Signed-off-by: thierry <thierry.balliau@inrae.fr>

replacmeent off aov by lm in mcq.compute.anova for cmcq.generic.pepq object and cmcq.generic.pepq.labeling object

modification of the code of selection of factorToColor in mcq.plot.intensity.profile that caused a bug if the factor is not in the metadata

removing the customize sentence for new graph. There no way to take in account all possible interrpution shortcut

add test to avoid computation of contrast if only one factor with 2 modality in tyhe experimental design there is no sens to do that

clean capture of error if the model cannot be estimated for all proteins (test for xic on prot, peptide, spectral conut and normalized spectral count on data test)

Signed-off-by: thierry <thierry.balliau@inrae.fr>