We note that these samples are considered technical replicates and therefore represent an ideal scenario with minimal variation.
Then, by using publicly available data that considers technical replicates we show how contrast- to-noise ratio enhancement is attained.
For arrays considered as partially exploitable based on several quality criteria additional hybridizations were done and considered as technical replicates.
In addition, three samples were run on several microarrays, and were considered as technical replicates: samples from CBA/J, C57BL/6, and BALB/c mice were run on 5, 2, and 2 microarrays, respectively.
Linear models were constructed to consider biological and/or technical replicates (when available); within-array replicate probes were averaged; and an additional between-array quantile-quantile normalization was performed to ensure the comparability between experiments.
For qPCR, considering the biological and technical replicates, the fungal biomass ratios were also arcsine square-root transformed, but were analyzed with a different linear least squares model of yijm = β0 + β1 (genotypei) + β2 (repj) + β3 (platem (repj)) + eijm (genotype i = B73, Tx303 and two derived NILs; rep j = 1, 2, 3; plate m = 1, 2, 3; eijm: random experimental error).
We considered the deviation between technical replicates and the bias between different probe types on the same sample as sources of unwanted variation and hence scored as positive any pipeline leading to a simultaneous reduction of both errors.
Since Group A/B was evaluated using the same concentrations and in the same panel of CEPH cell lines, we consider this a technical replicate of our previously studies.
The converted SAGE library could be considered as a technical replicate.
Every measure was always performed in duplicate (two PCR reactions from the same RT sample) and the average of those technical replicates was considered as the Ct for that particular sample.
At least two biological and two technical replicates were considered for calculation of results for every analysis.
