Summary

The main idea is to infer how the mechanisms defining the clusters could lead to a dysregulation in the pathways/biological processes that we have observed in the data. To infer the links, we employ NeuroMMSig, a catalog of networks representing mechanisms underlying neurodegenerative diseases. Due to the multiscale nature of this resource, we can map the biological processes coming from Gene Ontology (GO) and KEGG pathways (via genes) that we identified to be dysregulated in different cohorts (i.e., PPMI and ROSMAP). NeuroMMSig mechanisms are directly mapped since they were used to define the clusters. Thanks to these mappings, we can calculate the paths (shortest paths) from nodes representing NeuroMMSig mechanisms to all the genes implicated on the pathways or to the node representing the biological process. By looking at these paths, we can hypothesize possible biological routes that have led to the dysregulation observed in the data. Finally, it is important to note that since each pathway/biological process has been identified in different cohorts (i.e., PPMI and ROSMAP) and these cohorts are disease-specific, each of the paths is calculated using one of the two disease network (i.e., Alzheimer's disease (AD) and Parkinson's disease (PD)) to maintain this disease specificity.

Types of Paths

By default, the paths are calculated on a version of NeuroMMSig containing exclusively mechanistic relationships such as increases, decreases, etc. However, since not all the pathways and biological processes have a path to NeuroMMSig, we enable users to use the standard version of NeuroMMSig that all sort of relationships (both mechanistic and not mechanistic) such as association, correlations, etc. This option is available by clicking in the button placed in the green box (above the main table).