Fig. 3

Application of algorithm to Insilco data and comparison with other algorithms. a A schematic showing different Insilco datasets on which algorithms are applied. Different datasets had different levels of noise and overlap degree. b Comparison with other algorithms when applied on Insilco datasets as shown above. Both rows contain different graphs when algorithms are applied on different datasets with increasing noise levels (column wise graphs). In first row, in each graph recovery of biclusters score is plotted as a function of overlap degree of biclusters in datasets on which algorithm is applied. Lower row contains graphs where relevance of biclusters score is plotted. CG algorithm performs better in each scenario. In the recovery graphs, qubic and bimax algorithm overlap completely with CG algorithm and can’t be seen here. Similarly, in the relevance graphs, qubic algorithm overlap completely with CG algorithm (upto noise = 0.10) and hence can’t be seen here