Accuracy of Target Spike Change Call Comparisons
Change calls reported in the MAS software are indicators of differential gene expression. Obtaining a Change call requires comparison between single probe analyses that may contain different concentrations of certain transcripts. The spiked cRNA target samples permit comparisons that will generate predictable Change calls for these probe sets for both scanner types.
Comparisons across different target spike concentrations were performed on single array analysis (.chp) files generated from scans acquired on either scanner type at each external site: 0.00 pM versus 0.75 pM; 0.00 pM versus 1.50 pM; 0.00 pM versus 3.00 pM; 0.75 pM versus 1.50 pM; 0.75 pM versus 3.00 pM; and 1.50 pM versus 3.0 pM.
Figure 5 illustrates the average percent target spike Increase call accuracy for each comparison at each site using HG-U133A probe arrays. The number of Increase calls was calculated in the MAS software across all 53 spiked transcripts. Average Increase call accuracy was calculated for all possible pair-wise comparisons of the three replicates at two concentrations, resulting in nine comparison analyses.
These results illustrate that, in general, the GeneChip Scanner 3000 and the GeneArray 2500 perform similarly across the range of comparisons, accurately calling the majority of the spiked transcripts as Increases (Figure 5). Statistical analysis showed similar Increase call accuracy results between scanner types, with a potential bias toward the GeneChip Scanner 3000 in some cases (Figure 5, *)
Spiked target concentrations in thi