A DIRICHLET PROCESS MIXTURE OF HIDDEN MARKOV MODELS FOR PROTEIN STRUCTURE PREDICTION

By providing new insights into the distribution of a protein's torsion angles, recent statistical models for this data have pointed the way to more efficient methods for protein structure prediction. Most current approaches have concentrated on bivariate models at a single sequence position. There is, however, considerable value in simultaneously modeling angle pairs at multiple sequence positions in a protein. One area of application for such models is in structure prediction for the highly variable loop and turn regions. Such modeling is difficult due to the fact that the number of known protein structures available to estimate these torsion angle distributions is typically small. Furthermore, the data is "sparse" in that not all proteins have angle pairs at each sequence position. We propose a new semiparametric model for the joint distributions of angle pairs at multiple sequence positions. Our model accommodates sparse data by leveraging known information about the behavior of protein secondary structure. We demonstrate our technique by predicting the torsion angles in a loop from the globin fold family. Our results show that a template-based approach can now be successfully extended to modeling the notoriously difficult loop and turn regions.     

GENDER DIFFERENCES IN JOB AUTONOMY:

This research investigates the impact of gender, occupational segregation, and authority position on various forms of job autonomy. The empirical analysis was guided by theoretical concerns over the relevance of occupational characteristics and authority position in explaining the gender gap in job autonomy. Ordinary Least Squares regression results show that after taking various individual and employment-related factors into account, authority position, but not female-dominated occupation, explains a large portion of the gender gap in the degree of job autonomy. Furthermore, logistic regressions indicate that no gender differences in the likelihood of having any of five components of job autonomy remain, after taking all other factors in the model into account. It is concluded that authority position at work is more instrumental in explaining the gender gap in job autonomy than the gender composition of occupations.     

Order test for high-dimensional two-sample means

We propose a new method to test the order between two high-dimensional mean curves. The new statistic extends the approach of Follmann (1996) to high-dimensional data by adapting the strategy of Bai and Saranadasa (1996). The proposed procedure is an alternative to the non-negative basis matrix factorization (NBMF) based test of Lee et al. (2008) for the same hypothesis, but it is much easier to implement. We derive the asymptotic mean and variance of the proposed test statistic under the null hypothesis of equal mean curves. Based on theoretical results, we put forward a permutation procedure to approximate the null distribution of the new test statistic. We compare the power of the proposed test with that of the NBMF-based test via simulations. We illustrate the approach by an application to tidal volume traces.     

Resolving an Insurance Allocation Problem: A Procedural Approach

In this paper we study the problem of the determination of a fair allocation in a co-insurance problem, i.e., how some insurance companies have to share the risk and the premium. We develop two procedures that produce a proportional and an equitable allocation, respectively. The procedures are applied to a real situation arising from environmental risk and the resulting allocations are compared with the classical quota share allocation and with an envy-free allocation resulting from a procedure presented in Fragnelli and Marina (Insur. Math Econ 33:75–85, 2003).     

Infant‐directed speech to infants at risk for dyslexia: A novel cross‐dyad design

When mothers speak to infants at risk for developmental dyslexia, they do not hyperarticulate vowels in their infant‐directed speech (IDS). Here, we used an innovative cross‐dyad design to investigate whether the absence of vowel hyperarticulation in IDS to at‐risk infants is a product of maternal infant‐directed behavior or of infants’ parent‐directed cues. Interactions between mothers and infants who were at risk or not at risk for dyslexia were recorded in three conditions: when mothers interacted with (a) their own infants, (b) infants who were not their own but of the same risk status, and (c) infants who were not their own and of the opposite risk status. This design revealed both infant and parent effects. Mothers of not‐at‐risk infants hyperarticulated vowels significantly more when speaking to not‐at‐risk than to at‐risk infants. In contrast, mothers of at‐risk infants hyperarticulated vowels significantly less than NAR mothers, and this was irrespective of infant status. Mothers of not‐at‐risk infants thus adjusted their IDS to the infant's risk status, while mothers of at‐risk infants did not. We suggest that IDS is determined reciprocally by characteristics of both partners in the dyad: Both infant and maternal factors are essential for the vowel hyperarticulation component of IDS.     

