MALDI Deamidation Score (MDS): A fast and flexible method for assessing deamidation in ZooMS data and its application to the Denisova Cave bone assemblage

Estimating deamidation from ZooMS spectra has frequently been achieved using the q2e method due to its high-throughput capacity and ease of use. Despite this, q2e only works with spectral data in txt format, operates slowly, and employs a genetic algorithm for fitting, which is stochastic and less interpretable. Furthermore, q2e only estimates deamidation at the peptide level and does not generate a sample-wide summary. The introduction of the Parchment Glutamine Index (PQI) presented an alternative method. Initially designed for a large ZooMS dataset of parchment, it utilises weighted least squares and a linear mixed-effects model (LME) to generate a peptide deamidation estimation and a sample level index, respectively. To address the limitations of q2e and expand the applicability of PQI to a wider range of archaeological tissues and MALDI-TOF-derived data (such as ZooMS data on bone collagen), we developed the MALDI Deamidation Score (MDS), an iteration of the PQI method optimised for handling large-scale datasets. MDS is more streamlined for analysing multi-species data with customisable peptide lists, offering dramatically reduced processing time. Using the published Denisova Cave ZooMS assemblage, we demonstrate that different peptides exhibit varying deamidation patterns over time, making the use of a single peptide to represent overall deamidation potentially biased. Such information is invaluable for investigating key questions such as protein preservation and site formation processes. SIGNIFICANCE: We introduce the MALDI Deamidation Score (MDS), a new model that estimates deamidation at both peptide and sample levels from MALDI-derived datasets. MDS streamlines multi-species analyses with customisable peptide lists and markedly reduces processing time, enabling robust and large-scale deamidation estimation. Before MDS, deamidation in ZooMS has often been inferred from a single peptide common to Eurasian terrestrial mammals. Applied to the published Denisova Cave ZooMS assemblage, MDS reveals peptide-specific temporal deamidation patterns, showing that single-peptide proxies can be biased as peptides differ in deamidation rates and may be subject to different taphonomic processes. These results have broad relevance for studies of protein preservation and site-formation processes related to molecular taphonomy across archaeological and paleontological contexts. Looking ahead, as MDS allows a customisable peptide list, it has the potential to extend deamidation analysis of ZooMS data beyond terrestrial mammals to marine mammals and non-mammalian vertebrates (e.g., birds), and to different proteinaceous material such as keratin. It also has the potential to be compatible with different instruments used in peptide mass fingerprinting, such as MALDI-TOF, MALDI-FTICR, and TIMS-TOF. We believe that MDS will contribute greatly to our understanding of deamidation on different peptide sequences under various environmental and burial conditions.