Fig. 5: Using the platform for prioritizing genomic variants of a rare disease patient.

a Overview of the different steps of the procedure, from collecting relevant data in the web module to performing interactive network analyses in VR and finally post-processing the results. b–f Screenshots illustrating the individual prioritization steps. b The required data (patient phenotypes and candidate genes, seed genes, network data) are collected and curated using the web interface. All data are immediately available also in the VR environment thanks to the shared data module infrastructure. c Individual candidate genes (red) and seed genes are inspected in their network context to identify related biological functions and known disease associations. For example, the blue links show a connection between the candidate gene DOCK2 and cellular metabolic processes in the functional landscape corresponding to GO biological processes. d The broader network neighborhood around the seed genes is identified using a random walk with restart algorithm, providing a provisional and purely network-based ranking of the candidate genes. Left: inspection panel showing a bar plot of the resulting visiting frequencies of the 13 candidate genes. Right: the seed genes (yellow) are significantly connected, indicating a particular network neighborhood associated with the respective disease. e Detailed inspection of the most highly ranked candidate genes according to known functions and disease associations within their network neighborhood. Left: Inspection panel showing expression levels of DOCK2 across different tissues. Right: Connections of DOCK2 in the network view. f, Saving the key results of the VR exploration for further inspection and follow-up analyses in the web module. Left: Screenshot of the web module summary of the VR session. Right: Following up on a particular interaction partner of DOCK2, RAC1, in PubMed.