Evaluating multidisciplinary child abuse and neglect teams: a research agenda

A review of child welfare research literature reveals used to investigate and intervene in child abuse and neglect cases, the field does not know enough about their structural variations, implementation processes, or effectiveness. Moreover, although articles advocating multidisciplinary teams enumerate their apparent strengths, they lack attention to the teams' possible weaknesses. The article discusses implications for future evaluation studies.     

Bayesian analysis of generalized elliptical semi-parametric models

In this paper, we study the statistical inference based on the Bayesian approach for regression models with the assumption that independent additive errors follow normal, Student-t, slash, contaminated normal, Laplace or symmetric hyperbolic distribution, where both location and dispersion parameters of the response variable distribution include nonparametric additive components approximated by B-splines. This class of models provides a rich set of symmetric distributions for the model error. Some of these distributions have heavier or lighter tails than the normal as well as different levels of kurtosis. In order to draw samples of the posterior distribution of the interest parameters, we propose an efficient Markov Chain Monte Carlo (MCMC) algorithm, which combines Gibbs sampler and Metropolis–Hastings algorithms. The performance of the proposed MCMC algorithm is assessed through simulation experiments. We apply the proposed methodology to a real data set. The proposed methodology is implemented in the R package BayesGESM using the function gesm().     

Nonparametric Tree‐Based Predictive Modeling of Storm Outages on an Electric Distribution Network

This article compares two nonparametric tree‐based models, quantile regression forests (QRF) and Bayesian additive regression trees (BART), for predicting storm outages on an electric distribution network in Connecticut, USA. We evaluated point estimates and prediction intervals of outage predictions for both models using high‐resolution weather, infrastructure, and land use data for 89 storm events (including hurricanes, blizzards, and thunderstorms). We found that spatially BART predicted more accurate point estimates than QRF. However, QRF produced better prediction intervals for high spatial resolutions (2‐km grid cells and towns), while BART predictions aggregated to coarser resolutions (divisions and service territory) more effectively. We also found that the predictive accuracy was dependent on the season (e.g., tree‐leaf condition, storm characteristics), and that the predictions were most accurate for winter storms. Given the merits of each individual model, we suggest that BART and QRF be implemented together to show the complete picture of a storm's potential impact on the electric distribution network, which would allow for a utility to make better decisions about allocating prestorm resources.     

New Methods for Resolving Conflicting Requests with Examples from Medical Residency Scheduling

In scheduling medical residents, the objective is often to maximize resident satisfaction across the space of feasible schedules, relative to the many hard constraints that ensure appropriate patient coverage, adequate training opportunities, etc. A common metric of resident satisfaction is the number of time‐off requests that are granted. Simply maximizing this total, however, may lead to undesirable schedules since some requests have higher priority than others. For example, it might be better to grant one resident's request for a family member's wedding in place of two residents’ requests to attend a rugby game. Another approach is to assign a weight to each request and maximize the total weight of granted requests, but determining weights that accurately represent residents’ and schedulers’ preferences can be quite challenging. Instead, we propose to identify the exhaustive collection of maximally feasible and minimally infeasible sets of requests which can then be used by schedulers to select their preferred solution. Specifically, we have developed two algorithms, which we call Sequential Request Selection Via Cuts (Sequential RSVC) and Simultaneous Request Selection Via Cuts (Simultaneous RSVC), to identify these sets by solving two sequences of optimization problems. We present these algorithms along with computational results based on a real‐world problem of scheduling residents at the University of Michigan C.S. Mott Pediatric Emergency Department. Although we focus our exposition on the problem of resident scheduling, our approach is applicable to a broad class of problems with soft constraints.