Figure 4

Per-gene approach dominates per-probeset approach. The percentages of overlapped genes detected by both approaches are shown in the overlapped areas. The numbers of differentially expressed genes identified by the per-gene approach and the per-probeset approach are: a. Data Set1, 313 vs. 6, overlap: 83.3% b. Data Set2, 402 vs. 32, overlap: 96.9% c. Data Set3, MBEI pre-processing, 1828 vs. 126, overlap: 90.5% d. Data Set3, RMA pre-processing, 1592 vs. 90, overlap: 88.9% e. Data Set3, GCRMA pre-processing, 832 vs. 107, overlap: 87.9%.