Profiling Amino Acid Substitutions in Proteomic Datasets of Alzheimer’s Disease Reveals Biological Significance

Background: Amino acid (AA) substitutions in Alzheimer’s disease (AD) have been studied almost exclusively at the DNA and RNA levels, and the proteome-level substitution landscape remains largely uncharacterized. Methods: We applied an open-search substitutomics pipeline (PIPI-C with Comet re-scoring, 1% PSM-level FDR) to two independent AD brain proteomic datasets, the AMP-AD superior temporal gyrus cohort (244 individuals) and PXD013753, excluding substitutions whose mass shifts coincide with common post-translational modifications. Results: We quantified 789 substitution-bearing peptidoforms from 185 proteins, corresponding to 784 protein-level substitution events. N > M was the most frequent N > X substitution in both datasets. In an institution-stratified case–control meta-analysis, 138 peptidoforms were more and 142 less abundant in AD than in neuropathological controls. Of 774 genome-mappable events, 710 (91.7%) were recorded in neither UniProt nor gnomAD, and only five matched a population variant with allele frequency ≥ 0.1%. The unrestricted search recovered six literature-reported sites in GFAP, ACTG1, and HBB. Conclusions: Proteome-level profiling reveals substitution events in AD brains that catalogued germline variation does not explain; their molecular origin requires orthogonal validation.