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 the data has been normalized by a control sample. For example if we have a set of gene expression ratios tumor versus normal.
For each gene analyzed, the method uses Bootstrap techniques to approximate the distribution of the ratio, and therefore, the statistical significance of the Ratio.
Data must be prepared in a two-column format. First column refers to gene names whereas second column refers to ratio 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.
- When defining gene names blank characters are not allowed: "Cox2" is right but "Cox 2" is wrong. SEE EXAMPLE.
- If observed median ratio and bootstrap 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.