basis of this ontology, Professor Köhler plans to establish an open-source database for rare diseases that is accessible to both physicians and patients.

“On average, patients suffering from rare diseases wait between 5 to 30 years to be diagnosed and typically need to see three or more doctors during this time. In at least 40 percent of all cases, the initial tentative diagnosis proves to be false,” Berlin-born Köhler explains. “As part of a previous project, and based on this ontology, we have already entered a list of about 7,000 rare diseases into an open-source database for clinical symptoms that have been attributed to one of these diseases in the past. This system enables users to conduct specific searches based on the symptoms observed. From the information entered, the computer provides a list ranking the diseases that might come into consideration.”

As the third priority area of research, this Human Ontology developed at Charité is to be synchronized with data from precision medicine (genome sequencing) in order to gain new insights, especially in the field of rare diseases.

“Genome sequencing has become affordable and fast, which means that in many cases it is now also performed for individual patients. I am working on developing software that compares the genomic data with data from the ontology, pursuing questions such as: Which genetic variants are linked to certain phenotypic diagnostic features? Is the gene variant under consideration associated with a particular type of disease? Once we can answer these questions, we will be able to find the causes of diseases, and therefore also the corresponding diagnosis, more quickly and optimize therapies,” says Köhler.

For this purpose, he also cooperates with the European Reference Networks (ERNs), an organization under the umbrella of the European Commission that is dedicated to pooling expertise and research on rare diseases. The goal is to enable patients and scientists to find the right experts more easily. “The ERNs work with the Human Phenotype Ontology that we created at Charité,” explains Professor Köhler, a passionate cyclist who navigates Berlin’s streets exclusively by bicycle for nine months of the year. “On the one hand, we benefit from the know-how of the medical specialists. On the other, doctors benefit from our data, which enables them, for instance, in just a few steps to put together a suitable group of patients for a study.” (kj)