3 data normalization (2014 lab tutorial)
Download as PPTX, PDF3 likes1,027 views
The document discusses normalization methods for reducing analytical variance and addressing batch effects in replicated measurements. It covers topics such as sample normalization, variable transformation, and visualizations of batch effects. The analysis highlights that while normalization can reduce analytical variance and batch effects, it is crucial to determine the impact of different normalization techniques on the data.
3 data normalization (2014 lab tutorial)
- 1. Biology Chemistry Informatics Normalization methods for analytical variance reduction Statistics Goals: Evaluate batch effects in replicated measurements and overview normalization methods Topics: 1. Batch effects 2. Sample normalization 3. Variable transformation 4. Variable normalization
- 2. Biology Chemistry Informatics Identify the effects of sample drying Statistics Use DATA: Normalization Data.csv Visualize: 1. Batch effects in replicated measurements 2. The effect of normalization on samples and variables Questions: 1. How can batch and outlier effects be mitigated?
- 3. Biology Chemistry Informatics Question: Statistics Are there any batch effects in this data?
- 4. Biology Chemistry Informatics Answer: Statistics Are there any batch effects in this data? (Yes!)
- 5. Biology Chemistry Informatics Answer: Statistics Not all metabolites are similarly affected
- 6. Biology Chemistry Informatics Question: Statistics Can sample and variable normalizations reduce analytical variance and batch effects?
- 7. Biology Chemistry Informatics Answer: Statistics Sample and variable normalizations can reduce or increase analytical variance and batch effects. Sum raw normalization
- 8. Biology Chemistry Informatics Answer: Statistics Sample and variable normalizations can reduce analytical variance and batch effects. Sum normalization raw
- 9. Biology Chemistry Informatics QC sample based normalizations Statistics Can be useful for estimating and removing analytical variance
- 10. Biology Chemistry Informatics Example: qcLOESS workflow Statistics
- 11. Biology Chemistry Informatics Overview of Normalizations (PCA) Raw (%RSD = 13) qcISTD (9) LOESS (12) qcISTD + LOESS (8) Only LOESS included normalizations effectively remove analytical batch effects