Modeling the forensic two-trace problem with Bayesian networks

The forensic two-trace problem is a perplexing inference problem introduced by Evett (J Forensic Sci Soc 27:375-381, 1987). Different possible ways of wording the competing pair of propositions (i.e., one proposition advanced by the prosecution and one proposition advanced by the defence) led to different quantifications of the value of the evidence (Meester and Sjerps in Biometrics 59:727-732, 2003). Here, we re-examine this scenario with the aim of clarifying the interrelationships that exist between the different solutions, and in this way, produce a global vision of the problem. We propose to investigate the different expressions for evaluating the value of the evidence by using a graphical approach, i.e. Bayesian networks, to model the rationale behind each of the proposed solutions and the assumptions made on the unknown parameters in this problem.

Further insights on the French WISC-IV factor structure through Bayesian structural equation modeling

The interpretation of the Wechsler Intelligence Scale for Children-Fourth Edition (WISC-IV) is based on a 4-factor model, which is only partially compatible with the mainstream Cattell-Horn-Carroll (CHC) model of intelligence measurement. The structure of cognitive batteries is frequently analyzed via exploratory factor analysis and/or confirmatory factor analysis. With classical confirmatory factor analysis, almost all crossloadings between latent variables and measures are fixed to zero in order to allow the model to be identified. However, inappropriate zero cross-loadings can contribute to poor model fit, distorted factors, and biased factor correlations; most important, they do not necessarily faithfully reflect theory. To deal with these methodological and theoretical limitations, we used a new statistical approach, Bayesian structural equation modeling (BSEM), among a sample of 249 French-speaking Swiss children (8-12 years). With BSEM, zero-fixed cross-loadings between latent variables and measures are replaced by approximate zeros, based on informative, small-variance priors. Results indicated that a direct hierarchical CHC-based model with 5 factors plus a general intelligence factor better represented the structure of the WISC-IV than did the 4-factor structure and the higher order models. Because a direct hierarchical CHC model was more adequate, it was concluded that the general factor should be considered as a breadth rather than a superordinate factor. Because it was possible for us to estimate the influence of each of the latent variables on the 15 subtest scores, BSEM allowed improvement of the understanding of the structure of intelligence tests and the clinical interpretation of the subtest scores.

Resurgence of HIV infection among men who have sex with men in Switzerland: mathematical modelling study.

BACKGROUND: New HIV infections in men who have sex with men (MSM) have increased in Switzerland since 2000 despite combination antiretroviral therapy (cART). The objectives of this mathematical modelling study were: to describe the dynamics of the HIV epidemic in MSM in Switzerland using national data; to explore the effects of hypothetical prevention scenarios; and to conduct a multivariate sensitivity analysis. METHODOLOGY/PRINCIPAL FINDINGS: The model describes HIV transmission, progression and the effects of cART using differential equations. The model was fitted to Swiss HIV and AIDS surveillance data and twelve unknown parameters were estimated. Predicted numbers of diagnosed HIV infections and AIDS cases fitted the observed data well. By the end of 2010, an estimated 13.5% (95% CI 12.5, 14.6%) of all HIV-infected MSM were undiagnosed and accounted for 81.8% (95% CI 81.1, 82.4%) of new HIV infections. The transmission rate was at its lowest from 1995-1999, with a nadir of 46 incident HIV infections in 1999, but increased from 2000. The estimated number of new infections continued to increase to more than 250 in 2010, although the reproduction number was still below the epidemic threshold. Prevention scenarios included temporary reductions in risk behaviour, annual test and treat, and reduction in risk behaviour to levels observed earlier in the epidemic. These led to predicted reductions in new infections from 2 to 26% by 2020. Parameters related to disease progression and relative infectiousness at different HIV stages had the greatest influence on estimates of the net transmission rate. CONCLUSIONS/SIGNIFICANCE: The model outputs suggest that the increase in HIV transmission amongst MSM in Switzerland is the result of continuing risky sexual behaviour, particularly by those unaware of their infection status. Long term reductions in the incidence of HIV infection in MSM in Switzerland will require increased and sustained uptake of effective interventions.

