ROC-guided survival trees and ensembles

Tree-based methods are popular nonparametric tools in studying time-to-event outcomes. In this article, we introduce a novel framework for survival trees and ensembles, where the trees partition the dynamic survivor population and can handle time-dependent covariates. Using the idea of randomized tests, we develop generalized time-dependent receiver operating characteristic (ROC) curves for evaluating the performance of survival trees. The tree-building algorithm is guided by decision-theoretic criteria based on ROC, targeting specifically for prediction accuracy. To address the instability issue of a single tree, we propose a novel ensemble procedure based on averaging martingale estimating equations, which is different from existing methods that average the predicted survival or cumulative hazard functions from individual trees. Extensive simulation studies are conducted to examine the performance of the proposed methods. We apply the methods to a study on AIDS for illustration.     

Inference of Multisite Phosphorylation Rate Constants and Their Modulation by Pathogenic Mutations

1. Download : Download high-res image (170KB)
2. Download : Download full-size image

     

Dietary Intakes and Health Risk of Toxic and Essential Heavy Metals through the Food Chain in Agricultural,Industrial, and Coal Mining Areas of Northern India

Minerals can enter the food chain through industrial and mining activities. Soil-to-vegetable transfer is higher than soil-to-cereal, but human consumption of metals is attributable to balanced diets containing both vegetables and cereals and drinking water. However, the impact of location on intakes of metals from predominantly cereal-based Indian diets is not clear. Hence, the present study was undertaken in selected Agricultural, Industrial, and Coal Mining Areas (AA, IA, CMA) around the Allahabad District in Northern India to compare transfer of toxic metals, Pb, Cd, Cr and essential metals, Fe, Zn, Cu, Co in soil and water to common crops: cereals (rice, wheat, maize) and vegetables (spinach, potato), and to assess Daily Intake of Metal (DIM) and consequent Health Risk Index (HRI). The overall content of all metals, except Cu, in water, soils, and crops followed the pattern CMA > IA > AA. Transfer factors (TFs) followed the sequence spinach > potato > cereals. Quantitatively, however, cereals contribute maximally to a balanced diet, so DIM and HRI were higher from cereals than vegetables. Even though spinach had the highest TFs, cereals contributed maximally to HRI. CMA had the highest metal content so locally grown cereals contributed significantly to intake of both toxic and essential metals.     

Key canopy traits drive forest productivity

Quantifying the mechanistic links between carbon fluxes and forest canopy attributes will advance understanding of leaf-to-ecosystem scaling and its potential application to assessing terrestrial ecosystem metabolism. Important advances have been made, but prior studies that related carbon fluxes to multiple canopy traits are scarce. Herein, presenting data for 128 cold temperate and boreal forests across a regional gradient of 600 km and 5.4°C (from 2.4°C to 7.8°C) in mean annual temperature, I show that stand-scale productivity is a function of the capacity to harvest light (represented by leaf area index, LAI), and to biochemically fix carbon (represented by canopy nitrogen concentration, %N). In combination, LAI and canopy %N explain greater than 75 per cent of variation in above-ground net primary productivity among forests, expressed per year or per day of growing season. After accounting for growing season length and climate effects, less than 10 per cent of the variance remained unexplained. These results mirror similar relations of leaf-scale and canopy-scale (eddy covariance) maximum photosynthetic rates to LAI and %N. Collectively, these findings indicate that canopy structure and chemistry translate from instantaneous physiology to annual carbon fluxes. Given the increasing capacity to remotely sense canopy LAI, %N and phenology, these results support the idea that physiologically based scaling relations can be useful tools for global modelling.     

