User: Password:
Automated Clustering

An automated clustering algorithm (MINE) was used to identify gene clusters using 3 different Connection Specificity Index (CSI) thresholds (0.93, 0.95, 0.97); this automated clustering information can be viewed by clicking here. MINE iteratively identifies clusters of genes based on network topology and employs a small set of user-defined parameters for modularity, node degree, and clustering coefficient. During a first round of partitioning, nodes are allowed to be in multiple clusters. After this initial round, clusters were merged if the overlap fraction between them was > 0.75.

A heatmap of pairwise similarity for the sterile gene set was constructed using the CSI (See Below).

Download large version