Variance Reduction in Analytical Chemistry : New Numerical Methods in Chemometrics and Molecular Simulation

This thesis is based on five papers addressing variance reduction in different ways. The papers have in common that they all present new numerical methods. Paper I investigates quantitative structure-retention relationships from an image processing perspective, using an artificial neural network to preprocess three-dimensional structural descriptions of the studied steroid molecules. Paper II presents a new method for computing free energies. Free energy is the quantity that determines chemical equilibria and partition coefficients. The proposed method may be used for estimating, e.g., chromatographic retention without performing experiments. Two papers (III and IV) deal with correcting deviations from bilinearity by so-called peak alignment. Bilinearity is a theoretical assumption about the distribution of instrumental data that is often violated by measured data. Deviations from bilinearity lead to increased variance, both in the data and in inferences from the data, unless invariance to the deviations is built into the model, e.g., by the use of the method proposed in paper III and extended in paper IV. Paper V addresses a generic problem in classification; namely, how to measure the goodness of different data representations, so that the best classifier may be constructed. Variance reduction is one of the pillars on which analytical chemistry rests. This thesis considers two aspects on variance reduction: before and after experiments are performed. Before experimenting, theoretical predictions of experimental outcomes may be used to direct which experiments to perform, and how to perform them (papers I and II). After experiments are performed, the variance of inferences from the measured data are affected by the method of data analysis (papers III-V).

Progetto di reti Sensori Wireless e tecniche di Fusione Sensoriale

Ambient Intelligence (AmI) envisions a world where smart, electronic environments are aware and responsive to their context. People moving into these settings engage many computational devices and systems simultaneously even if they are not aware of their presence. AmI stems from the convergence of three key technologies: ubiquitous computing, ubiquitous communication and natural interfaces. The dependence on a large amount of fixed and mobile sensors embedded into the environment makes of Wireless Sensor Networks one of the most relevant enabling technologies for AmI. WSN are complex systems made up of a number of sensor nodes, simple devices that typically embed a low power computational unit (microcontrollers, FPGAs etc.), a wireless communication unit, one or more sensors and a some form of energy supply (either batteries or energy scavenger modules). Low-cost, low-computational power, low energy consumption and small size are characteristics that must be taken into consideration when designing and dealing with WSNs. In order to handle the large amount of data generated by a WSN several multi sensor data fusion techniques have been developed. The aim of multisensor data fusion is to combine data to achieve better accuracy and inferences than could be achieved by the use of a single sensor alone. In this dissertation we present our results in building several AmI applications suitable for a WSN implementation. The work can be divided into two main areas: Multimodal Surveillance and Activity Recognition. Novel techniques to handle data from a network of low-cost, low-power Pyroelectric InfraRed (PIR) sensors are presented. Such techniques allow the detection of the number of people moving in the environment, their direction of movement and their position. We discuss how a mesh of PIR sensors can be integrated with a video surveillance system to increase its performance in people tracking. Furthermore we embed a PIR sensor within the design of a Wireless Video Sensor Node (WVSN) to extend its lifetime. Activity recognition is a fundamental block in natural interfaces. A challenging objective is to design an activity recognition system that is able to exploit a redundant but unreliable WSN. We present our activity in building a novel activity recognition architecture for such a dynamic system. The architecture has a hierarchical structure where simple nodes performs gesture classification and a high level meta classifiers fuses a changing number of classifier outputs. We demonstrate the benefit of such architecture in terms of increased recognition performance, and fault and noise robustness. Furthermore we show how we can extend network lifetime by performing a performance-power trade-off. Smart objects can enhance user experience within smart environments. We present our work in extending the capabilities of the Smart Micrel Cube (SMCube), a smart object used as tangible interface within a tangible computing framework, through the development of a gesture recognition algorithm suitable for this limited computational power device. Finally the development of activity recognition techniques can greatly benefit from the availability of shared dataset. We report our experience in building a dataset for activity recognition. Such dataset is freely available to the scientific community for research purposes and can be used as a testbench for developing, testing and comparing different activity recognition techniques.

