MetaProViz::DMA - Add checks for variance assumptions

[ChristinaSchmidt1]

This is to check if the choosen parameters are good to be used in regards of the data.

• [x] Check variance assumptions for models. (equal variances, heteroscedasticity)

[dprymidis]

So here is what I found regarding the tests and assumptions. This is like a general consensus on what to use depending on the data but there are many opinions out there. For not normal data we will use the Kruskal.test as we do. For normal data (after Shapirto) We can perform Bartlett. test to assess the variance homogeneity of the groups. If the variances are equal then we use standard Anova. If they are not the we use Welch anova.

I added the bartlett test (which is to test the variance of the groups for normal data) as an "external" function (like Shapiro) in the DMA. As a result it return a message wi ththe percentage of metabolites that have equal group variances based on the test. I also made a density plot with the distribution of the pvalues but I dont think its nessesary so its not printed but its in the Plot list in the output. Now its just to add the welchanova and its posthoc test and it should be fairly quick

• [x] Variance homogeneity test
• [x] Welch's Anova
• [x] Game-Howell post hoc test

[dprymidis]

added vst tranformation and param VST =TRUE/FALSE