I have expression analysis data of 7 genes of two different genotypes and of 24 different tissues, can any one tell me about
1. Independent component analysis (ICA) & Principal component analysis
2. Should I use ICA or PCA for my data?
3. How to draw Venn diagrams for my data?
thank u Yasin. My data is of real time PCR. Is is possible to draw Venn diagrams for these coz' what papers I found, people have used Venn diagram for Microarray data, therefore in comparison mine is a small set of data
I suggest toi use Cardoso's JADE algorithme to do ICA on yr data
You Can find Matlab and R version on his website
The choice between both methods should be strongly related to the kind of your data. If your data can be physically modelized by a linear mixture of statistically independent "source" signals then ICA should provide an estimation of these physical sources (PCA won't). If it is not the case you should not use ICA and prefere PCA. PCA will give you a new representation of your data with some statiscal meaning.
Beg to differ - PCA just calculates the vectors describing the maximum dispersion of the samples in the space of the variables, using the variance-covariance matrix.
The statistical meaning of this is moderate.
If you want to interpret the contributions of your variables to the dispersion of your samples, ICA seems preferable to me.
ICA can be used not just for signals in the classical sense but also for sets of variables that are influenced by underlying phenomena.
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology