Quick access: MAIN || NO CONTROL || WITH CONTROL || EXAMPLES
This statistical analysis is the intended to assess the statistical significance of Ratios in those experiments where each gene of interest, say G, has one (or more) experimental conditions to be compared to a control condition.
For each gene analyzed, the method uses a permutationís test to approximate the distribution of the resulting ratio obtained from experimental condition expression values versus control condition expression values, and therefore, the statistical significance of each Ratio.
Data must be prepared in a three-column format. First column refers to gene names, second column refers to experimental condition (including control one) and the third column refers to expression values.
IMPORTANT INSTRUCTIONS FOR A SUCCESSFUL ANALYSIS:
- Each gene should have more than 1 data point (Null data is not admitted). We make note that Ratio distribution approximation improves when the number of data points is N>10.
- Data should be prepared in a tab separated file without headings. SEE EXAMPLE.
- Column 2 is the type of condition column. Control must be indicated for each gene as CTRL (all capital letters). When defining gene names blank characters are not allowed: "Cox2" is right but "Cox 2" is wrong. SEE EXAMPLE.
- If observed ratio of the mean values of expression and BootstRatio median ratio are far apart one from each other, the analysis should be USED WITH CAUTION since it could indicate a non valid analysis. This situation may arise in a small sample size situation which may entail large variability.