Compute summary scores
[V11/4/07+] dChip can compute a copy number summary score using inferred copy. At the chromosome view of inferred copy number, use "Chromosome/Compute Score" and set "Options/Threshold (for copy number, single sample)" to be a value not equal to 2 (such as 1.5 or 2.5). The score is approximately the proportion of samples with inferred copy <= threshold (when the threshold is < 2) or with inferred copy >= threshold (when the threshold is > 2). Use "Chromosome/Toggle Score" (key S) to toggle between the copy curve for a single sample and the copy summary score. If inferred copy method is median smoothing, "Options/Median smoothing window" will affect the smoothness of the copy summary score. A related function is copy number summary plot. [Forum thread]
Permutation to identify loci with significant
scores
After the allele sharing scores (ASS) or LOH
scores are obtained for a group of samples, we want to assess the
genome-wide significance of the peak score regions and report peak regions that
are highly unlikely to be due to random chance. Use “Chromosome/Compute
score" ("Permute data” or "Association" in version before
11/4/07) for significance analysis and set “Options/Gene” to be 0.01 or
0.05 as significance level. To
assess the allele sharing score curve (average of allele sharing across all the
samples), do not select a factor here. To contrast the allele sharing of the
non-GVHD and GVHD group, specify a “GVHD(numeric)” column (1 for non-GVHD and
–1 for GVHD) in the sample information
file and select this column as factor in the “Chromosome/Permute data”
dialog, and the score will be average (allele sharing * GVHD(numeric)) and
larger value indicates more sharing in the non-GVHD group and small value
indicates more sharing in the GVHD group. After 500 permutations, the min and
max of the shaded box on the right indicate the significance threshold for low
and high score values.
(Page created 11/4/07)