Les récents progrès instrumentaux en spectrométrie de masse, notamment en terme de- rapidité de balayage et de résolution, ont permis l'émergence de l'approche « data independent acquisition» (DIA). Cette approche promet de combiner les points forts des approches « shotgun » et ciblées,mais aujourd'hui l'analyse des données DIA reste compliquée. L'objectif de cette thèse a été de développer des méthodes innovantes de spectrométrie de masse, et en particulier d'améliorer l'analyse des données DIA. De plus, nous avons développé une approche originale Top 3-ID-DIA, permettant à la fois un profilage complet des protéines de la cellule hôte (HCP) ainsi qu'une quantification absolue d'HCP clés dans les échantillons d'anticorps monoclonaux (mAb), au sein d'une même analyse.Cette méthode est prête à être implémentée en industrie, et pourrait fournir un support en temps réel aux développements du procédé de production de mAb, ainsi que pour évaluer la pureté des biomédicaments. / Recent instrumental developments in mass spectrometry, notably in terms of scan speed and resolution, allowed the emergence of “data independent acquisition” (DIA) approach. This approach promises to combine the strengths of both shotgun and targeted proteomics, but today DIA data analysis remains challenging. The objective of my PhD was to develop innovative mass spectrometry approaches, and in particular to improve DIA data analysis. Moreover, we developed an original Top 3-ID-DIA approach, allowing both a global profiling of host cell proteins (HCP) and an absolute quantification of key HCP in monoclonal antibodies samples, within a single analysis. This method is ready to be transferred to industry, and could provide a real time support for mAb manufacturing process development, as well as for product purity assessment.

The main approach to studying the proteome is a technique called data dependent acquisition (DDA). In DDA, peptides are analyzed by mass spectrometry to determine the protein composition of a biological isolate. However, DDA is limited in its ability to analyze the proteome, in that it only selects the most abundant ions for analysis, and different protein identifications can result even if the same sample is analyzed multiple times in succession. Data independent acquisition (DIA) is a newly developed method that should be able to solve these limitations and improve our ability to analyze the proteome. We used an implementation of DIA (SWATH) to perform relative protein quantitation in the model bacterial system, Clostridium stercorarium, using two different carbohydrate sources, and found that it was able to provide precise quantitation of proteins and was overall more consistent in its ability to identify components of the proteome than DDA. Relative quantitation of proteins is an important method that can determine which proteins are important to a biochemical process of interest. How we determine which proteins are differentially regulated between different conditions is an important question in proteomic analysis. We developed a new approach to analyzing differential protein expression using variation between biological replicates to determine which proteins are being differentially regulated between two conditions. This analysis showed that a large proportion of proteins identified by quantitative proteomic analysis can be differentially regulated and that these proteins are in fact related to biological processes. Analyzing changes in protein expression is a useful tool that can pinpoint many key processes in biological systems. However, these techniques fail to take into account that enzyme activity is regulated by other factors than controlling their level of expression. Activity based protein profiling (ABPP) is a method that can determine the activity state of an enzyme in whole cell proteomes. We found that enzyme activity can change in response to a number of different conditions and that these changes do not always correspond with compositional changes. Mass spectrometry techniques were also used to identify serine hydrolases and characterize their expression in this organism. / February 2016

