Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 68-74). / Background: The commoditization of high-throughput gene expression sequencing and microarrays has led to a proliferation in both the amount of genomic and clinical data that is available. Descriptive textual information deposited with gene expression data in the Gene Expression Omnibus (GEO) is an underutilized resource because the textual information is unstructured and difficult to query. Rendering this information in a structured format utilizing standard medical terms would facilitate better searching and data reuse. Such a procedure would significantly increase the clinical utility of biomedical data repositories. Methods: The thesis is divided into two sections. The first section compares how well four medical terminologies were able to represent textual information deposited in GEO. The second section implements free-text search and faceted search and evaluates how well they are able to answer clinical queries with varying levels of complexity. Part I: 120 samples were randomly extracted from samples deposited in the GEO database from six clinical domains-breast cancer, colon cancer, rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), type I diabetes mellitus (IDDM), and asthma. These samples were previously annotated manually and structured textual information was obtained in a tag:value format. Data was mapped to four different controlled terminologies: NCI Thesaurus, MeSH, SNOMED-CT, and ICD- 10. The samples were assigned a score on a three-point scale that was based on how well the terminology was able to represent descriptive textual information. Part II: Faceted and free-text search tools were implemented, with 300 GEO samples included for querying. Eight natural language search questions were selected randomly from scientific journals. Academic researchers were recruited and asked to use the faceted and free-text search tools to locate samples matching the question criteria. Precision, recall, F-score, and search time were compared and analyzed for both free-text and faceted search. Results: The results show that the NCI Thesaurus consistently ranked as the most comprehensive terminology across all domains while ICD-10 consistently ranked as the least comprehensive. Using NCI Thesaurus to augment the faceted search tool, each researcher was able to reach 100% precision and recall (F-score 1.0) for each of the eight search questions. Using free-text search, test users averaged 22.8% precision, 60.7% recall, and an F-score of 0.282. The mean search time per question using faceted search and free-text search were 116.7 seconds, and 138.4 seconds, respectively. The difference between search time was not statistically significant (p=0. 734). However, paired t-test analysis showed a statistically signficant difference between the two search strategies with respect to precision (p=O.001), recall (p=O.042), and F-score (p<0. 001). Conclusion: This work demonstrates that biomedical terms included in a gene expression database can be adequately expressed using the NCI Thesaurus. It also shows that faceted searching using a controlled terminology is superior to conventional free-text searching when answering queries of varying levels of complexity. / S.M.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2010. / Page 84 blank. Cataloged from PDF version of thesis. / Includes bibliographical references. / The innate immune system is the first line of host defense, and its ability to propagate antimicrobial and inflammatory signals from the cellular microenvironment to the tissue at-large is critical for survival. In a remarkably complex microenvironment, cells are constantly processing external cues, initiating convoluted intracellular signaling cascades, and interacting with neighboring cells to generate a global, unified response. At the onset of infection or sterile injury, individual cells sense danger or damage signals and elicit innate immune responses that spread from the challenged cells to surrounding cells, thereby establishing an overall inflammatory state. However, little is known about how these dynamic spatiotemporal responses unfold. Through the use GFP reporters, in vitro transplant coculture systems, and in vivo models of infection and sterile injury, this thesis describes identification of a gap junction intercellular communication pathway for amplifying immune and inflammatory responses, and demonstrates its importance in host innate immunity. The first section describes development of stable GFP reporters to study the spatiotemporal activation patterns of two key transcription factors in inflammation and innate immunity: Nuclear factor-KappaB (NFKB) and Interferon regulatory factor 3 (IRF3). Stimulation of NFKB-GFP reporters resulted in a spatially homogeneous pattern of activation, found to be largely mediated by paracrine action of the pro-inflammatory cytokine TNFa. In contrast, the activation of IRF3 was spatially heterogeneous, resulting in the formation of multicellular colonies of activated cells in an otherwise latent background. This pattern of activation was demonstrated to be dependent on cell-cell contact mediated communication between neighboring cells, and not on paracrine signaling. The second section describes the discovery of a gap junction intercellular communication pathway responsible for the formation of IRF3 active colonies in response to immune activation. Cell sorting and gene expression profiling revealed that the activated reporter colonies, collectively, serve as the major source of critical antimicrobial and inflammatory cytokines. Using in vitro transplant coculture systems, colony formation was found to be dependent on gap junction communication. Blocking gap junctions with genetic specificity severely compromised the innate immune system's ability to mount antiviral and inflammatory responses. The third section illustrates an application of the gap junction-induced amplification of innate immunity