Computational methods in palaeoproteomics: making the most of mass spectrometry data

This thesis addresses the computational challenges that hinder progress in zooarchaeology by mass spectrometry (ZooMS) and liquid chromatography electrospray ionisation tandem mass spectrometry (LC-ESI MS/MS) based palaeoproteomics on ceramics, advocating for the development of open-source, community-driven, collaborative, and interoperable computational resources. Specifically, the aims include standardising data analysis for ZooMS, exploring optimal strategies for querying LC-ESI MS/MS spectra from degraded proteins, and investigating protein adsorption and survival on ceramics applying those strategies. These are described through three chapters based on articles, with an introduction outlining the state of the field. The first of these articles provides a comprehensive guide to data analysis workflows for matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI-TOF) based ZooMS, a high-throughput technique for taxonomic identification using collagen peptide mass fingerprinting (PMF). Here, I show other types of analysis that can be carried out on MALDI-TOF data. Despite the increasing application of ZooMS, the development of open-source software infrastructure for data analysis has lagged. The chapter addresses this by presenting and discussing various open-source data analysis modules and workflows for ZooMS, outlining five main modules: MS data preparation, preprocessing, deamidation, automatic species identification, and postprocessing clustering methods. It uses publicly available data to illustrate these modules and discusses current limitations and the impact of different practices. The next chapter reports on a systematic investigation of the performance of various sequencing tools and search engines (Mascot, MaxQuant, Metamorpheus, pFind, Fragpipe, and DeNovoGUI) for identifying a single degraded protein, bovine β-lactoglobulin (BLG). BLG was experimentally degraded by heating, and the resulting peptides were analysed using LC-ESI MS/MS. The study explores different database choices (targeted dairy vs. whole bovine proteome), digestion options (tryptic, semi-tryptic, non-specific) and explores open-search and de novo sequencing. Open-search is argued to be a highly valuable approach for post-translational modifications (PTM) discovery in palaeoproteomics. The final article reports on cooking experiments involving milk and fish on experimental ceramic briquettes with different compositions, focusing on protein adsorption, survival, and extraction from ceramics. This work uniquely investigates the ceramic fabric itself as a variable influencing protein preservation and extraction. Following the results from chapter 3, it employs workflows that utilise de novo sequencing and open-search. For milk samples, open-search with a wide precursor error tolerance and semi-tryptic digestion was used, which allowed studying PTMs such as lactosylation. For fish samples, de novo sequencing guided the building of a targeted search database. Overall, this work shows the need for the field to focus on building open-source, community-driven, collaborative and interoperable computational resources for ZooMS and LC-ESI MS/MS.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2025

Author dc:creator

• Rodriguez Palomo, Ismael

Advisor dc:contributor.advisor

• Collins, Matthew James

• Bioinformatics
• Ceramics
• Mass spectrometry
• Palaeoproteomics
• ZooMS

Licence

• https://www.repository.cam.ac.uk/bitstreams/21564432-c1bb-4abb-9335-c278a34de75e/download
• https://creativecommons.org/licenses/by-nc-sa/4.0/

Language dc:language

eng

Author Identifier

0000-0001-5313-9709

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/395248