The innate immune system is of central importance to the early containment of infection. When receptors of innate immunity recognize molecular patterns on pathogens, they initiate an immediate immune response by inducing the expression of cytokines and other host defense genes. Altered expression or function of the receptors, the molecules that mediate the signal transduction cascade, or the cytokines themselves can predispose individuals to infectious or autoimmune diseases. Here we used genomic approaches to uncover novel components underlying the innate immune response to cytosolic DNA and to characterize variation in the innate immune responses of human dendritic cells to bacterial and viral ligands. In order to identify novel genes involved in the cytosolic DNA sensing pathway, we first identified candidate proteins that interact with known signaling molecules or with dsDNA in the cytoplasm. We then knocked down 809 proteomic, genomic, or domain-based candidates in a high-throughput siRNA screen and measured cytokine production after DNA stimulation. We identified ABCF1 as a critical protein that associates with DNA and the known DNA-sensing components, HMGB2 and IFI16. We also found that CDC37 regulates stability of the signaling molecule, TBK1, and that chemical inhibition of CDC37 as well as of several other pathway regulators (HSP90, PPP6C, PTPN1, and TBK1) potently modulates the innate immune response to DNA and to retroviral infection. These proteins represent potential therapeutics targets for infectious and autoimmune diseases that are associated with the cytosolic DNA response. We also developed a high-throughput functional assay to assess variation in responses of human monocyte-derived dendritic cellsto LPS (receptor: TLR4) or influenza (receptors: RIG-I and TLR3), with the goal to ultimately map genetic variants that influence expression levels of pathogen-responsive genes. We compared the variation in expression between the dendritic cells of 30 different individuals, and within paired samples from 9 of these individuals collected several months later. We found genes that have significant inter- vs. intra-individual ariation in response to the stimuli, suggesting that there is a substantial genetic component underlying variation in these responses. Such variants may help to explain differences between individuals’ risk for infectious, autoimmune, or other inflammatory diseases.

The rise of high-throughput scientific experimentation and data collection has introduced new classes of statistical and computational challenges. The technologies driving this data explosion are subject to complex new forms of measurement error, requiring sophisticated statistical approaches. Simultaneously, statistical computing must adapt to larger volumes of data and new computational environments, particularly parallel and distributed settings. This dissertation presents several computational and theoretical contributions to these challenges. In chapter 1, we consider the problem of estimating the genome-wide distribution of nucleosome positions from paired-end sequencing data. We develop a modeling approach based on nonparametric templates that controls for variability due to enzymatic digestion. We use this to construct a calibrated Bayesian method to detect local concentrations of nucleosome positions. Inference is carried out via a distributed HMC algorithm that scales linearly in complexity with the length of the genome being analyzed. We provide MPI-based implementations of the proposed methods, stand-alone and on Amazon EC2, which can provide inferences on an entire S. cerevisiae genome in less than 1 hour on EC2. We then present a method for absolute quantitation from LC-MS/MS proteomics experiments in chapter 2. We present a Bayesian model for the non-ignorable missing data mechanism induced by this technology, which includes an unusual combination of censoring and truncation. We provide a scalable MCMC sampler for inference in this setting, enabling full-proteome analyses using cluster computing environments. A set of simulation studies and actual experiments demonstrate this approach's validity and utility. We close in chapter 3 by proposing a theoretical framework for the analysis of preprocessing under the banner of multiphase inference. Preprocessing forms an oft-neglected foundation for a wide range of statistical and scientific analyses. We provide some initial theoretical foundations for this area, including distributed preprocessing, building upon previous work in multiple imputation. We demonstrate that multiphase inferences can, in some cases, even surpass standard single-phase estimators in efficiency and robustness. Our work suggests several paths for further research into the statistical principles underlying preprocessing. / Statistics

The first end point of this study was to identify specific pro-apoptotic or anti-proliferative factors in the breast cancer cell secretome. To this end, we designed an in vitro screen that effectively cross-cultured 20 breast cancer cell lines in each other's conditioned media. We selected the strongest pro-apoptotic hits and performed further proteomic and biochemical characterization in order to analyze their composition. We determined that the pro-apoptotic activity resided in the soluble, exosome-free secreted fraction of triple negative breast cancer cell conditioned medium and used proteomic insights in order to narrow down the list of possible candidate molecules responsible for the apoptotic effect. The second endpoint of this study was to evaluate the particulate fraction found in breast cancer cell conditioned media for diagnostically significant molecules. We isolated cancer exosomes, employing a serial ultracentrifugation protocol, and were able to establish that the exosome cell surface receptors identically reflect the molecular identity of their cell lines of origin. However, downstream protein kinases within exosomes display patterns of depletion or enrichment in comparison to the corresponding cell lines. Overall, we found that the exosome protein composition in breast cancer is informative enough to guide the choice of specific inhibitor treatment in a clinical setting. / Engineering and Applied Sciences