phenomenon in an animal model of sterile injury. Drug-induced liver injury was shown to be dependent on gap junction communication for amplifying sterile inflammatory signals. Mice deficient in hepatic gap junction protein connexin 32 (Cx32) were protected against liver damage, inflammation, and death in response to hepatotoxic drugs. Co-administration of a selective pharmacologic Cx32 inhibitor with hepatotoxic drugs significantly limited hepatocyte damage and sterile inflammation, and completely abrogated mortality. These finds suggests that co-formulation of gap junction inhibitors with hepatotoxic drugs may prevent liver failure in humans, and potentially limit other forms of sterile injury. In summary, this thesis demonstrates the development of novel tools for investigating the spatiotemporal dynamics of cellular responses, describes how these tools were utilized to discover a basic gap junction communication pathway critical in innate immunity, and provides evidence for the clinical relevance of this pathway in sterile inflammatory injury. / by Suraj Jagdish Patel. / Ph.D.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2009. / Cataloged from PDF version of thesis. / Includes bibliographical references. / The integrated network of biochemical reactions known collectively as metabolism is essential for life, and dysfunction in parts of this network causes human disease - both rare, inherited disorders and common diseases such as diabetes mellitus. The study of metabolic disease depends upon quantitative methods which are traditionally custom-tailored to a given compound. Recent advances in technologies such as mass spectrometry now enable the simultaneous measurement of a diverse metabolite collection spanning multiple biological pathways, an approach known as metabolic profiling or metabolomics. This dissertation describes the development of one such metabolic profiling system and its application to the study of two major topics in human energy metabolism: the fasting:feeding transition and mitochondrial disease. In the first study, we profile human plasma in response to glucose ingestion, detecting dozens of metabolite changes and identifying several distinct effects of insulin. Based on these observations, we propose a multivariate view of insulin sensitivity, and show that individuals at risk for developing diabetes mellitus can differ in their insulin response profile, a concept of potential value for estimating disease risk and progression. In the second study, we elucidate a metabolic signature of human mitochondrial disease that reflects substrate oxidation, biosynthesis and energy charge. / (cont.) We demonstrate that the culture media profile of a cellular disease model of mitochondrial dysfunction reflects the plasma profile of human patients, an approach that could be applicable to other diseases as well. In addition, we show that a combination of metabolites distinguishes individuals with mitochondrial disease from healthy individuals better than the currently used diagnostic markers. Our findings provide insight into human disorders of energy metabolism, and demonstrate the utility of a profiling approach for the understanding of metabolic disease. / by Oded Shaham. / Ph.D.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2011. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Aneuploidy, an abnormal complement of chromosomes, is present in approximately 90% of human malignancies. Despite over 100 years of research, many questions remain regarding the contribution of aneuploidy to the cancer phenotype. In this thesis, we develop computational methods to infer the presence and specific patterns of aneuploidy across thousands of primary cancer tissue specimens. We then combine these inferences with clinical and genomic features of the cancer samples to refine our understanding of both the clinical implications of aneuploidy, and how it evolves in various human cancers. We identified a signature of chromosomal instability from specific genes whose expression was consistently correlated with aneuploidy in several cancer types, and which was predictive of poor clinical outcome multiple cancer types. Current genomic characterization techniques measure somatic alterations in a cancer sample in units of genomes (DNA mass). The meaning of such measurements is highly dependent on the tumors purity and its overall ploidy; they are hence complicated to interpret and compare across samples. Ideally, copy-number should be measured in copies-per-cancer-cell. Such measurements are straightforward to interpret and, for alterations that are fixed in the cancer cell population, are simple integer values. We develop two computational methods to infer tumor purity and malignant cell ploidy directly from allelic analysis of DNA. First we describe HAPSEG, a probabilistic method to interpret bi-allelic marker data in cancer samples in order to produce genome-wide estimates of homologue specific copy-ratios. Second, we describe ABSOLUTE, a method that infers purity, ploidy, and absolute copy-numbers from the estimates produced by HAPSEG. In addition, ABSOLUTE can analyze point mutations to detect subclonal heterogeneity and somatic homozygosity. We used ABSOLUTE to analyze ovarian cancer data and discovered that 54% of somatic point mutations were, in fact, subclonal. In contrast, mutations occurring in key tumor suppressor genes, TP53 and NF1 were predominantly clonal and homozygous. Analysis of absolute allelic copy-number profiles from 3,155 cancer specimens revealed that genome-doubling events are common in human cancer, and likely occur in already aneuploid cells in many cancer types. By correlating genome-doubling status with mutation data, we found that homozygous mutations in NF1 occurred predominantly in non-doubled samples. This finding suggests that genome doubling influences the pathways of tumor progression, with recessive inactivation being less common after genome doubling. / by Scott L. Carter. / Ph.D.

