Mapping complex traits using Random Forests

Random Forest is a prediction technique based on growing trees on bootstrap samples of data, in conjunction with a random selection of explanatory variables to define the best split at each node. In the case of a quantitative outcome, the tree predictor takes on a numerical value. We applied Random Forest to the first replicate of the Genetic Analysis Workshop 13 simulated data set, with the sibling pairs as our units of analysis and identity by descent (IBD) at selected loci as our explanatory variables. With the knowledge of the true model, we performed two sets of analyses on three phenotypes: HDL, triglycerides, and glucose. The goal was to approach the mapping of complex traits from a multivariate perspective. The first set of analyses mimics a candidate gene approach with a high proportion of true genes among the predictors while the second set represents a genome scan analysis using microsatellite markers. Random Forest was able to identify a few of the major genes influencing the phenotypes, such as baseline HDL and triglycerides, but failed to identify the major genes regulating baseline glucose levels.     

Arthropod Diversity and Conservation in Old-Growth Northwest Forests

Old-growth forests of the Pacific Northwest extend along thecoastal region from southern Alaska to northern California andare composed largely of conifer rather than hardwood tree species.Many of these trees achieve great age (500–1,000 yr).Natural succession that follows forest stand destruction normallytakes over 100 years to reach the young mature forest stage.This succession may continue on into old-growth for centuries.The changing structural complexity of the forest over time,combined with the many different plant species that characterizesuccession, results in an array of arthropod habitats. It isestimated that 6,000 arthropod species may be found in suchforests—over 3,400 different species are known from asingle 6,400 ha site in Oregon. Our knowledge of these speciesis still rudimentary and much additional work is needed throughoutthis vast region. Many of these species play critical rolesin the dynamics of forest ecosystems. They are important innutrient cycling, as herbivores, as natural predators and parasitesof other arthropod species. This faunal diversity reflects thediversity of the environment and the arthropod complex providesa sensitive barometer of the conditions of the forest. Conservationefforts for forest arthropods are limited at present and controlledlargely by land-use policies. For example, an effort is beingmade to include arthropods in conservation efforts for the NorthernSpotted Owl and arthropods will be included in the Forest HealthMonitoring program now underway by the U.S. Forest Service.Evidence from other parts of the world suggest that arthropodsthat depend upon large pieces of dead wood may be particularlythreatened by forest management practices. Much remains to bedone in the conservation of forest arthropods.     

Classification and Phytogeography of Larch Forests of Northeast Asia

This study focuses on a phytosociological system of the extensive Larix gmelinii and Larix cajanderi dominated forests and woodlands that span the nemoral, boreal and subarctic zones in northeast Asia in the Russian Far East, Eastern Siberia and northern China, an area that covers more than 5 million km². A total of 299 relevés representing different parts of the larch forest range were used to classify communities in this area into the boreal Rhytidio rugosi-Laricetea sibiricae and Vaccinio-Piceetea and to the nemoral Querco mongolicae-Betuletea davuricae. The hierarchical system of lower syntaxa of larch forests and woodlands of northeast Asia includes 24 associations and nine subassociations. Nineteen associations and two alliances – Rhododendro aurei-Laricion cajanderi and Rhododendro daurici-Laricion gmelinii – are described for the first time. Two associations were classified into Dictamno dasycarpi-Quercion mongolicae of Lespedezo bicoloris-Quercetalia mongolicae representing the nemoral class Querco mongolicae-Betuletea davuricae. All larch communities in the nemoral zone represent different post-fire-event stages. Eleven associations occurring in humid conditions of the boreal zone and in the areas influenced by the Pacific Ocean are classified into Abieti nephrolepidis-Piceion jezoensis, Piceion jezoensis and Pino pumilae-Piceion jezoensis representing boreal forests of humid maritime regions belonging to the order Abieti veitchii-Piceetalia jezoensis. In this area seral larch communities occupy zonal sites, and primeval larch forests may occur only in edaphic conditions restricting the distribution of zonal dominants, Picea jezoensis and the species of Abies. The area of continuous permafrost within the boreal zone having continental and ultracontinental climate conditions is occupied by primeval larch forests united in Ledo palustris-Laricion cajanderi and Rhododendro aurei-Laricion cajanderi of Ledo palustris-Laricetalia cajanderi, and on sites with dry soils with a deeper frozen horizon in Lathyro humilis-Laricion cajanderi of Lathyro humilis-Laricetalia cajanderi.     