Recent misconceptions about the "database search problem": a probabilistic analysis using Bayesian networks

This paper analyses and discusses arguments that emerge from a recent discussion about the proper assessment of the evidential value of correspondences observed between the characteristics of a crime stain and those of a sample from a suspect when (i) this latter individual is found as a result of a database search and (ii) remaining database members are excluded as potential sources (because of different analytical characteristics). Using a graphical probability approach (i.e., Bayesian networks), the paper here intends to clarify that there is no need to (i) introduce a correction factor equal to the size of the searched database (i.e., to reduce a likelihood ratio), nor to (ii) adopt a propositional level not directly related to the suspect matching the crime stain (i.e., a proposition of the kind 'some person in (outside) the database is the source of the crime stain' rather than 'the suspect (some other person) is the source of the crime stain'). The present research thus confirms existing literature on the topic that has repeatedly demonstrated that the latter two requirements (i) and (ii) should not be a cause of concern.

Comparative phylogeography of mutualists and the effect of the host on the genetic structure of its partners.

Whether or not species participating in specialized and obligate interactions display similar and simultaneous demographic variations at the intraspecific level remains an open question in phylogeography. In the present study, we used the mutualistic nursery pollination occurring between the European globeflower Trollius europaeus and its specialized pollinators in the genus Chiastocheta as a case study. Explicitly, we investigated if the phylogeographies of the pollinating flies are significantly different from the expectation under a scenario of plant-insect congruence. Based on a large-scale sampling, we first used mitochondrial data to infer the phylogeographical histories of each fly species. Then, we defined phylogeographical scenarios of congruence with the plant history, and used maximum likelihood and Bayesian approaches to test for plant-insect phylogeographical congruence for the three Chiastocheta species. We show that the phylogeographical histories of the three fly species differ. Only Chiastocheta lophota and Chiastocheta dentifera display strong spatial genetic structures, which do not appear to be statistically different from those expected under scenarios of phylogeographical congruence with the plant. The results of the present study indicate that the fly species responded in independent and different ways to shared evolutionary forces, displaying varying levels of congruence with the plant genetic structure

How quickly should we titrate antihypertensive medication? Systematic review modelling blood pressure response from trial data.

Context There are no evidence syntheses available to guide clinicians on when to titrate antihypertensive medication after initiation. Objective To model the blood pressure (BP) response after initiating antihypertensive medication. Data sources electronic databases including Medline, Embase, Cochrane Register and reference lists up to December 2009. Study selection Trials that initiated antihypertensive medication as single therapy in hypertensive patients who were either drug naive or had a placebo washout from previous drugs. Data extraction Office BP measurements at a minimum of two weekly intervals for a minimum of 4 weeks. An asymptotic approach model of BP response was assumed and non-linear mixed effects modelling used to calculate model parameters. Results and conclusions Eighteen trials that recruited 4168 patients met inclusion criteria. The time to reach 50% of the maximum estimated BP lowering effect was 1 week (systolic 0.91 weeks, 95% CI 0.74 to 1.10; diastolic 0.95, 0.75 to 1.15). Models incorporating drug class as a source of variability did not improve fit of the data. Incorporating the presence of a titration schedule improved model fit for both systolic and diastolic pressure. Titration increased both the predicted maximum effect and the time taken to reach 50% of the maximum (systolic 1.2 vs 0.7 weeks; diastolic 1.4 vs 0.7 weeks). Conclusions Estimates of the maximum efficacy of antihypertensive agents can be made early after starting therapy. This knowledge will guide clinicians in deciding when a newly started antihypertensive agent is likely to be effective or not at controlling BP.

Which is the optimal sampling strategy for habitat suitability modelling

Designing an efficient sampling strategy is of crucial importance for habitat suitability modelling. This paper compares four such strategies, namely, 'random', 'regular', 'proportional-stratified' and 'equal -stratified'- to investigate (1) how they affect prediction accuracy and (2) how sensitive they are to sample size. In order to compare them, a virtual species approach (Ecol. Model. 145 (2001) 111) in a real landscape, based on reliable data, was chosen. The distribution of the virtual species was sampled 300 times using each of the four strategies in four sample sizes. The sampled data were then fed into a GLM to make two types of prediction: (1) habitat suitability and (2) presence/ absence. Comparing the predictions to the known distribution of the virtual species allows model accuracy to be assessed. Habitat suitability predictions were assessed by Pearson's correlation coefficient and presence/absence predictions by Cohen's K agreement coefficient. The results show the 'regular' and 'equal-stratified' sampling strategies to be the most accurate and most robust. We propose the following characteristics to improve sample design: (1) increase sample size, (2) prefer systematic to random sampling and (3) include environmental information in the design'

Therapeutic Drug Monitoring (TDM) of Imatinib: Effectiveness of Bayesian dose adjustment for CML patients enrolled in the Imatinib Concentration Monitoring (I-COME) study

