Feature/dea functions

[DanielGarbozo]

feat: add TCGA-LUAD DEA example datasets and @examples to existing functions

New datasets (TCGA-LUAD, 16 tumor vs 16 normal samples):

• deseq2_results: DESeq2 results table (21,330 genes, tumor vs normal)
• norm_counts: DESeq2 size-factor normalized counts matrix (21,330 x 32)
• vst_counts: Variance-stabilized matrix for dimensionality reduction (21,330 x 32)
• data-raw/deseq2_results.R: reproducible generation script

New function:

• nice_VSB_DESeq2: BSV plot directly from a DESeq2 object (normalized counts extracted internally)

Updated functions (added @examples):

• nice_Volcano, nice_VSB, nice_PCA, nice_UMAP, nice_tSNE
• detect_filter, add_annotations, save_results, split_cases
• tpm, get_stars, power_analysis, get_annotations

Other changes:

• Updated @description of nice_VSB to clarify that input should be
  data frame like normalized counts (not VST), since log2 transformation is applied internally
• Updated DESCRIPTION: added LazyDataCompression xz, DESeq2 to Suggests,
  removed knitr/rmarkdown (vignette built by CI only)
• Updated NAMESPACE and man/ accordingly