The genetic signature of Neolithic in Greece

The Neolithic is characterized by the transition from a subsistence economy, based on hunting and gathering, to one based on food producing. This important change was paralleled by one of the most significant demographic increase in the recent history of European populations. The earliest Neolithic sites in Europe are located in Greece. However, the debate regarding the colonization route followed by the Middle-eastern farmers is still open. Based on archaeological, archaeobotanical, craniometric and genetic data, two main hypotheses have been proposed. The first implies the maritime colonization of North-eastern Peloponnesus from Crete, whereas the second points to an island hopping route that finally brought migrants to Central Greece. To test these hypotheses using a genetic approach, 206 samples were collected from the two Greek regions proposed as the arrival point of the two routes (Korinthian district and Euboea). Expectations for each hypothesis were compared with empirical observations based on the analysis of 60 SNPs and 26 microsatellite loci of Y-chromosome and mitochondrial DNA hypervariable region I. The analysis of Y-chromosome haplogroups revealed a strong genetic affinity of Euboea with Anatolian and Middle-eastern populations. The inferences of the time since population expansion suggests an earlier usage of agriculture in Euboea. Moreover, the haplogroup J2a-M410, supposed to be associated with the Neolithic transition, was observed at higher frequency and variance in Euboea showing, for both these parameters, a decreasing gradient moving from this area. The time since expansion estimates for J2a-M410 was found to be compatible with the Neolithic and slightly older in Euboea. The analysis of mtDNA resulted less informative. However, a higher genetic affinity of Euboea with Anatolian and Middle-eastern populations was confirmed. These results taken as a whole suggests that the most probable route followed by Neolithic farmers during the colonization of Greece was the island hopping route.

Stable isotope composition of atmospheric carbon monoxide : a modelling study

Stable isotope composition of atmospheric carbon monoxide: A modelling study.rnrnThis study aims at an improved understanding of the stable carbon and oxygen isotope composition of the carbon monoxide (CO) in the global atmosphere by means of numerical simulations. At first, a new kinetic chemistry tagging technique for the most complete parameterisation of isotope effects has been introduced into the Modular Earth Submodel System (MESSy) framework. Incorporated into the ECHAM/MESSy Atmospheric Chemistry (EMAC) general circulation model, an explicit treatment of the isotope effects on the global scale is now possible. The expanded model system has been applied to simulate the chemical system containing up to five isotopologues of all carbon- and oxygen-bearing species, which ultimately determine the δ13C, δ18O and Δ17O isotopic signatures of atmospheric CO. As model input, a new stable isotope-inclusive emission inventory for the relevant trace gases has been compiled. The uncertainties of the emission estimates and of the resulting simulated mixing and isotope ratios have been analysed. The simulated CO mixing and stable isotope ratios have been compared to in-situ measurements from ground-based observatories and from the civil-aircraft-mounted CARIBIC−1 measurement platform.rnrnThe systematically underestimated 13CO/12CO ratios of earlier, simplified modelling studies can now be partly explained. The EMAC simulations do not support the inferences of those studies, which suggest for CO a reduced input of the highly depleted in 13C methane oxidation source. In particular, a high average yield of 0.94 CO per reacted methane (CH4) molecule is simulated in the troposphere, to a large extent due to the competition between the deposition and convective transport processes affecting the CH4 to CO reaction chain intermediates. None of the other factors, assumed or disregarded in previous studies, however hypothesised to have the potential in enriching tropospheric CO in 13C, were found significant when explicitly simulated. The inaccurate surface emissions, likely underestimated over East Asia, are responsible for roughly half of the discrepancies between the simulated and observed 13CO in the northern hemisphere (NH), whereas the remote southern hemisphere (SH) compositions suggest an underestimated fractionation during the oxidation of CO by the hydroxyl radical (OH). A reanalysis of the kinetic isotope effect (KIE) in this reaction contrasts the conventional assumption of a mere pressure dependence, and instead suggests an additional temperature dependence of the 13C KIE, which is driven by changes in the partitioning of the reaction exit channels. This result is yet to be confirmed in the laboratory.rnrnApart from 13CO, for the first time the atmospheric distribution of the oxygen mass-independent fractionation (MIF) in CO, Δ17O, has been consistently simulated on the global scale with EMAC. The applicability of Δ17O(CO) observations to unravelling changes in the tropospheric CH4-CO-OH system has been scrutinised, as well as the implications of the ozone (O3) input to the CO isotope oxygen budget. The Δ17O(CO) is confirmed to be the principal signal for the CO photochemical age, thus providing a measure for the OH chiefly involved in the sink of CO. The highly mass-independently fractionated O3 oxygen is estimated to comprise around 2% of the overall tropospheric CO source, which has implications for the δ18O, but less likely for the Δ17O CO budgets. Finally, additional sensitivity simulations with EMAC corroborate the nearly equal net effects of the present-day CH4 and CO burdens in removing tropospheric OH, as well as the large turnover and stability of the abundance of the latter. The simulated CO isotopologues nonetheless hint at a likely insufficient OH regeneration in the NH high latitudes and the upper troposphere / lower stratosphere (UTLS).rn

Extraktion und nanoanalytische Charakterisierung refraktärer Nanometallpartikel frühester solarer Materie und Synthese metallischer Nanopartikel aus dotierten Ca-Mg-Al-Si-Schmelzen

Primitive kohlige Chondrite sind Meteorite, die seit ihrer Entstehung im frühen Sonnensystem kaum verändert wurden und dadurch einen Einblick in Prozesse geben, die zur Bildung und Veränderung der ersten festen Materie führten. Solche Prozesse können anhand von Bruchstücken dieser Meteorite detailliert im Labor studiert werden, sodass Rückschlüsse auf die Entwicklung unseres Sonnensystems im frühen Stadium getroffen werden können. Ca-, Al-reiche Einschlüsse (CAIs) aus chondritischen Meteoriten sind die ersten Festkörper des Sonnensystems und enthalten viele refraktäre Metallnuggets (RMNs), welche hauptsächlich aus den Elementen Os, Ir, Ru, Mo und Pt bestehen. Nach weit verbreiteter Ansicht sind diese Nuggets wahrscheinlich im Gleichgewicht mit dem solaren Nebel kondensiert, bereits früher oder gleichzeitig mit Oxiden und Silikaten. Die exakten Mechanismen, die zu ihren heute beobachteten Eigenschaften führten, sind allerdings unklar. Um frühere Arbeiten fortzuführen, wurde eine hohe Anzahl RMNs in vier unterschiedlichen Typen von Meteoriten detailliert studiert, darunter solche aus dem nahezu unveränderten Acfer 094, Allende (CV3ox), Leoville (CV3red) und Murchison (CM2). Die RMNs wurden in-situ, assoziiert mit ihren Wirtsmineralen und auch in Säurerückständen gefunden, deren Präparationsprozedur in dieser Arbeit speziell für RMNs durch eine zusätzliche Dichtetrennung verbessert wurde.rnDie Ergebnisse decken eine Reihe von Ungereimtheiten zwischen den beobachteten RMN-Eigenschaften und einer Kondensationsherkunft auf, sowohl für Kondensation in solarer Umgebung, als auch für Kondensation aus Material von Supernovae oder roten Riesen, für die die Kondensationssequenzen refraktärer Metalle speziell für diesen Vergleich berechnet wurden. Stattdessen wurden in dieser Arbeit neue Einblicke in die RMN-Entstehung und die Entwicklung der ersten Festkörper (CAIs) durch eine Kombination aus experimentellen, isotopischen, strukturellen und petrologischen Studien an RMNs gewonnen. Viele der beobachteten Eigenschaften sind mit Ausfällung der RMN aus einer CAI-Schmelze vereinbar. Ein solches Szenario wird durch entsprechende Untersuchungen an synthetisch hergestellten, mit refraktären Metallen im Gleichgewicht stehenden CAI-Schmelzen bestätigt. Es folgt aus den Ergebnissen, dass die Mehrzahl der RMNs isotopisch solar ist und alle untersuchten RMNs innerhalb von CAIs bei rascher Abkühlung (um bis zu 1000 °C/40 sek.) einer CAI-Schmelze gebildet wurden. rn

Blood aspiration as a vital sign detected by postmortem computed tomography imaging

Blood aspiration is a significant forensic finding. In this study, we examined the value of postmortem computed tomography (CT) imaging in evaluating findings of blood aspiration. We selected 37 cases with autopsy evidence of blood in the lungs and/or in the airways previously submitted to total-body CT scanning. The CT-images were retrospectively analyzed. In one case with pulmonary blood aspiration, biopsy specimens were obtained under CT guide for histological examination. In six cases, CT detected pulmonary abnormalities suggestive of blood aspiration, not mentioned in the autopsy reports. CT reconstructions provided additional data about the distribution and extent of aspiration. In one needle-biopsied case, the pulmonary specimens showed blood in the alveoli. We suggest the use of CT imaging as a tool complementary to traditional techniques in cases of blood aspiration to avoid misdiagnosis, to guide the investigation of lung tissue, and to allow for more evidence-based inferences on the cause of death.

Urothelial neoplasms of the urinary bladder occurring in young adult and pediatric patients: a comprehensive review of literature with implications for patient management

Bladder urothelial carcinoma is typically a disease of older individuals and rarely occurs below the age of 40 years. There is debate and uncertainty in the literature regarding the clinicopathologic characteristics of bladder urothelial neoplasms in younger patients compared with older patients, although no consistent age criteria have been used to define "younger" age group categories. Use of the World Health Organization 2004/International Society of Urological Pathology 1998 grading nomenclature and recent molecular studies highlight certain unique features of bladder urothelial neoplasms in young patients, particularly in patients below 20 years of age. In this meta-analysis and review, the clinical, pathologic, and molecular features and risk factors of bladder urothelial neoplasms in patients 40 years or less are presented and analyzed according to decades of presentation. Similar to older patients, bladder urothelial neoplasms in patients 40 years or younger occur more common in male patients, present mainly with gross painless hematuria, and are more commonly located at bladder trigone/ureteral orifices, but in contrast have a greater chance for unifocality. Delay in diagnosis of bladder urothelial neoplasms seems not to be uncommon in younger patients probably because of its relative rarity and the predominance of benign causes of hematuria in this age group causing hesitancy for an aggressive work-up. Most tumors in patients younger than 40 years were low grade. The incidence of low-grade tumors was the lowest in the first 2 decades of life, with incremental increase of the percentage of high-grade tumors with increasing age decades. Classification according to the World Health Organization 2004/International Society of Urological Pathology grading system identified papillary urothelial neoplasms of low malignant potential to be relatively frequent among bladder tumors of young patients particularly in the teenage years. Similar to grade, there was marked predominance of low stage tumors in the first 2 decades of life with gradual inclusion of few higher stage and metastatic tumors in the 2 older decades. Bladder urothelial neoplasms occurring in patients <20 years of age lack or have a much lower incidence of aberrations in chromosome 9, FGFR3, p53, and microsatellite instability and have fewer epigenetic alterations. Tumor recurrence and deaths were infrequent in the first 2 decades and increased gradually in each successive decade, likely influenced by the increased proportion of higher grade and higher stage tumors. Our review of the literature shows that urothelial neoplasms of the bladder occurring in young patients exhibit unique pathologic and molecular features that translate to its more indolent behavior; this distinction is most pronounced in patients <20 years. Our overall inferences have potential implications for choosing appropriate noninvasive diagnostic and surveillance modalities, whenever feasible, and for selecting suitable treatment strategies that factor in quality of life issues vital to younger patients.

The Spatial Signature of Biotic Interactions of a Clonal and a Non-clonal Palmetto in a Subtropical Plant Community

