Comparison of complex modeling strategies for prediction of a binary outcome based on a few,highly correlated predictors

Motivated by a clinical prediction problem, a simulation study was performed to compare different approaches for building risk prediction models. Robust prediction models for hospital survival in patients with acute heart failure were to be derived from three highly correlated blood parameters measured up to four times, with predictive ability having explicit priority over interpretability. Methods that relied only on the original predictors were compared with methods using an expanded predictor space including transformations and interactions. Predictors were simulated as transformations and combinations of multivariate normal variables which were fitted to the partly skewed and bimodally distributed original data in such a way that the simulated data mimicked the original covariate structure. Different penalized versions of logistic regression as well as random forests and generalized additive models were investigated using classical logistic regression as a benchmark. Their performance was assessed based on measures of predictive accuracy, model discrimination, and model calibration. Three different scenarios using different subsets of the original data with different numbers of observations and events per variable were investigated. In the investigated setting, where a risk prediction model should be based on a small set of highly correlated and interconnected predictors, Elastic Net and also Ridge logistic regression showed good performance compared to their competitors, while other methods did not lead to substantial improvements or even performed worse than standard logistic regression. Our work demonstrates how simulation studies that mimic relevant features of a specific data set can support the choice of a good modeling strategy.     

The hMLH1 promoter polymorphisms and cancer susceptibility in Asian populations: A meta-analysis

Background

A variety of studies have evaluated the associations between polymorphisms in the promoter regions of the hMLH1 and cancer risk. However, the results remain inconclusive. To better understand the roles of the hMLH1 polymorphisms and cancer risk, we conducted a comprehensive meta-analysis to investigate the association between the hMLH1 − 93G/A and 1151T/A (Val384Asp) polymorphisms and cancer risk in Asian population.

Methods

We performed a meta-analysis by conducting searches of the published studies in Pub Med, CNKI, CBM, ISI web of knowledge and Google scholar search databases. Finally, 12 studies were included into our meta-analysis. Overall and subgroup analyses were performed. Odds ratio (OR) and 95% confidence interval (CI) were used to evaluate the associations between hMLH1 polymorphisms and cancer risk. Statistical analysis was performed with Review Manager 5.0.

Results

Twelve studies addressing two hMLH1 polymorphisms were analyzed among a total of 4128 cancer cases and 4678 controls. For hMLH1 − 93G/A, there was no evidence that the hMLH1 − 93G/A polymorphism was significantly associated with an increased cancer risk (P > 0.05) in Asian populations (heterozygote comparison: OR = 0.89 [95% CI (0.75, 1.060)] P = 0.20; dominant model comparison: OR = 0.98 [95% CI (0.83, 1.15)] P = 0.79). In subgroup analysis based on cancer types and the sources of control, no associations were found in colorectal cancer, gastric cancer and “other cancers” under the any gene model except for lung cancer (recessive model comparison: OR = 1.69 [95% CI (1.30, 2.19)] P < 0.0001). For hMLH1 1151T/A, the polymorphism significantly associated with an increased cancer risk in Asians: OR = 1.88 [95% CI (1.49, 2.25)], P < 0.0001, and OR = 1.87 [95% CI (1.49, 2.25)], P < 0.0001.

Conclusions

Our investigations demonstrated that the hMLH1 − 93G/A polymorphism is not a candidate for susceptibility to overall cancers, and that the hMLH1 1151T/A polymorphism is significantly associated with higher cancer risk in Asian populations. Further studies with large sample size for hMLH1 should be conducted.     

Evaluation of glutathione S-transferase genetic variants affecting type 2 diabetes susceptibility: A meta-analysis

Genetic polymorphisms of glutathione S-transferases (GSTs) and type 2 diabetes mellitus (T2DM) risk have been widely studied, however, the results were somewhat conflicting. To evaluate the association of GSTs (GSTM1, GSTT1 and GSTP1) gene polymorphisms with T2DM, a meta-analysis was performed before October, 2012. ORs were pooled according to random-effects model. There were a total of 1354/1666 (n = 9) cases/controls (studies) for GSTM1, 1271/1470 (n = 8) for GSTT1, and 1205/1250 (n = 7) for GSTM1. There were significant associations between GSTM1 polymorphism, GSTT1 polymorphism and T2DM in the contrast of present genotype vs. null genotype, with pooled OR = 1.99 (95%CI = 1.46–2.71) and OR = 1.61 (95%CI = 1.19–2.17), respectively. Yet no significant association of GSTP1 polymorphism and T2DM was showed. When stratified by ethnicity, the significant associations were also existed in Asians for GSTM1 and GSTT1, but not GSTP1. No publication bias but some extent of heterogeneity was observed. Finally, the accumulated evidence proved the obvious associations of GSTM1 and GSTT1 polymorphisms with an increased risk of T2DM.     

