Review imputation for outliers methods

[ESCRI11]

https://github.com/bigomics/omicsplayground/pull/1221/files

So for playbase::detectOutlierSamples we do not need to impute. for singular value decomposition yes in this case.

Also, for corX we could use: corX=cor(t(scaledX), use="pairwise.complete.obs") which uses all complete cases with no NA.

Correlation here is at sample level (for outlier detection) for at feature level.

We could also remove results_outlier_methods entirely and handle differently. Let's focus on this later.

Thanks Xavi! I forgot to add that we had in use playbse::is.xxl(), however that was in imputed() and was for features not for samples in order to cover both. That was disabled as part of MPoC.