Figure 1

Research of the co-associated genes in clusters. (i) The original dataset is cleaned by removing all the genes with MVs. (ii) A hierarchical clustering is performed for a given clustering method and the clusters are selected to characterized the original co-association of the genes, (iii) MVs are generated with a τ rate per gene, and replaced by zero or by an estimated value computed with the kNN approach. (iv) Hierarchical clusterings are done for the pondered Euclidean distance, the kNN and zero replacement methods. clusters are selected to characterize the new co-association of the genes. From this last step, the Conserved Pairs Proportion (CPP) is computed (see Methods).