Two human genome probe array designs (HG-U95Av2 and HG-U133A) were used with three lots of probe arrays per design. Samples were run once on each array. Therefore, each external site generated 48 scans: 2 probe array designs x 3 lots of each probe array design x 1 sample for each lot x 4 spiked target samples x 2 scanners (Table 1).
Each probe array was scanned first on the GeneChip Scanner 3000, then on the GeneArray 2500. This scanning strategy permitted scanning a single probe array per pair of instruments, eliminating variation introduced by different probe arrays, and ensuring that any differences in performance were attributable to the scanner that generated the image.
Photobleaching that may occur due to scanning in the GeneChip Scanner 3000 prior to scanning in the GeneArray 2500 is negligible, and has been shown to have no impact on image quality or the resultant data (data not shown). Furthermore, the scanning order between the GeneChip Scanner 3000 and GeneArray 2500 was reversed in limited tests and showed no impact on results (data not shown).
Alpha Test Plan
An initial alpha test plan was conducted at Affymetrix and compared the performance of three pairs of scanners (one GeneChip Scanner 3000 and one GeneArray 2500). The alpha test plan had the same experimental design, probe arrays, supplies, and reagents used for the beta test plan. However, samples were run in triplicate on each array lot, therefore generating 144 internal site scans: 2 probe array designs x 3 lots of each probe array design x 3 replicate samples for each lot x 4 spi