Sequence trimming was conducted using LUCY – a program developed at TIGR [ 33] with a trimming standard of an overall base call error rate of <1%, free of vector- and E. coli sequences, and a trimmed sequence read length of > 100 bp.
Then, all data in that column is standardized by subtracting the trimmed mean and being divided by the trimmed standard deviation.
Probes exhibiting median signal intensities less than the trimmed mean (trim = 5%) plus three trimmed standard deviations of the 3xSLv1 probe were flagged as absent (within the background signal).
This trimmed standard deviation was used to remove outliers from the calculations.
Spots with median signal intensities less than the trimmed mean (trim = 5%) plus three trimmed standard deviations of the 3xSLv1 probe were flagged.
Spots with median signal intensities less than the trimmed mean plus three trimmed standard deviations of these negative controls were flagged as absent.
This involved shuffling the segmented regions in each sample, recalculating trimmed standard deviations, and calculating an empirical false discovery rate (FDR) [ 41].
These significantly variable CNVRs were determined by summarizing each window using a trimmed standard deviation (removing the minimum and maximum), followed by a permutation testing procedure to calculate the threshold used to identify potential CNVRs: any window with a trimmed standard deviation above this threshold was considered to lie within a CNVR.
Background fluorescence was measured using the 3xSLv1 probes; probes with median signal intensities less than the trimmed mean (trim = 5%) plus three trimmed standard deviations of the 3xSLv1 probes were flagged and called absent.
The background fluorescence was measured using the negative control 3xSLv1 probes; probes with median signal intensities less than the trimmed mean (trim = 5%) plus three trimmed standard deviations of the 3xSLv1 probe were flagged as absent (within the background signal).
Background fluorescence was measured using the 3xSLv1 negative control probes; probes with median signal intensities less than the trimmed mean (trim = 5%) plus three trimmed standard deviations of the 3xSLv1 probes were flagged as absent (within the background signal).
