Figure 4 illustrates average Detection call accuracy for the 53 transcripts at each spiked concentration. Target spike concentrations correspond to sample targets containing the 53 transcripts presumed to be Absent (0.00 pM) based on historical results from GeneChip HG-U133 expression assay analysis of this HeLa cell line, and then reconstituted to 0.75 pM, 1.50 pM, or 3.00 pM in the final Figure 3. Comparison of Average Percent False Change Between GeneChip Scanner 3000 and GeneArray 2500 at External Sites. False Change is described by the frequency of Increase calls greater than or equal to a Signal Log Ratio of 1 and Decrease calls less than or equal to a Signal Log Ratio of -1 for a comparison analysis between identical replicate samples hybridized to two probe arrays of the same design. Each histogram shown reflects the average percent False Change for 12 target samples hybridized to three lots of HG-U133A probe arrays scanned on a particular scanner. Error bars represent one standard deviation. See text and Figure 2 for sample description. 1.8% 1.6% 1.4% 1.2% 1% 0.8% 0.6% 0.4% 0.2% 0% Site 1 Site 2 GeneArray GeneChip Scanner 3000 y 2500 Site 3 Site 4 Site % False Change (Average) False Change Comparisons Between Scanner Types complex cRNA target. The histogram value is the average frequency for which a Present Call is made for the 53 transcripts at 0.75 pM, 1.50 pM, or 3.00 pM; or the average frequency for which an Absent call is made for the 0.00 pM target. For each scanner type at each site, the percent accuracy of Detection is an average of three hybridized target samples. In general, the analysis shows that accurate transcript detection is similar for both scanner types at most sites.
The 3.00 pM target spike was called Present with nearly 100% fidelity by all scanners at all sites. Detection accuracy for the