Infectious diseases caused by viruses such as influenza A virus (IAV) and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pose major threats to human health. Glycosylation, a post-translational modification critical for biological functions including receptor recognition and binding, cell adhesion, and protein folding, is a key mediator of the interaction between viruses and host cells. IAV and SARS-CoV-2 recognize and bind to glycans on host cells prior to uptake by the cells; by the same token, the glycoproteins hemagglutinin of IAV and the spike protein of SARS-CoV-2 are the targets of both host immune molecules and vaccines. The diversity of glycans, structures made up of oligosaccharide residues in complex, branched configurations, can in part be attributed to the push and pull of evolutionary pressures from infectious disease agents such as these viral pathogens. Evolving host glycans may gain the ability to evade recognition by viruses, and likewise, the evolution of viral glycans may result in viral evasion from immune responses. Thus, for a complete understanding of host-pathogen interactions, detailed characterization of glycoproteins that quantitatively measures changes in glycosylation is necessary. However, a number of factors makes quantitative characterization of glycoproteins difficult. Firstly, glycans are highly heterogeneous with dozens of possible glycans at a given glycosylation site and different occupancy levels at each site. Secondly, a particular glycoform may have very low abundance, making the signals difficult to detect. Thirdly, it is difficult to achieve deep, quantitative measurement of glycoprotein glycans using conventional liquid chromatography-mass spectrometry experiments. The usual mass spectrometry methods are not adequate because they are biased towards selecting higher abundance precursors, which leave many glycopeptide glycoforms undetected. This dissertation begins with an assessment of the current state-of-the-art of glycoproteomics using mass spectrometry to give context to our primary research discussed in subsequent chapters. Chapter 2 describes the use of a modified Tanimoto similarity coefficient to quantify the glycosylation similarity between two variants of a strain of IAV, wild-type and mutant, both expressed in embryonated chicken eggs. Our results indicate that even subtle changes in the amino acid sequence of hemagglutinin can result in measurably distinct glycosylation. Chapter 3 expands the number of comparisons of IAV strains made in the previous chapter to include strains produced in a mammalian expression vector, Madin-Darby canine kidney cells. We show that the choice of expression system can change the population of glycoforms at some but not necessarily all glycosylation sites. In addition, we explore data-independent acquisition mass spectrometry to improve upon sensitivity and selectivity of glycopeptide identification. In Chapter 4, this data-independent acquisition method is applied to the quantitative characterization of SARS-CoV-2 spike protein. The work presented here provides a significant contribution toward improving the confident detection and assignment of site-specific glycopeptides. Furthermore, understanding how to measure changes in glycosylation in related viral glycoprotein variants offers opportunities to include consideration of specific glycosylations in the design of vaccines to potentially improve efficacy against continually evolving viruses.

No / Aim: A diagnostic test that could detect Treponema pallidum antigens in urine would facilitate the prompt diagnosis of syphilis. Materials & methods: Urine from 54 individuals with various clinical stages of syphilis and 6 controls were pooled according to disease stage and interrogated with complementary mass spectrometry techniques to uncover potential syphilis biomarkers. Results & conclusion: In total, 26 unique peptides were uncovered corresponding to four unique T. pallidum proteins that have low genetic sequence similarity to other prokaryotes and human proteins. This is the first account of direct T. pallidum protein detection in human clinical samples using mass spectrometry. The implications of these findings for future diagnostic test development is discussed. Data are available via ProteomeXchange with identifier PXD009707.

