Improving the scaling normalization for high-density oligonucleotide GeneChip expression microarrays

BMC Bioinformatics

Table 1 Summary of signal data in 76 rat genome U34A GeneChip microarrays.

	Lowest	Highest	Mean	SD	CV (%)
Total signal	832,561.4	3,161,392.7	2,039,655.7	526,295.0	25.80%
Sum of signals used for SF	524,513.7	1,986,236.9	1,212,296.5	336,138.0	27.73%
Trimmed total	308,047.7	1,240,257.3	827,359.1	215,325.1	26.03%
Mean signal	94.6	359.3	231.0	59.8	25.80%
Median of signals	17.8	54.8	35.7	8.7	24.41%
Mean of log signals	4.3	5.8	5.1	0.4	7.17%
Trimmed percentage	34.4	54.1	40.7	4.4	10.70%

"Total signal" is the sum of all the signals on each array. "Sum of signals used for SF" is the sum of signals excluding the trimmed data and used to calculate SF. "Trimmed total" is the sum of the 2% probe sets with the highest signals on the array. "Mean of log signals" is the mean of log₂ transformed signals. "Trimmed percentage" = (Trimmed total/Total signal) × 100%. See also in Methods. The "lowest" and "highest" showed the lowest and highest number in the category among the 76 chips, respectively. The mean, standard deviation (SD) and coefficient of variation (CV) were also calculated.

ISSN: 1471-2105