Introduction: As imatinib pharmacokinetics are highly variable, plasma levels differ largely between patients under the same dosage. Retrospective studies in chronic myeloid leukemia (CML) patients showed significant correlations between low levels and suboptimal response, as well as between high levels and poor tolerability. Monitoring of trough plasma levels, targeting 1000 μg/L and above, is thus increasingly advised. Our study was launched to assess prospectively the clinical usefulness of systematic imatinib TDM in CML patients. This preliminary analysis addresses the appropriateness of the dosage adjustment approach applied in this study, which targets the recommended trough level and allows an interval of 4-24 h after last drug intake for blood sampling. Methods: Blood samples from the first 15 patients undergoing 1st TDM were obtained 1.5-25 h after last dose. Imatinib plasma levels were measured by LC-MS/MS and the concentrations were extrapolated to trough based on a Bayesian approach using a population pharmacokinetic model. Trough levels were predicted to differ significantly from the target in 12 patients (10 <750 μg/L; 2 >1500 μg/L along with poor tolerance) and individual dose adjustments were proposed. 8 patients underwent a 2nd TDM cycle. Trough levels of 1st and 2nd TDM were compared, the sample drawn 1.5 h after last dose (during distribution phase) was excluded from the analysis. Results: Individual dose adjustments were applied in 6 patients. Observed concentrations extrapolated to trough ranged from 360 to 1832 μg/L (median 725; mean 810, CV 52%) on 1st TDM and from 720 to 1187 μg/L (median 950; mean 940, CV 18%) on 2nd TDM cycle. Conclusions: These preliminary results suggest that TDM of imatinib using a Bayesian interpretation is able to target the recommended trough level of 1000 μg/L and to reduce the considerable differences in trough level exposure between patients (with CV decreasing from 52% to 18%). While this may simplify blood collection in daily practice, as samples do not have to be drawn exactly at trough, the largest possible interval to last drug intake yet remains preferable to avoid sampling during distribution phase leading to biased extrapolation. This encourages the evaluation of the clinical benefit of a routine TDM intervention in CML patients, which the randomized Swiss I-COME trial aims to.

Uncertainty Quantification with Support Vector Regression Prediction Models

Uncertainty quantification of petroleum reservoir models is one of the present challenges, which is usually approached with a wide range of geostatistical tools linked with statistical optimisation or/and inference algorithms. The paper considers a data driven approach in modelling uncertainty in spatial predictions. Proposed semi-supervised Support Vector Regression (SVR) model has demonstrated its capability to represent realistic features and describe stochastic variability and non-uniqueness of spatial properties. It is able to capture and preserve key spatial dependencies such as connectivity, which is often difficult to achieve with two-point geostatistical models. Semi-supervised SVR is designed to integrate various kinds of conditioning data and learn dependences from them. A stochastic semi-supervised SVR model is integrated into a Bayesian framework to quantify uncertainty with multiple models fitted to dynamic observations. The developed approach is illustrated with a reservoir case study. The resulting probabilistic production forecasts are described by uncertainty envelopes.

Postglacial recolonization at a snail's pace (Trochulus villosus): confronting competing refugia hypotheses using model selection.

The localization of Last Glacial Maximum (LGM) refugia is crucial information to understand a species' history and predict its reaction to future climate changes. However, many phylogeographical studies often lack sampling designs intensive enough to precisely localize these refugia. The hairy land snail Trochulus villosus has a small range centred on Switzerland, which could be intensively covered by sampling 455 individuals from 52 populations. Based on mitochondrial DNA sequences (COI and 16S), we identified two divergent lineages with distinct geographical distributions. Bayesian skyline plots suggested that both lineages expanded at the end of the LGM. To find where the origin populations were located, we applied the principles of ancestral character reconstruction and identified a candidate refugium for each mtDNA lineage: the French Jura and Central Switzerland, both ice-free during the LGM. Additional refugia, however, could not be excluded, as suggested by the microsatellite analysis of a population subset. Modelling the LGM niche of T. villosus, we showed that suitable climatic conditions were expected in the inferred refugia, but potentially also in the nunataks of the alpine ice shield. In a model selection approach, we compared several alternative recolonization scenarios by estimating the Akaike information criterion for their respective maximum-likelihood migration rates. The 'two refugia' scenario received by far the best support given the distribution of genetic diversity in T. villosus populations. Provided that fine-scale sampling designs and various analytical approaches are combined, it is possible to refine our necessary understanding of species responses to environmental changes.