As high-throughput biology experiments generate increasing amounts of data, the field is naturally turning to data-driven methods for the analysis and extraction of novel insights. These insights into biological systems are crucial for understanding disease progression, drug targets, treatment development, and diagnostics methods, ultimately leading to improving human health and well-being, as well as, deeper insight into cellular biology. Biological data sources such as the genome, transcriptome, proteome, metabolome, and metagenome provide critical information about biological system structure, function, and dynamics. The focus of this licentiate thesis is on proteomics, the study of proteins, which is a natural starting point for understanding biological functions as proteins are crucial functional components of cells. Proteins play a crucial role in enzymatic reactions, structural support, transport, storage, cell signaling, and immune system function. In addition, proteomics has vast data repositories and technical and methodological improvements are continually being made to yield even more data. However, generating proteomic data involves multiple steps, which are prone to errors, making sophisticated models essential to handle technical and biological artifacts and account for uncertainty in the data. In this licentiate thesis, the use of machine learning and probabilistic methods to extract information from mass-spectrometry-based proteomic data is investigated. The thesis starts with an introduction to proteomics, including a basic biological background, followed by a description of how massspectrometry-based proteomics experiments are performed, and challenges in proteomic data analysis. The statistics of proteomic data analysis are also explored, and state-of-the-art software and tools related to each step of the proteomics data analysis pipeline are presented. The thesis concludes with a discussion of future work and the presentation of two original research works. The first research work focuses on adapting Triqler, a probabilistic graphical model for protein quantification developed for data-dependent acquisition (DDA) data, to data-independent acquisition (DIA) data. Challenges in this study included verifying that DIA data conformed with the model used in Triqler, addressing benchmarking issues, and modifying the missing value model used by Triqler to adapt for DIA data. The study showed that DIA data conformed with the properties required by Triqler, implemented a protein inference harmonization strategy, and modified the missing value model to adapt for DIA data. The study concluded by showing that Triqler outperformed current protein quantification techniques. The second research work focused on developing a novel deep-learning based MS2-intensity predictor by incorporating the self-attention mechanism called transformer into Prosit, an established Recurrent Neural Networks (RNN) based deep learning framework for MS2 spectrum intensity prediction. RNNs are a type of neural network that can efficiently process sequential data by capturing information from previous steps, in a sequential manner. The transformer self-attention mechanism allows a model to focus on different parts of its input sequence during processing independently, enabling it to capture dependencies and relationships between elements more effectively. The transformers therefore remedy some of the drawbacks of RNNs, as such, we hypothesized that the implementation of MS2-intensity predictor using transformers rather than RNN would improve its performance. Hence, Prosit-transformer was developed, and the study showed that the model training time and the similarity between the predicted MS2 spectrum and the observed spectrum improved. These original research works address various challenges in computational proteomics and contribute to the development of data-driven life science. / Allteftersom high-throughput experiment genererar allt större mängder data vänder sig området naturligt till data-drivna metoder för analys och extrahering av nya insikter. Dessa insikter om biologiska system är avgörande för att förstå sjukdomsprogression, läkemedelspåverkan, behandlingsutveckling, och diagnostiska metoder, vilket i slutändan leder till en förbättring av människors hälsa och välbefinnande, såväl som en djupare förståelse av cell biologi. Biologiska datakällor som genomet, transkriptomet, proteomet, metabolomet och metagenomet ger kritisk information om biologiska systems struktur, funktion och dynamik. I licentiatuppsats fokusområde ligger på proteomik, studiet av proteiner, vilket är en naturlig startpunkt för att förstå biologiska funktioner eftersom proteiner är avgörande funktionella komponenter i celler. Dessa proteiner spelar en avgörande roll i enzymatiska reaktioner, strukturellt stöd, transport, lagring, cellsignalering och immunsystemfunktion. Dessutom har proteomik har stora dataarkiv och tekniska samt metodologiska förbättringar görs kontinuerligt för att ge ännu mer data. Men för att generera proteomisk data krävs flera steg, som är felbenägna, vilket gör att sofistikerade modeller är väsentliga för att hantera tekniska och