3.3.11.1 Calculating correlations

Assuming that such paired profiles have been loaded within VAMP (see section 4.4 for an example), the analysis can be launched from the menu Tools $\rightarrow$ Correlation Analysis $\rightarrow$ Compute:

Figure 3.67: - ``Compute'' Dialog for Correlation Analysis.

Several parameters can be chosen to run a Correlation Analysis

scope:

should correlation coefficients be calculated on all data or only on selected regions (if any) ?

criteria:

should correlation coefficients be calculated from copy number ratios, or from smoothed copy number values given by GLAD ?

correlation:

which type of correlation coefficient should be used ?The Pearson correlation coefficient measures the extent to which the association between copy number and expression is linear; its sensitivity to outliers makes it suitable to detect associations within regions that are amplified in only few samples. The Spearman correlation coefficient is the Pearson coefficient between measurements ranks; it is therefore robust to outliers and able to detect non-linear associations