Nudivirus Genomics： Diversity and Classification

Nudiviruses represent a diverse group of arthropod specific, rod-shaped and dsDNA viruses. Due to similarities in pathology and morphology to members of the family Baculoviridae, they have been previously classified as the so-called “non-occluded” baculoviruses. However, presently they are taxonomically orphaned and are not assigned to any virus family because of the lack of genetic relatedness to Baculoviridae,. Here, we report on recent progress in the genomic analysis of Heliothis zea nudivirus 1 (HzNV-1), Oryctes rhinoceros nudivirus (OrNV), Gryllus bimaculatus nudivirus (GbNV) and Heliotis zea nudivirus 2 (HzNV-2). Gene content comparison and phylogenetic analyses indicated that the viruses share 15 core genes with baculoviruses and form a monophyletic sister group to them. Consequences of the genetic relationship are discussed for the classification of nudiviruses.     

Hydrologic Landscape Regionalisation Using Deductive Classification and Random Forests

Landscape classification and hydrological regionalisation studies are being increasingly used in ecohydrology to aid in the management and research of aquatic resources. We present a methodology for classifying hydrologic landscapes based on spatial environmental variables by employing non-parametric statistics and hybrid image classification. Our approach differed from previous classifications which have required the use of an a priori spatial unit (e.g. a catchment) which necessarily results in the loss of variability that is known to exist within those units. The use of a simple statistical approach to identify an appropriate number of classes eliminated the need for large amounts of post-hoc testing with different number of groups, or the selection and justification of an arbitrary number. Using statistical clustering, we identified 23 distinct groups within our training dataset. The use of a hybrid classification employing random forests extended this statistical clustering to an area of approximately 228,000 km² of south-eastern Australia without the need to rely on catchments, landscape units or stream sections. This extension resulted in a highly accurate regionalisation at both 30-m and 2.5-km resolution, and a less-accurate 10-km classification that would be more appropriate for use at a continental scale. A smaller case study, of an area covering 27,000 km², demonstrated that the method preserved the intra- and inter-catchment variability that is known to exist in local hydrology, based on previous research. Preliminary analysis linking the regionalisation to streamflow indices is promising suggesting that the method could be used to predict streamflow behaviour in ungauged catchments. Our work therefore simplifies current classification frameworks that are becoming more popular in ecohydrology, while better retaining small-scale variability in hydrology, thus enabling future attempts to explain and visualise broad-scale hydrologic trends at the scale of catchments and continents.     

Mapping Global Diversity Patterns for Migratory Birds

Nearly one in five bird species has separate breeding and overwintering distributions, and the regular migrations of these species cause a substantial seasonal redistribution of avian diversity across the world. However, despite its ecological importance, bird migration has been largely ignored in studies of global avian biodiversity, with few studies having addressed it from a macroecological perspective. Here, we analyse a dataset on the global distribution of the world’s birds in order to examine global spatial patterns in the diversity of migratory species, including: the seasonal variation in overall species diversity due to migration; the contribution of migratory birds to local bird diversity; and the distribution of narrow-range and threatened migratory birds. Our analyses reveal a striking asymmetry between the Northern and Southern hemispheres, evident in all of the patterns investigated. The highest migratory bird diversity was found in the Northern Hemisphere, with high inter-continental turnover in species composition between breeding and non-breeding seasons, and extensive regions (at high latitudes) where migratory birds constitute the majority of the local avifauna. Threatened migratory birds are concentrated mainly in Central and Southern Asia, whereas narrow-range migratory species are mainly found in Central America, the Himalayas and Patagonia. Overall, global patterns in the diversity of migratory birds indicate that bird migration is mainly a Northern Hemisphere phenomenon. The asymmetry between the Northern and Southern hemispheres could not have easily been predicted from the combined results of regional scale studies, highlighting the importance of a global perspective.     

