Figure 1

Outline of feature selection and cluster analysis process. A flow chart of the analysis process is shown in (A) and is composed of 3 steps. Step 1 is the significant differential expression filtering where each gene was required to have at least one time point where the signal's change p-value and the fold change met the indicated criteria. Step 2 is the consistency filtering of the results from Step 1 where genes that exhibit consistency of expression level and profile pattern between experiments are identified. Step 3 is the cluster analysis of all data from Step 2 and includes a FOM analysis followed by consensus clustering. (B) A representation of all datasets generated from Step 1 in (A) and their overlaps. The gray circle represents all 45,101 genes contained on the Affymetrix GeneChip; the green (7982) and blue (2469) circles on the left are the number of probe sets that met the 2-fold and 4-fold filtering criteria respectively for experiment 1; the yellow (8076) and red (2642) circles on the right represent the number of probe sets that met the 2-fold and 4-fold criteria respectively for experiment 2. (C) An expanded view of the dotted box in (B) showing the results of Step 2 in (A): the purple (1000) and brown (4439) circles represent the common gene lists for the 4-fold and 2-fold criteria between both experiments (step 2a). The inner yellow (395) and green (2142) circles represent the two final sets of genes that were determined to have the most significant and consistent differential expression for the 4-fold and 2-fold filtered data, respectively.