Fluctuated water depth with high nutrient concentrations promote the invasiveness of Wedelia trilobata in Wetland

The distribution of invasive and native species in wetlands is determined by hydrological conditions; whereas conditions such as water depth fluctuations, variations in the nutrient concentrations are expected to affect the growth and physiological traits of plants. For the assessment of such effects, we conduct greenhouse experiment with three factors; 1) water depth of 5 cm and 15 cm (static and fluctuated); 2) three levels of nutrient concentrations (i) full‐strength Hoagland solution (N1), (ii) ¼‐strength Hoagland solution (N2), and (iii) ¹/₈‐strength Hoagland solution (N3); and 3) species, invasive Wedelia trilobata (L.) and its congener, native Wedelia chinensis (Osbeck.) under mono and mixed culture. Water depth of 5 cm combined with any of the nutrient treatments significantly restrained the photosynthesis, intracellular CO₂ concentration and leaf chlorophyll of both W. trilobata and W. chinensis. Increase in the water depth to 15 cm with low‐nutrient treatment N3 did not sustain the physiological traits of W. chinensis under mono and mixed planting. A great loss was noted in the growth of W. chinensis at 15 cm static and fluctuated water depth with low‐nutrient treatment (N3) and under mixed culture. In addition, water depth fluctuations with both low‐ and high‐nutrient treatments significantly affected the root‐shoot ratio, relative growth rate, and interspecific interaction among these two species. W. trilobata benefited more from competitive interaction index (CII) under fluctuated water depth at 15 cm with high nutrients, and the value of CII was clearly positive. Therefore, higher competitive ability may contribute to the invasiveness of W. trilobata in wetlands.     

Evaluating the wind cooling potential on outdoor thermal comfort in selected Iranian climate types

Wind is one of the main factors affecting people's outdoor thermal sensation. Ongoing urbanization and urban densification are transforming the urban climate and complicating the pedestrian-level wind environment. Therefore, the main aim of this research is to evaluate the potential wind-cooling effect on human outdoor thermal conditions. Accordingly, the current research attempts determine the best wind directions for thermal comfort at the studied stations and how these factors will be changed under the effects of global warming. Outdoor thermal conditions were modeled based on the physiologically equivalent temperature (PET) thermal index using RayMan software for the decades of the 2000s and the 2040s in different climate types of Iran (Csb, BWh, Csa, and BSh) To estimate the potential cooling effect of wind, the PET was calculated (1) under actual wind conditions, and (2) under calm wind (0.05 m/s) conditions. Then, the ΔPET for these two conditions, which indicates the cooling potential effect (CPE) of the wind, was calculated for four representative stations (Ardebil, Bandar Abbas, Gorgan, and Shiraz). In comparison with the 2000s, the results indicated that by the 2040s, the predicted wind cooling potential will have increased in Ardebil, Shiraz, Bandar Abbas and Gorgan (CPE of 13.2 °C, 13.1 °C, 11.2 °C, and 11 °C, respectively). Based on the overall average of two climate change scenarios (A2 and B1) used in this study, the occurrence of “comfortable” conditions by the 2040s will have increased in Bandar Abbas, Shiraz, and Ardebil by 1.1%, 0.4%, and 0.3%, respectively, while it will have decreased in Gorgan by 1.5%. Accounting for the cooling effect of wind, the comfort cooling potential of wind is predicted to rise by an average of 1.6 °C in the 2040s compared with the 2000s in all the studied stations. Therefore, this will affect the microclimates positively and could reduce the urban heat island effects.     

Microclimate modeling in naturally ventilated dairy barns during the hot season: Checking the accuracy of forecasts

Monitoring and predicting the microclimate in naturally ventilated barns (NVB) is important given the adverse effects of high summer temperatures on dairy cows in the context of global climate change. The aim of the work was to verify the accuracy of the microclimate forecast in a NVB using linear regression (LR). Our working hypothesis suggested that multiple periodic measurements of air temperature and relative humidity outside and inside the barns at the same time will allow us to build LR models for predicting the temperature-humidity index (THI). This was done not only for a specific dairy barn based on this indicator outside, but also in other dairy barns with a similar design, located in similar weather conditions. The results of the research indicate that the use of LR had a high accuracy of forecasting (93–96%) the THI in NVB of various designs during the summer heat. At the same time, differences were found between traits (air temperature, relative humidity as well as resulting THI) provided by meteorological weather stations and these data measured simultaneously next to the dairy barns. The proposed LR models can be used to predict THI in NVBs of different designs.     

