Capturing the synaptic proteome: approaches for measuring and defining the synapse

Chapter 2 first establishes that the probability of detection in proteomics depends on the concentration of the analyte. This causes undersampling of quantitative data on medium to low abundant proteins that results in missing data, which can be substantial among replicates (>30%) and reduces the number of quantifiable proteins. A censoring model coined EBRCT is introduced that takes the pattern of missingness into account in differential expression analysis. Application to a benchmark dataset demonstrates improved performance of the EBRCT model in comparison with alternative models. Chapter 3 explores an alternative mode of operation, coined WiSIM-DIA, for mass spectrometers and evaluates results together with state-of-the-art SWATH-MS. WiSIM-DIA combines SWATH-MS and wide-SIM (wide selected-ion monitoring) windows to partition the precursor mass-over-charge space to produce high-quality precursor ion chromatograms. This improves MS1 peak area-based quantification in a DIA strategy, in contrast to the SWATH-MS strategy that utilizes MS2 peak areas. Both strategies show strong overlap in the set of quantified peptides, but also exhibit unique advantages. Chapter 4 is aimed at delineating biochemical impurities from protein constituents of a synaptic subcellular fraction of interest by application of a combination of quantitative proteomics and correlation-based data analysis to a series of related biochemical subfractions. Here, we do not rely on the protein identity list of the respective biochemical subfraction alone, but instead consider the protein abundances relative to related subfractions such as synaptosomes and synaptic membranes. Using canonical PSD proteins as a reference, which are enriched in the PSD biochemical subfraction and relatively low abundant in other subfractions, we searched for proteins that exhibit a strong correlation profile over all samples. The candidate protein list found through this bioinformatics approach indeed contained previously established PSD-enriched proteins among the top scoring proteins, validating the approach. We identified multiple candidate proteins that are likely synaptic and validated two using high-resolution microscopy. Chapter 5 is aimed at using SWATH-MS to quantify levels of hippocampal synaptic proteins of four species, the rodents; mouse and rat, and the primates; marmoset and human. This enabled reliable detection of many protein abundance differences between species with mostly small fold changes. We used a set of proteins regulated by a strong fear learning paradigm impacting on the hippocampus to represent synaptic plasticity related proteins. Using this set of proteins we asked whether its constituents would belong to the rodent-primate conserved or rather the differentially expressed part of the synaptic proteome. The latter was true. This indicates that within the synaptic proteome those proteins of which expression differences maximally evolved during evolution are overrepresented in the plasticity response. Chapter 6 describes an evidence-based, expert-curated knowledgebase for the synapse coined SynGO. First, an extensive ontology was developed to define synaptic processes and cellular components. After achieving consensus on this framework within the SynGO consortium, a worldwide collaboration of many expert laboratories in the synapse research community, domain experts systematically described synaptic protein functions and localizations for 1112 synaptic genes. This data was then used to empower bioinformatic analyses of the synapse that were previously not feasible due to a lack of high quality data at such large scale. We found that SynGO genes are exceptionally large, well conserved, and intolerant to mutations (as compared to other genes). Furthermore, a strong enrichment among genes associated with brain disorders was observed. All data was integrated into the Gene Ontology database and an online data analysis platform was developed to facilitate usage of the SynGO knowledgebase