NADIA: Missing Value-Aware Differential Abundance Analysis of DIA Proteomics Data
First version submitted to Bioconductor. NADIA covers the differential abundance analysis of protein quantification data from the raw report to the figure, treating the missing values that dominate data-independent acquisition as the central problem rather than a nuisance to be filtered away. Reports from Spectronaut, DIA-NN and Proteome Discoverer (TMT and label-free) parsed into one common structure. 13 normalisation methods, optional batch correction, 20 imputation methods (including two-stage MAR/MNAR hybrids and the probabilistic model limpa), and testing with limma or limpa. Benchmarking that ranks method combinations by ground-truth simulation and on spike-in datasets, so a pipeline can be justified on the data at hand rather than by convention. Results carry the per-group missingness measured before imputation alongside the fold change, which is what tells a genuine on/off signal from a fold change built entirely on imputed values. Nine vignettes, all executing real code on the shipped dataset. Documentation: https://sciordia.github.io/NADIA/ Full changelog in NEWS.md.