A comparison of Markov and generalized linear models of hospital mortality

This paper reports a comparison of three modeling strategies for the analysis of hospital mortality in a sample of general medicine inpatients in a Department of Veterans Affairs medical center. Logistic regression, a Markov chain model, and longitudinal logistic regression were evaluated on predictive performance as measured by the c-index and on accuracy of expected numbers of deaths compared to observed. The logistic regression used patient information collected at admission; the Markov model was comprised of two absorbing states for discharge and death and three transient states reflecting increasing severity of illness as measured by laboratory data collected during the hospital stay; longitudinal regression employed Generalized Estimating Equations (GEE) to model covariance structure for the repeated binary outcome. Results showed that the logistic regression predicted hospital mortality as well as the alternative methods but was limited in scope of application. The Markov chain provides insights into how day to day changes of illness severity lead to discharge or death. The longitudinal logistic regression showed that increasing illness trajectory is associated with hospital mortality. The conclusion is reached that for standard applications in modeling hospital mortality, logistic regression is adequate, but for new challenges facing health services research today, alternative methods are equally predictive, practical, and can provide new insights. ^

A re-examination of petrogenesis and 40Ar/39Ar systematics in the the Chain of Ponds K-feldspar: "diffusion domain" archetype versus polyphase hygrochronology

K-feldspar (Kfs) from the Chain of Ponds Pluton (CPP) is the archetypal reference material, on which thermochronological modeling of Ar diffusion in discrete “domains” was founded. We re-examine the CPP Kfs using cathodoluminescence and back-scattered electron imaging, transmission electron microscopy, and electron probe microanalysis. 40Ar/39Ar stepwise heating experiments on different sieve fractions, and on handpicked and unpicked aliquots, are compared. Our results reproduce the staircase-shaped age spectrum and the Arrhenius trajectory of the literature sample, confirming that samples collected from the same locality have an identical Ar isotope record. Even the most pristine-looking Kfs from the CPP contains successive generations of secondary, metasomatic/retrograde mineral replacements that post-date magmatic crystallization. These chemically and chronologically distinct phases are responsible for its staircase-shaped age spectra, which are modified by handpicking. While genuine within-grain diffusion gradients are not ruled out by these data, this study demonstrates that the most important control on staircase-shaped age spectra is the simultaneous presence of heterochemical, diachronous post-magmatic mineral growth. At least five distinct mineral species were identified in the Kfs separate, three of which can be traced to external fluids interacting with the CPP in a chemically open system. Sieve fractions have size-shifted Arrhenius trajectories, negating the existence of the smallest “diffusion domains”. Heterochemical phases also play an important role in producing non-linear trajectories. In vacuo degassing rates recovered from Arrhenius plots are neither related to true Fick’s Law diffusion nor to the staircase shape of the age spectra. The CPP Kfs used to define the "diffusion domain" model demonstrates the predominance of metasomatic alteration by hydrothermal fluids and recrystallization in establishing the natural Ar distribution amongst different coexisting phases that gives rise to the staircase-shaped age spectrum. Microbeam imaging of textures is as essential for 40Ar-39Ar hygrochronology as it is for U-Pb geochronology.

Bayesian generalized linear models for meta-analysis of diagnostic tests

With the recognition of the importance of evidence-based medicine, there is an emerging need for methods to systematically synthesize available data. Specifically, methods to provide accurate estimates of test characteristics for diagnostic tests are needed to help physicians make better clinical decisions. To provide more flexible approaches for meta-analysis of diagnostic tests, we developed three Bayesian generalized linear models. Two of these models, a bivariate normal and a binomial model, analyzed pairs of sensitivity and specificity values while incorporating the correlation between these two outcome variables. Noninformative independent uniform priors were used for the variance of sensitivity, specificity and correlation. We also applied an inverse Wishart prior to check the sensitivity of the results. The third model was a multinomial model where the test results were modeled as multinomial random variables. All three models can include specific imaging techniques as covariates in order to compare performance. Vague normal priors were assigned to the coefficients of the covariates. The computations were carried out using the 'Bayesian inference using Gibbs sampling' implementation of Markov chain Monte Carlo techniques. We investigated the properties of the three proposed models through extensive simulation studies. We also applied these models to a previously published meta-analysis dataset on cervical cancer as well as to an unpublished melanoma dataset. In general, our findings show that the point estimates of sensitivity and specificity were consistent among Bayesian and frequentist bivariate normal and binomial models. However, in the simulation studies, the estimates of the correlation coefficient from Bayesian bivariate models are not as good as those obtained from frequentist estimation regardless of which prior distribution was used for the covariance matrix. The Bayesian multinomial model consistently underestimated the sensitivity and specificity regardless of the sample size and correlation coefficient. In conclusion, the Bayesian bivariate binomial model provides the most flexible framework for future applications because of its following strengths: (1) it facilitates direct comparison between different tests; (2) it captures the variability in both sensitivity and specificity simultaneously as well as the intercorrelation between the two; and (3) it can be directly applied to sparse data without ad hoc correction. ^