避难所对七星瓢虫捕食桃蚜功能反应的影响

自然情况下的捕食是以植物为中介发生的。作者采用植物模型作为捕食环境,测试了单纯的物理因素(避难所)对七星瓢虫Coccinella septempunctata L.捕食桃蚜Myzus persicae(Sulzer)的功能反应的影响。结果表明,避难所的存在能造成七星瓢虫捕食量和捕食率极显著下降,对雄虫的影响显著比雌虫强,且在中等猎物密度时影响更为突出。Logistic回归分析发现,七星瓢虫雌虫和雄虫无论避难所是否存在,功能反应类型均为HollingⅡ型。避难所不存在的情况下,雌、雄成虫的圆盘方程分别为N a=1.0259 N0/(1+0.003078 N0)和N a=0.9246 N0/(1+0.002127 N0),避难所存在的情况下,其圆盘方程分别为N a=0.9500 N0/(1+0.00342 N0)和N a=0.6463N0/(1+0.002262 N0)。避难所的存在能导致两性成虫瞬时攻击率极显著下降,但仅雄虫处置时间极显著延长,说明雄虫在环境异质性较高时捕获猎物更困难。     

Morphometric and molecular variation in concert: taxonomy and genetics of the reticulate Pyrenean and Iberian alpine spiny fescues (Festuca eskia complex,Poaceae)

The Iberian mountain spiny fescues are a reticulate group of five diploid grass taxa consisting of three parental species and two putative hybrids: F. × souliei (F. eskia × F. quadriflora) and F. × picoeuropeana (F. eskia × F. gautieri). Phenotypic and molecular studies were conducted with the aim of determining the taxonomic boundaries and genetic relationships of the five taxa and disentangling the origins of the two hybrids. Statistical analyses of 31 selected phenotypic traits were conducted on individuals from 159 populations and on nine type specimens. Molecular analyses of random amplified polymorphic DNA (RAPD) markers were performed on 29 populations. The phenotypic analyses detected significant differences between the five taxa and demonstrated the overall intermediacy of the F. × picoeuropeana and F. × souliei between their respective parents. The RAPD analysis corroborated the genetic differentiation of F. eskia, F. gautieri and F. quadriflora and the intermediate nature of the two hybrids; however, they also detected genetic variation within F. × picoeuropeana. These results suggest distinct origins for F. × picoeuropeana in the Cantabrian and Pyrenean mountains, with the sporadic Pyrenean populations having potentially resulted from recent hybridizations and the stabilized Cantabrian ones from older events followed by potential displacements of the parents. © 2013 The Linnean Society of London, Botanical Journal of the Linnean Society, 2013, 173 , 676–706.     

Current methods for molecular typing of Campylobacter species

Campylobacter remains one of the most common bacterial causes of gastroenteritis worldwide. Tracking sources of this organism is challenging due to the large numbers of human cases, and the prevalence of this organism throughout the environment due to growth in a wide range of animal species. Many molecular subtyping methods have been developed to characterize Campylobacter species, but only a few are commonly used in molecular epidemiology studies. This review examines the applicability of these methods, as well as the role that emerging whole genome sequencing technologies will play in tracking sources of Campylobacter spp. infection.     

Environmental correlates of annual survival differ between two ecologically similar and congeneric owls

Understanding how survival is affected by the environment is essential to gain insight into population dynamics and the evolution of life‐history traits as well as to identify environmental selection pressures. However, we still have little understanding of the relative effect of different environmental factors and their interactions on demographic traits and population dynamics. Here we used two long‐term, individual‐based datasets on Tawny Owl Strix aluco (1981–2010) and Ural Owl S. uralensis (1986–2010) to undertake capture‐mark‐recapture analysis of annual survival of adult females in response to three biologically meaningful environmental variables and their two‐way interactions. Despite the similar ecology of these two species, their survival was associated with different and uncorrelated environmental drivers. The main correlate of Tawny Owl survival was an inverse association with snow depth (winter severity). For Ural Owl, high food (vole) abundance improved survival during years with deep snow, but was less important during years with little snow. In addition, Ural Owl survival was strongly density‐dependent, whereas Tawny Owl survival was not. Our findings advise caution in extrapolating demographic inferences from one species to another, even when they are very closely related and ecologically similar. Analyses including only one or few potential environmental drivers of a species' survival may lead to incomplete conclusions because survival may be affected by several factors and their interactions.     

