Skip to main content
Figure 2 | BMC Immunology

Figure 2

From: Dynamics of dendritic cell maturation are identified through a novel filtering strategy applied to biological time-course microarray replicates

Figure 2

Figure of merit analysis on both significant datasets indicates optimal clustering parameters. The FOM scores over a range of cluster numbers for the 4-fold filtered data (A, 395 genes) and the 2-fold filtered data (B, 2142 genes) have been graphed. Both (A) and (B) were generated using the data from experiment 1 (experiment 2 exhibited similar results; not shown). This analysis indicates that k-means is the optimal clustering algorithm for both datasets since it has the lowest average FOM score (though this was not significant for the 4-fold data). The box on each graph indicates the optimal range for the ideal number of clusters is between 6 and 8; 8 was chosen as optimal for both data sets.

Back to article page