DDD Conference

Mr. Tomas Konecny

Self-Organizing Maps-Based Machine Learning Method Boosts the Study of (Epi-)Genetics of Vine Vitis Vinifera

Armenian Bioinformatics Institute, Czech Republic


Cultural crops, like vine (Vitis vinifera), have a high impact on many aspects of our civilization, such as economy, agriculture, food production, and other human activities. Grapevine is one of the most popular crops consumed in many cultures. The oldest evidence of wine production can be traced to ancient Rome and Greece, although the origin of the domesticated vine species goes more far to the East, to the Caucasian region (including Georgia and Armenia).

We present Self-Organizing Maps (SOM), a novel machine learning method, to map, categorize and analyze large and complex omics data with the focus on genetic diversity and epigenetic as well as transcriptomic regulation of gene activity to understand not only the dissemination paths of vine but also the functional relationships with an impact on the plant response to the environment.

Application of the SOM-based method to the whole genomes of about 800 vine accessions revealed details of the cultivation history of vine in space and time in the Near East, Caucasian region, and around the Mediterranean Sea. Application to the transcriptomic data enables us to study how plants escape from unfavorable external conditions using defensive mechanisms to cope with excessive environmental stimuli, like light intensity, temperature, humidity, drought, etc. Our SOM portrayal approach enables studying different molecular aspects of vine genetics (and epigenetics) with impact for ongoing studies in the context of changing climate.

Skip to content