biologiska artefakter och för att ta hänsyn till osäkerhet i data. I denna licentiatuppsats undersöks användningen av maskininlärning och probabilistiska metoder för att extrahera information från masspektrometribaserade proteomikdata. Avhandlingen börjar med en introduktion till proteomik, inklusive en grundläggande biologisk bakgrund, följt av en beskrivning av hur masspektrometri-baserade proteomikexperiment utförs och utmaningar i proteomisk dataanalys. Statistiska metoder för proteomisk dataanalys utforskas också, och state-of-the-art mjukvara och verktyg som är relaterade till varje steg i proteomikdataanalyspipelinen presenteras. Avhandlingen avslutas med en diskussion om framtida arbete och presentationen av två original forskningsarbeten. Det första forskningsarbetet fokuserar på att anpassa Triqler, en probabilistisk grafisk modell för proteinkvantifiering som utvecklats för datadependent acquisition (DDA) data, till data-independent acquisition (DIA) data. Utmaningarna i denna studie inkluderade att verifiera att DIA-datas egenskaper överensstämde med modellen som användes i Triqler, att hantera benchmarking-frågor och att modifiera missing-value modellen som användes av Triqler till DIA-data. Studien visade att DIA-data överensstämde med de egenskaper som krävdes av Triqler, implementerade en proteininferensharmoniseringsstrategi och modifierade missing-value modellen till DIA-data. Studien avslutades med att visa att Triqler överträffade nuvarande state-of-the-art proteinkvantifieringsmetoder. Det andra forskningsarbetet fokuserade på utvecklingen av en djupinlärningsbaserad MS2-intensitetsprediktor genom att inkorporera self-attention mekanismen som kallas för transformer till Prosit, en etablerad Recurrent Neural Network (RNN) baserad djupinlärningsramverk för MS2 spektrum intensitetsprediktion. RNN är en typ av neurala nätverk som effektivt kan bearbeta sekventiell data genom att bevara och använda dolda tillstånd som fångar information från tidigare steg på ett sekventiellt sätt. Självuppmärksamhetsmekanismen i transformer tillåter modellen att fokusera på olika delar av sekventiellt data samtidigt under bearbetningen oberoende av varandra, vilket gör det möjligt att fånga relationer mellan elementen mer effektivt. Genom detta lyckas Transformer åtgärda vissa nackdelar med RNN, och därför hypotiserade vi att en implementation av en ny MS2-intensitetprediktor med transformers istället för RNN skulle förbättra prestandan. Därmed konstruerades Prosit-transformer, och studien visade att både modellträningstiden och likheten mellan predicerat MS2-spektrum och observerat spektrum förbättrades. Dessa originalforskningsarbeten hanterar olika utmaningar inom beräkningsproteomik och bidrar till utvecklingen av datadriven livsvetenskap. / <p>QC 2023-05-22</p>

<p>Urine-based biomarkers offer numerous advantages for clinical analysis, including non-invasive collection, a suitable sample source for longitudinal disease monitoring, a better screenshot of disease heterogeneity, higher sample volumes, faster processing times, and lower rejection rates and costs. They will be extremely useful in a clinical trial context, which can be applied alone or in combination with other methods as long as they demonstrate clear reproducibility across cohorts. While biofluids such as urine present enormous challenges with a wide dynamic range and extreme complex typically dominated by a few highly abundant proteins, we have demonstrated that the analytical issue can be efficiently addressed by focusing on extracellular vesicles (EVs), tiny packages released by all kinds of cells. These tiny packages contain different kinds of molecules from inside the cells. Here, we established a robust EV isolation and characterization platform to screen and validate Parkinson’s Disease (PD) and Renal Cell Carcinoma (RCC) biomarkers from urine. PD is a progressive neurological disorder affecting body movement because some brain cells stop producing dopamine. PD is often not diagnosed until it has advanced, making early detection crucial. We investigated urinary EVs from 138 individuals to enable early detection and found several proteins involved in PD development that could be biological indicators for early disease detection. Several biochemical techniques were applied to verify our findings. In the second project, we attempted to develop a novel diagnostic technique for early intervention of RCC. Here, we made our efforts to develop a quantitative method based on data-independent acquisition (DIA) mass spectrometry to analyze urinary EV phosphoproteomics for non-invasive RCC biomarker screening. Combined with our in-house EVtrap method for EV isolation and PolyMAC enrichment of phosphopeptides, we quantified 2,584 unique phosphosites. We observed unique upregulated phosphosites and pathways differentiating healthy control (HC), chronic kidney disease (CKD), low-grade, and high-grade clear cell RCC. These applications have a significant promise for early PD and RCC diagnosis and monitoring based on actual functional proteins with urine as the source. These studies might provide a viable path to developing urinary EV-based disease diagnosis.</p>