Spatial analyses of plant-distribution patterns can provide inferences about intra- and interspecific biotic interactions. Yet, such analyses are rare for clonal plants because effective tools (i.e., molecular markers) needed to map naturally occurring clonal individuals have only become available recently. Clonal plants are unique in that a single genotype has a potential to spatially place new individuals (i.e., ramets) in response to intra- and interspecific biotic interactions. Laboratory and greenhouse studies suggest that some clonal plants can avoid intra-genet, inter-genet, and inter-specific competition via rootplacement patterns. An intriguing and yet to be explored question is whether a spatial signature of such multi-level biotic interactions can be detected in natural plant communities. The facultatively clonal Serenoa repens and non-clonal Sabal etonia are ecologically similar and co-dominant palmettos that sympatrically occur in the Florida peninsula. We used amplified fragment length polymorphisms (AFLPs) to identify Serenoa genets and also to assign field-unidentifiable small individuals as Sabal seedlings, Serenoa seedlings, or Serenoa vegetative sprouts. Then, we conducted univariate and bivariate multi-distance spatial analyses to examine the spatial interactions of Serenoa (n=271) and Sabal (n=137) within a 20x20 m grid at three levels, intragenet, intergenet and interspecific. We found that spatial interactions were not random at all three levels of biotic interactions. Serenoa genets appear to spatially avoid self-competition as well as intergenet competition. Furthermore, Serenoa and Sabal were spatially negatively associated with each other. However, this negative association pattern was also evident in a spatial comparison between non-clonal Serenoa and Sabal, suggesting that Serenoa genets’ spatial avoidance of Sabal through placement of new ramets is not the explanation of the interspecific-level negative spatial pattern. Our results emphasize the importance of investigating spatial signatures of biotic as well as abiotic interactions at multiple levels in understanding spatial distribution patterns of clonal plants in natural plant communities.

MR imaging in Crohn's disease: correlation of MR motility measurement with histopathology in the terminal ileum

BACKGROUND: The objective of the study was to correlate MR-detectable motility alterations of the terminal ileum with biopsy-documented active and chronic changes in Crohn's disease. METHODS: This IRB approved retrospective analysis of 43 patients included magnetic resonance enterography (MRE) and terminal ileum biopsies (<2 weeks apart). Motility was measured at the terminal ileum using coronal 2D trueFISP pulse sequences (1.5T MRI,TR 83.8,TE1.89) and dedicated motility assessment software. Motility grading (hypermotility, normal, hypomotility, complete arrest) was agreed by two experienced readers. Motility was compared and correlated with histopathology using two-tailed Kruskal-Wallis test and paired Spearman Rank-Order Correlation tests. KEY RESULTS: Motility abnormalities were present in 27/43 patients: nine hypomotility and 18 complete arrest. Active disease was diagnosed on 15 biopsies: eight moderate and seven severe inflammatory activity. Chronic changes were diagnosed on 17 biopsies: 13 moderate and four severe cases. In four patients with normal motility alterations on histopathology were diagnosed. Histopathology correlated with presence (P = 0.0056 for hypomotility and P = 0.0119 for complete arrest) and grade (P < 0.0001; P = 0.0004) of motility alterations. A significant difference in the motility was observed in patients with active or chronic CD compared with patients without disease (P < 0.001; P = 0.0024). CONCLUSIONS & INFERENCES: MR-detectable motility changes of the terminal ileum correlate with histopathological findings both in active and chronic CD. Motility changes may indicate the presence pathology, but do not allow differentiation of active and chronic disease.

Comparative Ultrastructure and Molecular Phylogeny of Selenidium melongena n. sp and S. terebellae Ray 1930 Demonstrate Niche Partitioning in Marine Gregarine Parasites (Apicomplexa)

Gregarine apicomplexans are a diverse group of single-celled parasites that have feeding stages (trophozoites) and gamonts that generally inhabit the extracellular spaces of invertebrate hosts living in marine, freshwater, and terrestrial environments. Inferences about the evolutionary morphology of gregarine apicomplexans are being incrementally refined by molecular phylogenetic data, which suggest that several traits associated with the feeding cells of gregarines arose by convergent evolution. The study reported here supports these inferences by showing how molecular data reveals traits that are phylogenetically misleading within the context of comparative morphology alone. We examined the ultrastructure and molecular phylogenetic positions of two gregarine species isolated from the spaghetti worm Thelepus japonicus: Selenidium terebellae Ray 1930 and S. melongena n. sp. The ultrastructural traits of S. terebellae were very similar to other species of Selenidium sensu stricto, such as having vermiform trophozoites with an apical complex, few epicytic folds, and a dense array of microtubules underlying the trilayered pellicle. By contrast, S. melongena n. sp. lacked a comparably discrete assembly of subpellicular microtubules, instead employing a system of fibrils beneath the cell surface that supported a relatively dense array of helically arranged epicytic folds. Molecular phylogenetic analyses of small subunit rDNA sequences derived from single-cell PCR unexpectedly demonstrated that these two gregarines are close sister species. The ultrastructural differences between these two species were consistent with the fact that S. terebellae infects the inner lining of the host intestines, and S. melongena n. sp. primarily inhabits the coelom, infecting the outside wall of the host intestine. Altogether, these data demonstrate a compelling case of niche partitioning and associated morphological divergence in marine gregarine apicomplexans. (C) 2014 Elsevier GmbH. All rights reserved.

Mechanical Properties of Plant Underground Storage Organs and Implications for Dietary Models of Early Hominins

The diet of early human ancestors has received renewed theoretical interest since the discovery of elevated d13C values in the enamel of Australopithecus africanus and Paranthropus robustus. As a result, the hominin diet is hypothesized to have included C4 grass or the tissues of animals which themselves consumed C4 grass. On mechanical grounds, such a diet is incompatible with the dental morphology and dental microwear of early hominins. Most inferences, particularly for Paranthropus, favor a diet of hard or mechanically resistant foods. This discrepancy has invigorated the longstanding hypothesis that hominins consumed plant underground storage organs (USOs). Plant USOs are attractive candidate foods because many bulbous grasses and cormous sedges use C4 photosynthesis. Yet mechanical data for USOs—or any putative hominin food—are scarcely known. To fill this empirical void we measured the mechanical properties of USOs from 98 plant species from across sub-Saharan Africa. We found that rhizomes were the most resistant to deformation and fracture, followed by tubers, corms, and bulbs. An important result of this study is that corms exhibited low toughness values (mean = 265.0 J m-2) and relatively high Young’s modulus values (mean = 4.9 MPa). This combination of properties fits many descriptions of the hominin diet as consisting of hard-brittle objects. When compared to corms, bulbs are tougher (mean = 325.0 J m-2) and less stiff (mean = 2.5 MPa). Again, this combination of traits resembles dietary inferences, especially for Australopithecus, which is predicted to have consumed soft-tough foods. Lastly, we observed the roasting behavior of Hadza hunter-gatherers and measured the effects of roasting on the toughness on undomesticated tubers. Our results support assumptions that roasting lessens the work of mastication, and, by inference, the cost of digestion. Together these findings provide the first mechanical basis for discussing the adaptive advantages of roasting tubers and the plausibility of USOs in the diet of early hominins.

A Cox Model for Biostatistics of the Future

Professor Sir David R. Cox (DRC) is widely acknowledged as among the most important scientists of the second half of the twentieth century. He inherited the mantle of statistical science from Pearson and Fisher, advanced their ideas, and translated statistical theory into practice so as to forever change the application of statistics in many fields, but especially biology and medicine. The logistic and proportional hazards models he substantially developed, are arguably among the most influential biostatistical methods in current practice. This paper looks forward over the period from DRC's 80th to 90th birthdays, to speculate about the future of biostatistics, drawing lessons from DRC's contributions along the way. We consider "Cox's model" of biostatistics, an approach to statistical science that: formulates scientific questions or quantities in terms of parameters gamma in probability models f(y; gamma) that represent in a parsimonious fashion, the underlying scientific mechanisms (Cox, 1997); partition the parameters gamma = theta, eta into a subset of interest theta and other "nuisance parameters" eta necessary to complete the probability distribution (Cox and Hinkley, 1974); develops methods of inference about the scientific quantities that depend as little as possible upon the nuisance parameters (Barndorff-Nielsen and Cox, 1989); and thinks critically about the appropriate conditional distribution on which to base infrences. We briefly review exciting biomedical and public health challenges that are capable of driving statistical developments in the next decade. We discuss the statistical models and model-based inferences central to the CM approach, contrasting them with computationally-intensive strategies for prediction and inference advocated by Breiman and others (e.g. Breiman, 2001) and to more traditional design-based methods of inference (Fisher, 1935). We discuss the hierarchical (multi-level) model as an example of the future challanges and opportunities for model-based inference. We then consider the role of conditional inference, a second key element of the CM. Recent examples from genetics are used to illustrate these ideas. Finally, the paper examines causal inference and statistical computing, two other topics we believe will be central to biostatistics research and practice in the coming decade. Throughout the paper, we attempt to indicate how DRC's work and the "Cox Model" have set a standard of excellence to which all can aspire in the future.

