During meta-analysis, we have to do normalization data in preprocessing data steps and then we will analyze the different expressed genes. But after we do the normalization, how can we know that the data we have is good or not? I mean do we have any validation step for normalization? If yes, how do we do validation for normalization?
Thank you!
Are you using microarray data? One common way to check if your data has been normalized correctly is to compare expression distributions with boxplot. If it's normalized correctly the boxes should be even.
(image source :http://telemedicina.med.muni.cz/genomic-proteomic-analysis/index-en.php?pg=E-learning--2-data-analysis--2-2-analysis-of-genomic-data--2-2-1-DNA-microarrays-1--2-2-1-6-normalization)
If you're new, I think this slide might help. (It did for me :) : http://barc.wi.mit.edu/education/bioinfo2007/arrays/slides/arrays_lecture1-color.pdf
Mainly by diagnostic plots such as boxplots (not in case of quantile normalization), Mean-Sd-Plots (homoscedasticity), similarity maps (outliers), etc.
Best Matthias
In my experience, depending on your data type you may validate it or not. If you work searching for differential gene expression, usually quantile does the work and validation using boxplot should be enough. However if you work with data with exponential decay, only your final result may indicate if your normalization was fine and no prior validation would be possible.
Caveat: In case of quantile normalization the boxplots will always look perfect, as the idea of quantile normalization is to become normalized distributions that are identical for all the arrays.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis