Figure 3
From: A comparison of four clustering methods for brain expression microarray data
Flowchart summarising the method used by ISA. tG and tC are user-defined threshold parameters. They determine how great the level of expression for a gene or sample (defined in standard deviations from the weighted mean of all genes over those samples, or all samples over those genes) needs to be for selection in the cluster. Higher values lead to more, smaller clusters, lower values to fewer, larger clusters. A preliminary run at a low number of iterations, with a wide range of values for tG and tC, is used to determine a sensible range of tG and tC values for use in the main run.