Thesis (Ph. D.)--Harvard--Massachusetts Institute of Technology Division of Health Sciences and Technology, 2002. / Includes bibliographical references (leaves 106-120). / Our lab's long-term goal is to elucidate the circuitry of the visual cortex, to develop quantitative computational models of neuronal function in the visual cortex, and to establish how these models may relate to visual perception and visually guided behavior. Central to this goal is the analysis of functional architecture, which is crucial to an understanding of how the brain works. In my thesis research, I applied behavioral and microstimulation techniques to demonstrate the causal connections between neural activity and behavior. Understanding these relationships is one of the fundamental issues needed to be addressed in Neurobiology. Specifically, I focused on the functional analysis of the middle temporal visual area (MT) and the behavior associated with it. MT is an extrastriate area that is primarily involved in visual motion processing. A very important function within MT is a segregation of center-surround interactions which plays a critical role in processing visual motion cues. There are two types of motion center-surround interactions in MT neurons: surrounds may reinforce (at wide-field sites) or suppress (at local-motion sites) the centers' directional responses. They are important in representing the initial stages of a functional segregation between wide-field and local-contrast motion processing. To further study the computational model used by the brain to readout sensory information, I conducted microstimulation experiments in MT by changing stimulation amplitudes (from 10/LA to 160tA) and frequencies (from 25Hz to 500Hz). Microstimulation can introduce an additional velocity signal into MT and the pursuit and saccadic systems usually compute a vector average of the visually evoked and microstimulation-induced velocity signals. / (cont.) Increasing either amplitude or frequency generally increases the relative weight of the electrical velocity signal,' with the effects of amplitude being slightly more prominent. In addition, applying higher current fre-quencies appears to preserve the directionality of microstimulation better than does applying higher current amplitudes. With increasing frequencies, the magnitude of the electrical velocity signal either increased or remained constant, while its direction remained consistent. In contrast, increasing current amplitude tended to decrease the magnitude of the signal and increased its variability in direction. This finding is consistent with the idea that large current amplitudes, which presumably activate many MT columns signaling different directions, introduce noise into the behavior. My preliminary results have demonstrated that microstimulation in MT can also introduce an additional positional signal into the saccade system and that this electrical signal is combined with the visually evoked signal through a vector summation mechanism. The direction of this electrical signal is highly correlated with the position of the receptive field relative to the fixation point. To test the notion that the center-surround properties of MT neurons may be important for signaling the relative motion between object and background, we conducted behavioral experiments by using real background motion to simulate the microstimulation experiments at wide-field sites ... / by Ruilin Zhao. / Ph.D.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, 2007. / Includes bibliographical references. / Drug eluting stents have revolutionized the practice of medicine and the landscape of medical devices. Yet, more than four years after introduction clinical trial data and clinical use have still not fully clarified what drives the safety and efficacy of these devices. The goal of this thesis was to help fill this void by describing the mechanisms by which stent-eluted drugs are distributed within healthy and atherosclerotic vascular models. In the first part of the thesis we investigated the effect of drug physicochemical properties on drug deposition, retention, and distribution in a healthy vascular model. We found that hydrophobic drugs are deposited to a far greater degree than hydrophilic drugs, with longer retention times, and distribution patterns that likely track with specific and general binding sites. The second part of the thesis investigated how arterial ultrastructure in health and disease modulates the arterial deposition and distribution of hydrophobic antiproliferative drugs used with drug-eluting stents. We tracked the distribution of radiolabeled and FITC-labeled compounds and demonstrated that macrostructural changes in arterial architecture led to profound changes in drug deposition. Paclitaxel in particular was sensitive to tissue state. / (cont.) This drug binds specifically to tubulin and to lesser extent in a general manner to elastic. Drug levels fell as paclitaxel, tubulin and elastin were displaced by lipid and collagen. These observations might well explain how drugs may partition within different arterial lesions as determined by lesion composition. Finally, we demonstrated that association with these binding sites was governed by association kinetics that reflects the different components of the arterial wall compartments. Slower release kinetics yielded up to 64% higher deposition of a drug from stents implanted in rabbit iliac arteries over a 28-day period. Mathematical modeling illustrates that the dependence of drug deposition on stent release kinetics is contingent on drug retention. Further model development is implicated for predicting drug deposition profiles for different types of drugs, arterial states, and stent release kinetics. / by Neda Vukmirovic. / Ph.D.