Mapping the Diversity of Maize Races in Mexico

Traditional landraces of maize are cultivated throughout more than one-half of Mexico''s cropland. Efforts to organize in situ conservation of this important genetic resource have been limited by the lack of knowledge of regional diversity patterns. We used recent and historic collections of maize classified for race type to determine biogeographic regions and centers of landrace diversity. We also analyzed how diversity has changed over the last sixty years. Based on racial composition of maize we found that Mexico can be divided into 11 biogeographic regions. Six of these biogeographic regions are in the center and west of the country and contain more than 90% of the reported samples for 38 of the 47 races studied; these six regions are also the most diverse. We found no evidence of rapid overall decline in landrace diversity for this period. However, several races are now less frequently reported and two regions seem to support lower diversity than in previous collection periods. Our results are consistent with a previous hypothesis for diversification centers and for migration routes of original maize populations merging in western central Mexico. We provide maps of regional diversity patterns and landrace based biogeographic regions that may guide efforts to conserve maize genetic resources.     

Effects of Previous Land-Use on Plant Species Composition and Diversity in Mediterranean Forests

At some point in their history, most forests in the Mediterranean Basin have been subjected to intensive management or converted to agriculture land. Knowing how forest plant communities recovered after the abandonment of forest-management or agricultural practices (including livestock grazing) provides a basis for investigating how previous land management have affected plant species diversity and composition in forest ecosystems. Our study investigated the consequences of historical “land management” practices on present-day Mediterranean forests by comparing species assemblages and the diversity of (i) all plant species and (ii) each ecological group defined by species’ habitat preferences and successional status (i.e., early-, mid-, and late-successional species). We compared forest stands that differed both in land-use history and in successional stage. In addition, we evaluated the value of those stands for biodiversity conservation. The study revealed significant compositional differentiation among stands that was due to among-stand variations in the diversity (namely, species richness and evenness) of early-, intermediate-, and late-successional species. Historical land management has led to an increase in compositional divergences among forest stands and the loss of late-successional forest species.     

Ground Layer Plant Species Turnover and Beta Diversity in Southern-European Old-Growth Forests

Different assembly processes may simultaneously affect local-scale variation of species composition in temperate old-growth forests. Ground layer species diversity reflects chance colonization and persistence of low-dispersal species, as well as fine-scale environmental heterogeneity. The latter depends on both purely abiotic factors, such as soil properties and topography, and factors primarily determined by overstorey structure, such as light availability. Understanding the degree to which plant diversity in old-growth forests is associated with structural heterogeneity and/or to dispersal limitation will help assessing the effectiveness of silvicultural practices that recreate old-growth patterns and structures for the conservation or restoration of plant diversity. We used a nested sampling design to assess fine-scale species turnover, i.e. the proportion of species composition that changes among sampling units, across 11 beech-dominated old-growth forests in Southern Europe. For each stand, we also measured a wide range of environmental and structural variables that might explain ground layer species turnover. Our aim was to quantify the relative importance of dispersal limitation in comparison to that of stand structural heterogeneity while controlling for other sources of environmental heterogeneity. For this purpose, we used multiple regression on distance matrices at the within-stand extent, and mixed effect models at the extent of the whole dataset. Species turnover was best predicted by structural and environmental heterogeneity, especially by differences in light availability and in topsoil nutrient concentration and texture. Spatial distances were significant only in four out of eleven stands with a relatively low explanatory power. This suggests that structural heterogeneity is a more important driver of local-scale ground layer species turnover than dispersal limitation in southern European old-growth beech forests.     