Bayesian Hidden Markov Modeling of Array CGH Data

Genomic alterations have been linked to the development and progression of cancer. The technique of Comparative Genomic Hybridization (CGH) yields data consisting of fluorescence intensity ratios of test and reference DNA samples. The intensity ratios provide information about the number of copies in DNA. Practical issues such as the contamination of tumor cells in tissue specimens and normalization errors necessitate the use of statistics for learning about the genomic alterations from array-CGH data. As increasing amounts of array CGH data become available, there is a growing need for automated algorithms for characterizing genomic profiles. Specifically, there is a need for algorithms that can identify gains and losses in the number of copies based on statistical considerations, rather than merely detect trends in the data. We adopt a Bayesian approach, relying on the hidden Markov model to account for the inherent dependence in the intensity ratios. Posterior inferences are made about gains and losses in copy number. Localized amplifications (associated with oncogene mutations) and deletions (associated with mutations of tumor suppressors) are identified using posterior probabilities. Global trends such as extended regions of altered copy number are detected. Since the posterior distribution is analytically intractable, we implement a Metropolis-within-Gibbs algorithm for efficient simulation-based inference. Publicly available data on pancreatic adenocarcinoma, glioblastoma multiforme and breast cancer are analyzed, and comparisons are made with some widely-used algorithms to illustrate the reliability and success of the technique.

Semiparametric Theory for Causal Mediation Analysis: efficiency bounds, multiple robustness, and sensitivity analysis

Whilst estimation of the marginal (total) causal effect of a point exposure on an outcome is arguably the most common objective of experimental and observational studies in the health and social sciences, in recent years, investigators have also become increasingly interested in mediation analysis. Specifically, upon establishing a non-null total effect of the exposure, investigators routinely wish to make inferences about the direct (indirect) pathway of the effect of the exposure not through (through) a mediator variable that occurs subsequently to the exposure and prior to the outcome. Although powerful semiparametric methodologies have been developed to analyze observational studies, that produce double robust and highly efficient estimates of the marginal total causal effect, similar methods for mediation analysis are currently lacking. Thus, this paper develops a general semiparametric framework for obtaining inferences about so-called marginal natural direct and indirect causal effects, while appropriately accounting for a large number of pre-exposure confounding factors for the exposure and the mediator variables. Our analytic framework is particularly appealing, because it gives new insights on issues of efficiency and robustness in the context of mediation analysis. In particular, we propose new multiply robust locally efficient estimators of the marginal natural indirect and direct causal effects, and develop a novel double robust sensitivity analysis framework for the assumption of ignorability of the mediator variable.

Semiparametric Estimation of Models for Natural Direct and Indirect Effects

In recent years, researchers in the health and social sciences have become increasingly interested in mediation analysis. Specifically, upon establishing a non-null total effect of an exposure, investigators routinely wish to make inferences about the direct (indirect) pathway of the effect of the exposure not through (through) a mediator variable that occurs subsequently to the exposure and prior to the outcome. Natural direct and indirect effects are of particular interest as they generally combine to produce the total effect of the exposure and therefore provide insight on the mechanism by which it operates to produce the outcome. A semiparametric theory has recently been proposed to make inferences about marginal mean natural direct and indirect effects in observational studies (Tchetgen Tchetgen and Shpitser, 2011), which delivers multiply robust locally efficient estimators of the marginal direct and indirect effects, and thus generalizes previous results for total effects to the mediation setting. In this paper we extend the new theory to handle a setting in which a parametric model for the natural direct (indirect) effect within levels of pre-exposure variables is specified and the model for the observed data likelihood is otherwise unrestricted. We show that estimation is generally not feasible in this model because of the curse of dimensionality associated with the required estimation of auxiliary conditional densities or expectations, given high-dimensional covariates. We thus consider multiply robust estimation and propose a more general model which assumes a subset but not all of several working models holds.