Thesis (Ph. D.)--Harvard-MIT Division of Health Sciences and Technology, February 2004. / Includes bibliographical references (p. 167-175). / This thesis explores two important factors that constrain the syntactic parser of the sentence processing mechanism, syntactic storage costs and plausibility information. It uses behavioral methods to explore the characteristics of the two factors and neuroimaging to explore the underlying neurological substrates associated with these aspects of syntactic processing. Experiment 1 behaviorally demonstrated the presence of syntactic storage costs for predictions of verbs, filler-gaps, and subcategorized prepositional phrases. It is argued that the data support the Dependency Locality Theory (Gibson, 2000) supposition of stored predicted heads as well as a theory of syntax that includes empty categories. Experiment 2 demonstrated brain regions associated with storage and integration cost demands in the contrast of subject-object (SO) and object-subject (OS) sentence structures. The results indicate that the inferior parietal cortex is part of a larger network of cortex, including inferior frontal perisylvian areas, that is involved in the processing of SO vs. OS sentences. However, the involvement is not identical to that of the inferior frontal areas and has a distinct hemodynamic character. Experiment 3 explored regions of the brain involved in the resolution of the main verb/reduced relative (MV/RR) ambiguity. Activation was seen in portions of the angular gyrus and the middle temporal gyrus for a contrast in subject noun plausibility, but not structure ambiguity, indicating that the MV interpretation was still considered even in unambiguously relative clause sentence structures. The unexpected results could imply that syntax is not the only factor that determines [theta]-role assignment and ultimately provide evidence about the brain regions involved in / (cont.) the process of plausibility information resolution in sentence interpretation. / by Evan Chen. / Ph.D.

Thesis (S.M.)--Harvard-MIT Division of Health Sciences and Technology, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 87). / Medicare coverage policies of medical procedures can be promulgated at a national level by the Centers of Medicare and Medicaid Services (CMS) as National Coverage Determinations (NCDs) or at a local level by Medicare contractors as Local Coverage Determinations (LCDs). Although LCDs shouldn't contradict NCDs, they can differ. In the present study, I analyze some factors that could partially account for the differences between NCDs and LCDs. Using the Medicare Coverage Database from CMS, I searched for differences between NCDs and LCDs in five benefit categories: inpatient hospital services, durable medical devices, diagnostic laboratory tests, physician's services and other diagnostic tests. There is a reasonable degree of homogeneity in coverage policies for procedures for which an NCD has been issued: 82% are exactly the same. Most of the differences took the form of exclusions from LCDs, but not from NCDs. For each state, I computed the number of times that LCDs were issued and the number of times that LCDs differed from NCDs and searched for possible linear or exponential correlation models. The following factors were initially hypothesized to account for these differences: number of Schools of Medicine, number of physicians, GDP per capita by state, state ranking according to number of Level 1 and Level 2 Trauma Centers and the profile of MEDCAC members. At a national level, I found no correlation between the number of LCDs issued or the number of differences between LCDs and NCDs and any of these variables. However, on a sub-analysis at a local level, in some regions I found a positive correlation (r2 >.94) between the following three variables: 1) number of Schools of Medicine, 2) number of physicians, and 3) state ranking according to the number of Level 1 and Level 2 Trauma Centers and the following two parameters: 1) the level of LCD issuing activity, and 2) the number of times that LCDs differ from NCDs. The correlations shown by the performed sub-analysis within regions may imply that more LCDs are issued to restrict coverage when there is a local need to control the excessive demand partially driven by the higher number of hospitals and physicians that are active in pursuing their interests. The fact that these correlations were shown only at a regional level may indicate that when local factors are disregarded, the original hypothesized factors do influence LCD activity, however, at a national level, other hypothetical local factors may have a greater influence on LCD activity and policy discrepancies. In order to have a better understanding of my results and the factors that could account for the differences between NCDs and LCDs, I interviewed four Contractor Medical Directors (CMDs). These interviews indicated that other factors could account for these differences, including the following: a history of abuse and fraud, contractor's budgets, the CMD's special interests and experience, data analysis capabilities, the number of claims and the novelty of the procedure. The impact of these variables on the differences between national and local coverage policies can be an interesting topic for future research on the subject. / by Rafael Díaz Treviño. / S.M.