Establishment of a New Cross of the Rice Blast Fungus Derived from Japanese Differential Strain Ina168 and Hermaphroditic Rice Pathogen Guy11

Mating experiments between Magnaporthe grisea Japanese rice pathogens and Guy11, a hermaphroditic fertile rice pathogen, were done aimed at identification of avirulence genes. A cross named cross 2107 with thirty-six random progenies was obtained. Segregation analyses of genetic markers found that the cross was less suitable for genetic analysis. Backcrosses with cross 2107 progenies and Guy11 were done and another cross named cross 5307 with sixty-five progenies was obtained. A locus controlling kasugamycin resistance named Ksg1R was identified and used for a model case of genetic mapping. Bulked segregant analysis was done to find adjacent RAPD markers for mapping of the gene. Three adjacent markers to Ksg1R were obtained and a genetic map around the Ksg1R was made, but these markers were not located on a single chromosome. These results suggest that genetic analysis to identify a gene locus is available in cross 5307. Infection assay of parental strains of cross 5307 to Japanese differential rice cultivars suggested the possibility of genetic analysis of cultivar specificity toward four rice cultivars: Aichiasahi, Kusabue, Tsuyuake, and K59.     

Dynamic occupancy models for analyzing species' range dynamics across large geographic scales

Large‐scale biodiversity data are needed to predict species' responses to global change and to address basic questions in macroecology. While such data are increasingly becoming available, their analysis is challenging because of the typically large heterogeneity in spatial sampling intensity and the need to account for observation processes. Two further challenges are accounting for spatial effects that are not explained by covariates, and drawing inference on dynamics at these large spatial scales. We developed dynamic occupancy models to analyze large‐scale atlas data. In addition to occupancy, these models estimate local colonization and persistence probabilities. We accounted for spatial autocorrelation using conditional autoregressive models and autologistic models. We fitted the models to detection/nondetection data collected on a quarter‐degree grid across southern Africa during two atlas projects, using the hadeda ibis (Bostrychia hagedash) as an example. The model accurately reproduced the range expansion between the first (SABAP1: 1987–1992) and second (SABAP2: 2007–2012) Southern African Bird Atlas Project into the drier parts of interior South Africa. Grid cells occupied during SABAP1 generally remained occupied, but colonization of unoccupied grid cells was strongly dependent on the number of occupied grid cells in the neighborhood. The detection probability strongly varied across space due to variation in effort, observer identity, seasonality, and unexplained spatial effects. We present a flexible hierarchical approach for analyzing grid‐based atlas data using dynamical occupancy models. Our model is similar to a species' distribution model obtained using generalized additive models but has a number of advantages. Our model accounts for the heterogeneous sampling process, spatial correlation, and perhaps most importantly, allows us to examine dynamic aspects of species ranges.     

Band reporting rates of waterfowl: does individual heterogeneity bias estimated survival rates?

In capture–recapture studies, the estimation accuracy of demographic parameters is essential to the efficacy of management of hunted animal populations. Dead recovery models based upon the reporting of rings or bands are often used for estimating survival of waterfowl and other harvested species. However, distance from the ringing site or condition of the bird may introduce substantial individual heterogeneity in the conditional band reporting rates (r), which could cause bias in estimated survival rates (S) or suggest nonexistent individual heterogeneity in S. To explore these hypotheses, we ran two sets of simulations (n = 1000) in MARK using Seber''s dead recovery model, allowing time variation on both S and r. This included a series of heterogeneity models, allowing substantial variation on logit(r), and control models with no heterogeneity. We conducted simulations using two different values of S: S = 0.60, which would be typical of dabbling ducks such as mallards (Anas platyrhynchos), and S = 0.80, which would be more typical of sea ducks or geese. We chose a mean reporting rate on the logit scale of −1.9459 with SD = 1.5 for the heterogeneity models (producing a back-transformed mean of 0.196 with SD = 0.196, median = 0.125) and a constant reporting rate for the control models of 0.196. Within these sets of simulations, estimation models where σ_S = 0 and σ_S > 0 (σ_S is SD of individual survival rates on the logit scale) were incorporated to investigate whether real heterogeneity in r would induce apparent individual heterogeneity in S. Models where σ_S = 0 were selected approximately 91% of the time over models where σ_S > 0. Simulation results showed < 0.05% relative bias in estimating survival rates except for models estimating σ_S > 0 when true S = 0.8, where relative bias was a modest 0.5%. These results indicate that considerable variation in reporting rates does not cause major bias in estimated survival rates of waterfowl, further highlighting the robust nature of dead recovery models that are being used for the management of harvested species.