Figure 4. Effect of model parameters on features of optimal bi-cluster.

(a-b) Variation of bi-cluster volume with change in row and column weights (W_r, W_c) for 2-D (a) and 3-D (b) experiments. In both cases changing W_r was found to change the bi-cluster volume considerably, while it was less sensitive to changes in W_c. The bi-cluster volume was further analyzed separately as rows and columns depicting genes and conditions. (c-f) Variation of number of genes and conditions in the optimum bi-cluster as a function of row weight (c,d) and column weight (e,f) for 2-dimensional culture (c,e) and 3-dimensional culture (d,f). This indicates the possibility of tailoring the bi-clusters by biasing the analysis towards genes or conditions by modifying the row and column weights.