Thesis (Ph.D.)--Harvard--Massachusetts Institute of Technology Division of Health Sciences and Technology, 2002. / Vita. Page 202 blank. / Includes bibliographical references (p. 197-201). / In the clinical management of voice patients, quantifying vocal function is becoming increasingly important both for corroborating clinicians' subjective impressions during a voice evaluation and for assessing the effectiveness of surgery or voice therapy. Current devices for quantifying vocal function measure acoustic, aerodynamic or electric signals (i.e., sound pressure, airflow, or electroglottography) during short tasks such as reading. One technique that has shown potential for measuring vocal function but has been mostly used to quantify speech-related behaviors besides phonation is measuring the acceleration of the skin near the larynx. The acceleration of the skin on the neck between the cricoid cartilage and the sternal notch arises from the airflow pulses that result from vocal-fold vibration. At least two sets of structures play a role in transforming these airflow pulses into the measured acceleration: the subglottal system, and the tissues between the subglottal airspace and the accelerometer (e.g., tracheal cartilage, skin, etc.). Advantages of measuring acceleration over current techniques include 1) the structures that filter the glottal pulses vary less over time than the vocal tract and thus they may be adequately modeled as time-invariant, making signal processing potentially easier; 2) environmental acoustic noise has a minimal influence on the measured acceleration; and 3) the accelerometer's size and placement make it more comfortable and unobtrusive for extended recordings than current techniques. This thesis work investigated the potential of using the measured acceleration for quantifying vocal function. / by Harold Arthur Cheyne, II. / Ph.D.

Thesis (M. Eng.)--Harvard-MIT Division of Health Sciences and Technology, 2004. / Includes bibliographical references (p. 54-55). / (cont.) nucleus and at the nuclear surface. / Mutations in the lamin A/C gene (Lmna) and the lamin-associated protein emerin gene (EM) cause a variety of human diseases including Emery-Dreifuss muscular dystrophy, dilated cardiomyopathy, familial partial lipodystrophy, Charcot-Marie-Tooth Neuropathy and Hutchinson-Gilford progeria syndrome. The molecular mechanisms underlying the varied phenotypes are unknown, and both a mechanical stress hypothesis and an altered gene expression hypothesis have been proposed to explain the tissue specific effects observed in laminopathies. To investigate the role of emerin in mechanotransduction, lamin A/C deficient (Lmna⁻/⁻) fibroblasts, and emerin deficient (EM⁻/y) fibroblasts were studied for nuclear mechanical properties, cytoskeletal stiffness, and mechanical strain-induced signaling. EM⁻/y fibroblasts exhibited similar cell sensitivity, nuclear and cytoskeletal properties compared to wild type cells under stress and strain. Interestingly, both Lmna⁻/⁻ and EM⁻/y fibroblasts had impaired mechanotransduction, characterized by attenuated expression of the mechanosensitive genes egr-1, iex-1, and txnip in response to mechanical stimulation. In addition, NF-rB signaling appeared disturbed in Lmna⁻/⁻ cells, but normal in EM⁻/y fibroblasts. The relationship between changes in cytoskeletal stiffness recently discovered in Lmna⁻/⁻ cells and nuclear mechanics under strain was explored using a computational finite elemental model. Analysis of the several models using variations in material properties and cell geometry revealed that nuclear shape, material properties of the cytoskeleton and nucleus, as well as the size and location of strain application on the cell are important parameters in determining the magnitude of stress and strain within the / by Janet Hsiao. / M.Eng.