New methods for quantification and analysis of quantitative real-time polymerase chain reaction data

Quantitative real-time polymerase chain reaction (qPCR) is a sensitive gene quantitation method that has been widely used in the biological and biomedical fields. The currently used methods for PCR data analysis, including the threshold cycle (CT) method, linear and non-linear model fitting methods, all require subtracting background fluorescence. However, the removal of background fluorescence is usually inaccurate, and therefore can distort results. Here, we propose a new method, the taking-difference linear regression method, to overcome this limitation. Briefly, for each two consecutive PCR cycles, we subtracted the fluorescence in the former cycle from that in the later cycle, transforming the n cycle raw data into n-1 cycle data. Then linear regression was applied to the natural logarithm of the transformed data. Finally, amplification efficiencies and the initial DNA molecular numbers were calculated for each PCR run. To evaluate this new method, we compared it in terms of accuracy and precision with the original linear regression method with three background corrections, being the mean of cycles 1-3, the mean of cycles 3-7, and the minimum. Three criteria, including threshold identification, max R2, and max slope, were employed to search for target data points. Considering that PCR data are time series data, we also applied linear mixed models. Collectively, when the threshold identification criterion was applied and when the linear mixed model was adopted, the taking-difference linear regression method was superior as it gave an accurate estimation of initial DNA amount and a reasonable estimation of PCR amplification efficiencies. When the criteria of max R2 and max slope were used, the original linear regression method gave an accurate estimation of initial DNA amount. Overall, the taking-difference linear regression method avoids the error in subtracting an unknown background and thus it is theoretically more accurate and reliable. This method is easy to perform and the taking-difference strategy can be extended to all current methods for qPCR data analysis.^

Long chain 1,13- and 1,15-diols in surface sediments ans SST reconstruction for sediment core GeoB6518-1

Although commonly reported in marine and freshwater environments, little is known about the biological sources of long chain alkyl 1,13- and 1,15-diols, and factors controlling their distributions. Here we analyzed the occurrence and distribution of these lipids in a comprehensive set of marine surface sediments and compare their distributions with environmental conditions like sea surface temperature (SST), salinity and nutrient concentrations. Fractional abundances of the C28 1,13-, C30 1,13- and C30 1,15-diols show a strong correlation with SST and based on these results, we propose the Long chain Diol Index (LDI), which expresses the C30 1,15-diol abundance relative to those of C28 1,13-, C30 1,13- and C30 1,15-diols. The LDI shows a strong linear correlation with SST (LDI = 0.033 × SST + 0.095; R2 = 0.969, n = 162) over a temperature range of -3 to 27 °C. Long chain diol distributions in sediments from the South Atlantic close to the Congo River outflow (West Africa) provided a 43 kyr LDI SST record. This record reflects several known climatic events and shows similarities with an alkenone-derived SST record obtained using the same suite of sediments, both in trend and in terms of absolute SST. This confirms the potential of the LDI as a proxy for palaeo-SST reconstruction.

Sources and proxy potential of long chain alkyl diols in lacustrine environments

Long chain 1,13- and 1,15-alkyl diols form the base of a number of recently proposed proxies used for climate reconstruction. However, the sources of these lipids and environmental controls on their distribution are still poorly constrained. We have analyzed the long chain alkyl diol (LCD) composition of cultures of ten eustigmatophyte species, with three species from different families grown at various temperatures, to identify the effect of species composition and growth temperature on the LCD distribution. The results were compared with the LCD distribution of sixty-two lake surface sediments, and with previously reported LCD distributions from marine environments. The different families within the Eustigmatophyceae show distinct LCD patterns, with the freshwater family Eustigmataceae most closely resembling LCD distributions in both marine and lake environments. Unlike the other two eustigmatophyte families analyzed (Monodopsidaceae and Goniochloridaceae), C28 and C30 1,13-alkyl diols and C30 and C32 1,15-alkyl diols are all relatively abundant in the family Eustigmataceae, while the mono-unsaturated C32 1,15-alkyl diol was below detection limit. In contrast to the marine environment, LCD distributions in lakes did not show a clear relationship with temperature. The Long chain Diol Index (LDI), a proxy previously proposed for sea surface temperature reconstruction, showed a relatively weak correlation (R2 = 0.33) with mean annual air temperature used as an approximation for annual mean surface temperature of the lakes. A much-improved correlation (R2 = 0.74, p-value<0.001) was observed applying a multiple linear regression analysis between LCD distributions and lake temperatures reconstructed using branched tetraether lipid distributions. The obtained regression model provides good estimates of temperatures for cultures of the family Eustigmataceae, suggesting that algae belonging to this family have an important role as a source for LCDs in lacustrine environments, or, alternatively, that the main sources of LCDs are similarly affected by temperature as the Eustigmataceae. The results suggest that LCDs may have the potential to be applicable as a palaeotemperature proxy for lacustrine environments, although further calibration work is still required.

Three functional luciferase domains in a single polypeptide chain

We report a unique case of a gene containing three homologous and contiguous repeat sequences, each of which, after excision, cloning, and expression in Escherichia coli, is shown to code for a peptide catalyzing the same reaction as the native protein, Gonyaulax polyedra luciferase (Mr = 137). This enzyme, which catalyzes the light-emitting oxidation of a linear tetrapyrrole (dinoflagellate luciferin), exhibits no sequence similarities to other luciferases in databases. Sequence analysis also reveals an unusual evolutionary feature of this gene: synonymous substitutions are strongly constrained in the central regions of each of the repeated coding sequences.

The α chain of laminin-1 is independently secreted and drives secretion of its β- and γ-chain partners

A mammalian recombinant strategy was established to dissect rules of basement membrane laminin assembly and secretion. The α-, β-, and γ-chain subunits of laminin-1 were expressed in all combinations, transiently and/or stably, in a near-null background. In the absence of its normal partners, the α chain was secreted as intact protein and protein that had been cleaved in the coiled-coil domain. In contrast, the β and γ chains, expressed separately or together, remained intracellular with formation of ββ or βγ, but not γγ, disulfide-linked dimers. Secretion of the β and γ chains required simultaneous expression of all three chains and their assembly into αβγ heterotrimers. Epitope-tagged recombinant α subunit and recombinant laminin were affinity-purified from the conditioned medium of αγ and αβγ clones. Rotary-shadow electron microscopy revealed that the free α subunit is a linear structure containing N-terminal and included globules with a foreshortened long arm, while the trimeric species has the typical four-arm morphology of native laminin. We conclude that the α chain can be delivered to the extracellular environment as a single subunit, whereas the β and γ chains cannot, and that the α chain drives the secretion of the trimeric molecule. Such an α-chain-dependent mechanism could allow for the regulation of laminin export into a nascent basement membrane, and might serve an important role in controlling basement membrane formation.

Quantitation of Pseudomonas aeruginosa in wound biopsy samples: from bacterial culture to rapid `real-time' polymerase chain reaction

We developed a real-time detection (RTD) polymerase chain reaction (PCR) with rapid thermal cycling to detect and quantify Pseudomonas aeruginosa in wound biopsy samples. This method produced a linear quantitative detection range of 7 logs, with a lower detection limit of 103 colony-forming units (CFU)/g tissue or a few copies per reaction. The time from sample collection to result was less than 1h. RTD-PCR has potential for rapid quantitative detection of pathogens in critical care patients, enabling early and individualized treatment.

Handling of Uncertainty in Life Cycle Inventory by Correlated Multivariate Lognormal Distributions: Application to the Design of Supply Chain Networks

Poster presented in the 24th European Symposium on Computer Aided Process Engineering (ESCAPE 24), Budapest, Hungary, June 15-18, 2014.

Handling of Uncertainty in Life Cycle Inventory by Correlated Multivariate Lognormal Distributions: Application to the Design of Supply Chain Networks

In this work, we analyze the effect of incorporating life cycle inventory (LCI) uncertainty on the multi-objective optimization of chemical supply chains (SC) considering simultaneously their economic and environmental performance. To this end, we present a stochastic multi-scenario mixed-integer linear programming (MILP) coupled with a two-step transformation scenario generation algorithm with the unique feature of providing scenarios where the LCI random variables are correlated and each one of them has the desired lognormal marginal distribution. The environmental performance is quantified following life cycle assessment (LCA) principles, which are represented in the model formulation through standard algebraic equations. The capabilities of our approach are illustrated through a case study of a petrochemical supply chain. We show that the stochastic solution improves the economic performance of the SC in comparison with the deterministic one at any level of the environmental impact, and moreover the correlation among environmental burdens provides more realistic scenarios for the decision making process.

