#33 · Proteomics

Analytical choices shape accuracy in MS-based strain, variant and quantity inference

The dominant thread is evaluation of how analytical decisions in computational proteomics workflows affect accuracy, from raw mass spectrometry data to protein identifications, strain assignments, or quantities. The DIA-NN workflow for Legionella pneumophila integrates allelic protein sequence variability through clustering and protein inference, targeting single amino acid variants and accessory proteome variability in data-independent acquisition. NovoTax combines de novo peptide sequencing with proteome database search to identify prokaryotic strains directly from raw bottom-up proteomics data. In metaproteomics, MS1 area-under-the-curve quantification is benchmarked against spectral counting, and AUC without match-between-runs is reported to give better accuracy and dynamic range. The items address distinct problems (variant-aware database construction, strain-level identification, and quantification strategy), so the shared element is methodological comparison rather than a single new algorithm class. Novelty lies in the specific design choices examined.