The cummerbund package cannot properly handle condition (given by the 'sample_1' and 'sample_2' columns) names starting with numerics.

Some methods handles it standard R style by adding an 'X' before the condition/sample name. Examples include:

cuffDB <- readCufflinks(...)

diffData(genes(cuffDB))

Other methods fails and end up with NA's ('<NA>' to be precise, the factor levels of these are however correct). I found this for methods for CuffGeneSet classes (result of getGenes() function). Example:

myCuffGeneSet <- getGenes(cuffDB, featureNames(genes(cuffDB))

cuffSplicing <- diffData(splicing(myCuffGeneSet))

The files outputtet by CuffDIff does not have this problem.

Hope you can fix this.

Kindest regards

Kristoffer