- 1. Have unified input file with separate sheets for each method and only sample names, no group names needed here
- 2. Plot ISTDs
- 3. Normalize each method by ISTD or IQR
- 4. Merge methods to make a combined normalized dataset
- 5. Plot QCs
- 6. Drop QCs from normalized dataset
- 7. Save excel file with normalized data
- 8. Extract comparisons and settings from a separate excel file with a separate sheet for the settings and then separate sheets for each comparison containing
- a. Comparison – A_over_B
- b. Statistical test – t-test or ANOVA
- c. Group order – B, A
- d. Group and sample names
- i. This helps define new groups when there are overlapping groups in comparisons
- 9. Perform statistical tests for each comparison and save
- a. t-test reports with
- i. LogFoldChange
- ii. LinearFoldChange
- iii. p-value
- iv. FDR
- b. ANOVA reports with
- i. p-value
- ii. FDR
- c. Boxplots of each feature for each comparison
- d. Heatmaps for each comparison of
- i. All features
- ii. Only significant features passing FDR threshold
- e. Excel files containing underlying data used to generate heatmaps
- a. t-test reports with
- 10. Generate pptx report summarizing results, similar to python pipeline