The browser you are using is not supported by this website. All versions of Internet Explorer are no longer supported, either by us or Microsoft (read more here:

Please use a modern browser to fully experience our website, such as the newest versions of Edge, Chrome, Firefox or Safari etc.

Improved and tailored prediction methods for cancer

Illustration of cellular networks in different forms of cancer
Harmful variants in cellular networks divided by different forms of cancer

Lund researcher Mauno Vihinen has received a grant from the Cancer Foundation to develop AI methods that in the long run enable improved cancer diagnosis.

One of the researchers who has received a grant from the Cancer Foundation is Mauno Vihinen, professor of medical structural biology at the Department of Experimental Medical Science and leader of the research group Protein Structure and Bioinformatics. He has received a grant of SEK 600,000 per year for two years for the project "Prediction of effects of cancer-related genetic variants".
– The grant will make it possible to develop prediction methods tailored to cancer. This is, of course, an important area for effectively defining sequences of genes and genomes for cancer patients. Interpretation of the variants is the bottleneck, but this is precisely our specialty.

Understanding disease mechanisms

With the grant, the research program can be realized, says Mauno Vihinen, and benefit cancer patients. With the help of the grant, Mauno Vihinen will recruit a bioinformatics officer who will be commissioned to develop methods and analyze genetic variants from patients to sift out which variants are related to the cancer disease.Mauno Vihinen hasresearched variants in principle throughout his research career.
– In my dissertation I used protein engineering to modify properties of proteins. It isinteresting to understand the mechanisms of disease. You have to use computer methods because it is impossible to analyze them all in the lab. A patient may have more than one million variants in hisoch hers DNA.

Development of AI methods

The research field is multidisciplinary, both focused and broad, and combines knowledge from different angles; bioinformatics, genetics, biochemistry, medicine and immunology. The research is about variations, also called mutations.
– We use artificial intelligence, machine learning and other computer techniques to develop methods for screening variants that are disease-related. There are many variations in most cancers, but only a few of them are related to the disease. We train computer programs to find harmful variants. Furthermore, we research in which way disease-related variants are harmful. For this we look at their effects on DNA, RNA and protein levels, functional effects and how protein structures are affected.
The project that the grant will be used for is about developing AI methods for analysis.-We will develop AI methods to analyze which variants related to the cancer are harmful. This in turn enables improved diagnostics. More specifically, we want to develop new methods for identifying pathogenicity, that is, potential capacity to cause disease, in protein function-related variants.The research group's previous prediction methods are already used today in clinics and research.
–This will also be the case with the new methods that we develop in this project. When you know which variants are harmful and in which cellular processes, you can personalize treatments for patients, which is to the benefit of the patient.

LInk to original article published by Agata Garpenlind 19 January 2021