RESUMO
The southeastern region of the United States exhibits an unusual trend of decreasing tree species richness (TSR) from higher to lower latitudes over the Florida peninsula. This trend contradicts the widely marked latitudinal diversity gradient where species richness is highest in tropical zones and decreases towards extratropical regions. This study aims to assess the environmental factors that prompt this atypical inverse latitudinal gradient seen in TSR using the USDA Forest Service's Forest Inventory and Analysis (FIA) database. Fifteen variables under four categories of forested area, groundwater, soil properties, and climate groups were examined to model TSR in the region. Generalized linear models (GLMs) with Poisson distributions first assessed individual variables to test explanatory power then the LASSO regularization method was utilized to extract two subsets of the most influential variables to predict TSR. Forest area and four climate variables (mean annual temperature, precipitation seasonality, mean temperature of coldest quarter, and mean precipitation of driest quarter) were the top five variables during the initial GLM assessment implying their potential individual influence in regulating TSR. Two subsets of LASSO models contained seven and three predictor variables, respectively. Frist subset includes seven predictors, presented in highest to low standardized coefficient, mean temperature of coldest quarter, forested area, precipitation seasonality, mean precipitation of driest quarter, water table depth, spodosol, and available water storage. The other subset further excluded four lowest influential variables from the first set, leaving the top three variables from the first subset. The first subset of the LASSO model predicted TSR with 63.4% explained deviance while the second subset reproduced 60.2% of deviance explained. With only three variables used, the second model outperformed the first model evaluated by the AIC value. We conclude that forest patch area, mean temperature of coldest quarter, and precipitation seasonality are the highly influential variables of TSR among environmental factors in the southeastern region of U.S., but evolutionary or historic cause should be further incorporated to fully understand tree species diversity pattern in this region.
RESUMO
OBJECTIVE: Although immunoglobulin (Ig)E-mediated allergies are readily identifiable, non-IgE-mediated allergies present more diagnostic difficulty. We performed a formal retrospective analysis to determine whether there is a recognizable clinical pattern in children. METHODS: We studied 121 children (mean age, 17.3 months) with multiple food allergies who were recruited on the basis of adequate immunological assessment by using case notes and parental questionnaire. RESULTS: Group 1 (n=44) had rapid reactions to dietary antigens, of whom 41 also showed delayed reactions. Group 2 (n=77) had delayed reactions only. Mean IgE was increased in group 1 but both groups otherwise shared a pattern of increased IgG1, decreased IgG2/4, and low-normal IgA. Lymphocyte subsets were skewed, with an increased percentage of CD4 and CD19 and decreased CD8 and natural killer cells. Gastroesophageal reflux, esophagitis, subtle enteropathy, and constipation were frequent in both groups. Of 55 exclusively breast-fed infants, 44 sensitized before weaning. Twenty-one of the mothers suffered from autoimmunity. CONCLUSIONS: There appears to be a recognizable pattern of immune deviation and minor enteropathy in children with multiple food allergy, irrespective of the speed of reactions. Disturbed gut motility is particularly common, as is a maternal history of autoimmunity.
