R Methy VII: ChAMP

In this part we will use another dataset:

library(ChAMP)
testDir=system.file("extdata",package="ChAMPdata")

Champ Analysis

See the manual http://www.bioconductor.org/packages/release/bioc/vignettes/ChAMP/inst/doc/ChAMP.html for ChAMP.

Perform a full pipeline analysis of the data with ChAMP. The latest version of ChAMP is able to use several cores. If your server or computer has more cores, you may specify more threads at the same time to make function faster, but which may cost more memory. To detect how many cores your computer has, type detectCores().

To do the full analysis will take several hours, so do the analysis step by step:

• load the data using champ.load()
• Perform QC with champ.QC()
• normalize the data using champ.norm( , cores= ), use several cores if your computer allows it.
• perform a SVD analysis champ.SVD(beta=myNorm,pd=myLoad$pd)
• if needed, perform a batch effect correction using champ.runCombat()
  • this one takes very long time som you might want to skip it for now and do it some other time.
• perform a single site analysis using champ.MVP()
• Perform a regional analysis champ.DMR(cores = )

Questions

11. How many probes are filtered based on detection p-value? How many probes remain after filtering?
12. Hand in QC-plots: MDS-plot, density plot and sample dendogram?
13. Which normalization method was used?
14. Hand in SVD-plot, is a batch correction needed for this data?
15. How many significant MVPs (with an adjusted p-value below 0.05) did you find?
16. How many significant DMRs did you find?

---

Next: ChAMP extra

Previous: Analysis