Visualising personal genomics - machine learning tools in action
09 June 2019 | talks
Prof. Wim Van Criekinge shares his thoughts on visualisation tools for personal genomics.
The primary purpose of exploring personal genomics is to gain knowledge and understanding from the wealth of information encoded in an individual’s genome. Visualisation is a critical aspect for exploration and interpretation of this personal genomic information. However, due to the quantity and richness of the data – a complete human genome contains around 2x3.2 billion letters or nucleotides – specialised tools and techniques are needed.
There is a need in the healthcare sector to develop tools such as a genome browser or mobile device application specifically designed for physicians to easily interpret, deliver and communicate patient’s genomic information and its significance through effective visualisation. Applications such as Genewall-tensor provide browsable access to patient’s raw genomic data in parallel with the genomic sequence allowing straightforward visualisation of disease-associated variant locations.
Access to personal genomics data and tools for its interpretation should become accessible to everyone.
Importantly, one main aspect of such browser is the ability to integrate multiple forms of genomic information in the same field of view: next to a person’s genome sequence, different data types providing additional information (i.e. transcription factor binding, proteins, etc.) can be rendered in parallel tracks to visually assess possible functional implications of a specific genomic variant. Note that this information is not limited to one patient (sequence), but that it is possible to upload multiple genomic profiles. This property facilitates comparative visual assessment between for example personal genomes of family members, patient’s with similar symptoms/diseases, or even between the somatic DNA sequence of a cancer patient compared and the DNA sequence his tumour genome.
Adapted from ICML 2019 Expo Talk by Wim Van Criekinge. Written by Kathryn Morrissey for BioLizard.