基于“源-汇”生态过程的长江上游农业非点源污染

景观空间格局是农业非点源污染的主要影响因素之一,关于二者的相互关系缺乏定量研究.针对长江上游的农业非点源污染问题,应用基于"源-汇"生态过程理论提出的景观空间负荷对比指数,选取9个典型的行政单元,探讨了长江上游流域景观空间格局和非点源污染之间的定量关系.结果表明,景观空间负荷对比指数对非点源污染负荷有显著的响应关系,说明景观空间负荷对比指数可作为非点源污染空间风险评价的有用方法之一.在此基础上,进一步分析了长江上游典型行政单元景观空间负荷对比指数时空演变规律.探讨了指数演变时空差异原因,认为该指数主要受到区域景观的坡度、与污染出口相对距离、高程、土地利用类型比重和农业与农村经济政策的影响.最后提出了长江上游非点源污染空间风险控制与管理的对策.     

Identifying patterns in the diet of mackerel icefish (<Emphasis Type="Italic">Champsocephalus gunnari)</Emphasis> at South Georgia using bootstrapped confidence intervals of a dietary index

Ontogenetic, inter-annual and regional variations in diet were investigated for mackerel icefish, Champsocephalus gunnari, in three successive summer seasons around South Georgia. Stomach contents from 2239 C. gunnari (130–560 mm total length) were examined. A bootstrapping technique was used to calculate confidence intervals for an index of relative importance of prey categories (% IRI_DC). Diet varied significantly between years and age classes but there was little regional difference in diet. In general, diet was dominated by krill, Euphausia superba and by the amphipod Themisto gaudichaudii. Smaller (younger) fish tended to prey on a higher proportion of T. gaudichaudii and small euphausiids such as Thysanoessa sp. and took smaller quantities of E. superba. In a season of poor krill availability (summer of 2003–2004) the proportion of krill in the diet, stomach fullness and fish condition (indicated by length–weight relationships) were significantly lower than in the other summer seasons. A large reduction (>80%) in the estimated annual (2005) biomass of the C. gunnari stock directly followed the season of poor krill availability. This decline was largely because of mortality of 2+ and 3+ fish, which were more krill dependent than 1+ fish. Younger fish appear to have survived, leading to an increase in the estimated population biomass in 2006.     

Alcohol Intake and 8‐Year Weight Gain in Women: A Prospective Study

Objective: To examine prospectively the relationship between alcohol and 8‐year weight gain in women. Research Methods and Procedures: A prospective study of 49, 324 women 27 to 44 years old who did not have a history of cardiovascular disease, cancer, or diabetes, who were not pregnant during the study period, and who reported weights in 1991 and 1999. Results: In cross‐sectional analyses, there was a significant inverse relationship between alcohol and BMI even after adjustment for dietary factors and a wide range of confounders. In multivariate prospective analyses, a nonlinear relationship was seen between alcohol and weight gain (≥5 kg) in all women. Compared with nondrinkers, the adjusted relative odds [95% confidence interval (CI)] of weight gain according to grams per day were 0.94 (0.89, 0.99) for those consuming 0.1 to 4.9 g/d, 0.92 (0.85, 0.99) for 5 to 14.9 g/d, 0.86 (0.76, 0.78) for 15 to 29.9 g/d, and 1.07 (0.89, 1.28) for those consuming 30+ g/d (p < 0.0001 for quadratic trend). Women who continued to drink heavily and those who became heavy drinkers showed similar increased odds of weight gain. The increased odds of weight gain associated with heavy drinking (30+ g/d) were most marked in the younger women (<35 years) (odds ratio 1.64; 5% CI 1.03 to 2.61). In African‐American women, light drinking was associated with increased odds of weight gain compared with nondrinkers (odds ratio = 2.43; 95% CI 1.22 to 4.82) Discussion: Our data suggest that light to moderate drinking (up to 30 g/d) is not associated with weight gain in women except possibly in African‐American women. Heavier drinking may promote weight gain in women.