AD BEL Model subgraph statistics
Statistics for each AD subgraph are shown in table 1 and 2. There are in total over 100 different subgraphs in AD. Most of the hallmark pathways incorporated in Alzpathway are covered and represented as subgraphs in our mechanism repository NeuroMMSig (Ogishima et al., 2016). For example, "Amyloidogenic subgraph" was one of the biggest network and contains knowledge about amyloid plaque formation, likewise, "Non-amyloidogenic subgraph" contains information about how amyloid beta is degraded. These two subgraphs respectively correspond to the pathways "Amyloid degradation" and "Amyloid cleavage" pictured in figure 4. However, the size of Alzpathway’s networks is smaller in comparison with their related NeuroMMSig subgraphs. "ApoE subgraph" or "Beta secretase subgraph" subgraphs are also representing well characterized players in AD. It is important to mention that the AD model also focuses in the role of neuroinflammation in the disease, as increasing evidences suggest it (Heneka, et al., 2015). Hence, that many subgraphs are based on literature reporting knowledge about inflammation or immunological related processes. For example, NeuroMMSig’s networks: "Inflammatory response subgraph", "Tumor necrosis factor subgraph" or "Prostaglandin subgraph". Other subgraphs capture all the knowledge around proteins involved in familiar AD like "Gamma secretase subgraph". Gamma secretase, as mentioned in section 1.1.1, contains proteins (PSEN1/PSEN2) with pathogenic mutations linked to early-onset AD. Apoptotic processes are also described in NeuroMMSig (e.g., "XIAP subgraph", "Bcl-2 subgraph").