Defining the mechanics and consequences of protein adduction is crucial to understanding the toxicity of reactive electrophiles. Application of tandem mass spectrometry and data analysis algorithms enables detection and mapping of chemical adducts at the level of amino acid sequence. Nevertheless, detection of adducts does not indicate relative reactivity of different sites. In this dissertation I describe a method to measure the kinetics of competing adduction reactions at different sites on the same protein using quantitative mass spectrometry. Adducts are formed by electrophiles at Cys-14 and Cys-47 on the metabolic enzyme glutathione-S-transferase P1-1 and accompanied by a loss of enzymatic activity. Relative quantitation of protein adducts was done by tagging N-termini of peptide digests with isotopically labeled phenyl isocyanate and tracking the ratio of light-tagged peptide adducts to heavy-tagged reference samples. This method was used to measure rate constants for adduction at both positions with two different model electrophiles, IAB and BMCC. The results indicate that Cys-47 was approximately 2-3-fold more reactive toward both electrophiles than was Cys-14. This result was consistent with the relative reactivity of these electrophiles in a complex proteome system. Quantitative analyses of protein modifications provide a means of determining the reactivity and selectivity of damaging protein modifications in chemical toxicity.Another area of study explored in this dissertation is looking at the effects of protein alkylation on activating cellular signaling pathways, specifically the JNK signaling pathway. Protein adduction has been shown to be selective between different alkylating agents. It would then be reasonable to think this selectivity of adduction translates to selectivity of downstream consequences or cellular events directly tied to specific adductions. My work will show how treatment of HEK293 cells with either IAB or BMCC leads to differences in activation of JNK signaling. In addition, I've been able to show a difference in selectivity of a number of adducted targets by each alkylating agent, which are directly involved in regulation of the JNK signaling pathway. These studies illustrate not only the significance of protein adduction, but the importance for continual research to better understand their behavior in living systems.

Two popular workflows exist for quantitative proteome analysis: two-dimensional polyacrylamide gel electrophoresis (2D-PAGE), with staining to visualize proteins, and multidimensional solution phase separations of isotopically labelled peptides coupled to mass spectrometry (MS). However, the development of an alternative strategy, which combines easy-to-read differential profiling as seen in 2D-PAGE, with the sensitivity of MS for detection and identification, is needed. This thesis presents work towards the development of a workflow for high-throughput protein biomarker discovery. Multidimensional separations are vital to obtain sufficient fractionation of complex proteome mixtures. As a first dimension of separation, ion exchange chromatography (IEC) is a common choice, though it has yet to be thoroughly evaluated in terms of its effectiveness as a proteome prefractionation tool. This study used a defined set of protein standards to establish the resolution and proteome yield obtained through IEC. The evaluation uncovered significant bias in terms of protein recovery and separation. To improve throughput of a multidimensional separation strategy, a multiplexed (8-column) reversed phase liquid chromatography (RPLC) platform was constructed. The system design allowed for even distribution of flow across all columns with limited cross-loading during sample loading. This system was directly coupled to matrix-assisted laser desorption/ionization (MALDI) through a novel well plate device. The Teflon wells allowed for high recovery and no cross-contamination during collection/spotting, improved throughput, and greatly reduced the number of sample manipulation steps. An evaluation of MALDI MS, using the ThermoFisher vMALDI LTQ, for quantitative profiling was performed, employing the multiplexed LC-MALDI platform. The use of MALDI MS allowed for fast (< 5.5 hours) acquisition of quantitative data from isotopically differentiated samples partitioned over 640 fractions from two-dimensional LC. Proteins comprising 0.1% of the proteome were detected and quantified using this method. Finally, the effects of varying concentrations of acetonitrile (ACN) upon the products generated from tryptic digestions were explored. Poor enzymatic efficiency in 80% ACN was found to be responsible for an increased concentration of peptides containing missed cleavage sites. These peptides often contained unique amino acid sequences, which were not detected from complete digestions, resulting in improved protein sequence coverage following MS analysis.

Comparing protein structures based on their contact maps is an important problem in structural proteomics. Building a system for reconstructing protein tertiary structures from their contact maps is one of the motivations for devising novel contact map comparison algorithms. Several methods that address the contact map comparison problem have been designed which are briefly discussed in this thesis. However, they suggest scoring schemes that do not satisfy the two characteristics of “metricity” and “universality”. In this research we investigate the applicability of the Universal Similarity Metric (USM) to the contact map comparison problem. The USM is an information theoretical measure which is based on the concept of Kolmogorov complexity. The ultimate goal of this research is to use the USM in case-based reasoning system to predict protein structures from their predicted contact maps. The fact that the contact maps that will be used in such a system are the ones which are predicted from the protein sequences and are not noise-free, implies that we should investigate the noise-sensitivity of the USM. This is the first attempt to study the noise-tolerance of the USM. In this research, as the first implementation of the USM we converted the two-dimensional data structures (contact maps) to one-dimensional data structures (strings). The results of this implementation motivated us to circumvent the dimension reduction in our second attempt to implement the USM. Our suggested method in this thesis has the advantage of obtaining a measure which is noise tolerant. We assess the effectiveness of this noise tolerance by testing different USM implementation schemes against noise-contaminated versions of distinguished data-sets. / Thesis (Master, Computing) -- Queen's University, 2008-09-27 05:53:31.988