Controlled Radical Polymerization of Styrene and Methyl Acrylate in the Presence of Reversible Addition-Fragmentation Chain Transfer Agents, Phenylethyl Phenyl Dithioacetate and Phenyldithioacetic Acid

The use of phenyldithioacetic acid (PDA) in homopolymerizations of styrene or methyl acrylate produced only a small fraction of chains with dithioester end groups. The polymerizations using 1-phenylentyl phenyldithioacetate (PEPDTA) and PDA in the same reaction showed that PDA had little or no influence on the rate or molecular weight distribution even when a 1:1 ratio is used. The mechanistic pathway for the polymerizations in the presence of PDA seemed to be different for each monomer. Styrene favors addition of styrene to PDA via a Markovnikov type addition to form a reactive RAFT agent. The polymer was shown by double detection SEC to contain dithioester end groups over the whole distribution. This polymer was then used in a chain extension experiment and the M-n was close to theory. A unique feature of this work was that PDA could be used to form a RAFT agent in situ by heating a mixture of styrene and PDA for 24 h at 70 degrees C and then polymerizing in the presence of AIBN to give a linear increase in Mn and low values of PDI (< 1.14). In the case of the polymerization of MA with PDA, the mechanism was proposed to be via degradative chain transfer. (c) 2005 Wiley Periodicals, Inc.

Constriction flows of monodisperse linear entangled polymers: Multiscale modeling and flow visualization

We explore both the rheology and complex flow behavior of monodisperse polymer melts. Adequate quantities of monodisperse polymer were synthesized in order that both the materials rheology and microprocessing behavior could be established. In parallel, we employ a molecular theory for the polymer rheology that is suitable for comparison with experimental rheometric data and numerical simulation for microprocessing flows. The model is capable of matching both shear and extensional data with minimal parameter fitting. Experimental data for the processing behavior of monodisperse polymers are presented for the first time as flow birefringence and pressure difference data obtained using a Multipass Rheometer with an 11:1 constriction entry and exit flow. Matching of experimental processing data was obtained using the constitutive equation with the Lagrangian numerical solver, FLOWSOLVE. The results show the direct coupling between molecular constitutive response and macroscopic processing behavior, and differentiate flow effects that arise separately from orientation and stretch. (c) 2005 The Society of Rheology.

Synthesis of monocyclic and linear polystyrene using the reversible coupling/cleavage of thiol/disulfide groups

By carefully controlling the concentration of alpha,omega-thiol polystyrene in solution, we achieved formation of unique monocyclic polystyrene chains (i.e., polymer chains with only one disulfide linkage). The presence of cyclic polystyrene was confirmed by its lower than expected molecular weight due to a lower hydrodynamic volume and loss of thiol groups as detected by using Ellman's reagent. The alpha,omega-thiol polystyrene was synthesized by polymerizing styrene in the presence of a difunctional RAFT agent and subsequent conversion of the dithioester end groups to thiols via the addition of hexylamine. Oxidation gave either monocyclic polymer chains (i.e., with only one disulfide linkage) or linear multiblock polymers with many disulfide linkages depending on the concentration of polymer used with greater chance of cyclization in more dilute solutions. At high polymer concentrations, linear multiblock polymers were formed. To control the MWD of these linear multiblocks, monofunctional X-PSTY (X = PhCH2C(S)-S-) was added. It was found that the greatest ratio of X-PSTY to X-PSTY-X resulted in a low M-n and PDI. We have shown that we can control both the structure and MWD using this chemistry, but more importantly such disulfide linkages can be readily reduced back to the starting polystyrene with thiol end groups, which has potential use for a recyclable polymer material.

Measuring and predicting the dynamics of linear monodisperse entangled polymers in rapid flow through an abrupt contraction. A small angle neutron scattering study

Small-angle neutron scattering measurements on a series of monodisperse linear entangled polystyrene melts in nonlinear flow through an abrupt 4:1 contraction have been made. Clear signatures of melt deformation and subsequent relaxation can be observed in the scattering patterns, which were taken along the centerline. These data are compared with the predictions of a recently derived molecular theory. Two levels of molecular theory are used: a detailed equation describing the evolution of molecular structure over all length scales relevant to the scattering data and a simplified version of the model, which is suitable for finite element computations. The velocity field for the complex melt flow is computed using the simplified model and scattering predictions are made by feeding these flow histories into the detailed model. The modeling quantitatively captures the full scattering intensity patterns over a broad range of data with independent variation of position within the contraction geometry, bulk flow rate and melt molecular weight. The study provides a strong, quantitative validation of current theoretical ideas concerning the microscopic dynamics of entangled polymers which builds upon existing comparisons with nonlinear mechanical stress data. Furthermore, we are able to confirm the appreciable length scale dependence of relaxation in polymer melts and highlight some wider implications of this phenomenon.