Check Normality - hisl6802/ClusteringToolbox GitHub Wiki

Checking data normality

A common practice when performing metabolomics analysis is to log-transform (base-10) and auto-scale also known as z-score or standard scaling. This practice is important when applying t-tests, and ANOVA to understand dysregulation of specific metabolites. Thus, to ensure that users can also verify the normality of the data prior to performing clustering analysis the following functionality is provided.

Transformations Options

• Log-10 transformation
• Square-root transformation
• Cube-root transformation

Data Scaling Options

• Mean Centering
• Auto Scaling
• Pareto Scaling
• Range Scaling

We acknowledge that additional transformation and scaling options are available and will continue to add.

NOTE: Transformations and scaling will impact clustering results. Therefore, ensure you document and consider multiple transformations and data scaling types.