东北山地次生林鸟类多样性的研究

山地次生林是东北地区半山区重要的生态环境。随着森林的采伐,次生林越来越多,与人类的关系也越来越密切。有关次生林鸟类多样性的研究,国内报道甚少。我们于1983—1984年4—7月在吉林省蛟     

Newer Classification and Regression Tree Techniques: Bagging and Random Forests for Ecological Prediction

The task of modeling the distribution of a large number of tree species under future climate scenarios presents unique challenges. First, the model must be robust enough to handle climate data outside the current range without producing unacceptable instability in the output. In addition, the technique should have automatic search mechanisms built in to select the most appropriate values for input model parameters for each species so that minimal effort is required when these parameters are fine-tuned for individual tree species. We evaluated four statistical models—Regression Tree Analysis (RTA), Bagging Trees (BT), Random Forests (RF), and Multivariate Adaptive Regression Splines (MARS)—for predictive vegetation mapping under current and future climate scenarios according to the Canadian Climate Centre global circulation model. To test, we applied these techniques to four tree species common in the eastern United States: loblolly pine (Pinus taeda), sugar maple (Acer saccharum), American beech (Fagus grandifolia), and white oak (Quercus alba). When the four techniques were assessed with Kappa and fuzzy Kappa statistics, RF and BT were superior in reproducing current importance value (a measure of basal area in addition to abundance) distributions for the four tree species, as derived from approximately 100,000 USDA Forest Service’s Forest Inventory and Analysis plots. Future estimates of suitable habitat after climate change were visually more reasonable with BT and RF, with slightly better performance by RF as assessed by Kappa statistics, correlation estimates, and spatial distribution of importance values. Although RTA did not perform as well as BT and RF, it provided interpretive models for species whose distributions were captured well by our current set of predictors. MARS was adequate for predicting current distributions but unacceptable for future climate. We consider RTA, BT, and RF modeling approaches, especially when used together to take advantage of their individual strengths, to be robust for predictive mapping and recommend their inclusion in the ecological toolbox.     

Disaggregating Census Data for Population Mapping Using Random Forests with Remotely-Sensed and Ancillary Data

High resolution, contemporary data on human population distributions are vital for measuring impacts of population growth, monitoring human-environment interactions and for planning and policy development. Many methods are used to disaggregate census data and predict population densities for finer scale, gridded population data sets. We present a new semi-automated dasymetric modeling approach that incorporates detailed census and ancillary data in a flexible, “Random Forest” estimation technique. We outline the combination of widely available, remotely-sensed and geospatial data that contribute to the modeled dasymetric weights and then use the Random Forest model to generate a gridded prediction of population density at ~100 m spatial resolution. This prediction layer is then used as the weighting surface to perform dasymetric redistribution of the census counts at a country level. As a case study we compare the new algorithm and its products for three countries (Vietnam, Cambodia, and Kenya) with other common gridded population data production methodologies. We discuss the advantages of the new method and increases over the accuracy and flexibility of those previous approaches. Finally, we outline how this algorithm will be extended to provide freely-available gridded population data sets for Africa, Asia and Latin America.     

抚育间伐对上海水源涵养林林分生长及其林下植物多样性的影响

对不同强度间伐后上海水源涵养林林分的生长及其林下植物多样性变化情况进行调查。结果表明,间伐2年后,不同强度间伐对林分胸径和树高生长的影响差异不明显;间伐4年后,间伐样地的胸径和树高生长量表现不一,不同强度间伐对3种林分生长和林下植物多样性的影响不同。不同的林分应采用不同的间伐强度。杜英纯林宜采用65%(T3)间伐强度;香樟重阳木混交林采用30%(T5)间伐强度;多树种混交林采用20%~30%(T7~T8)间伐强度。     

Species Composition, Diversity and Local uses of Tropical Dry Deciduous and Gallery Forests in Nicaragua

The floristic composition and diversity of tropical dry deciduous and gallery forests were studied in Chacocente Wildlife Refuge, located on the Pacific coast in Nicaragua during 1994 and 2000. Density, dominance and frequency as well as species and family important values were computed to characterize the floristic composition. A variety of diversity measures were also calculated to examine heterogeneity in each forest community. A total of 29 families, 49 genera and 59 species were represented in 2 ha dry deciduous forest. In the gallery forest, the number of families, genera and species recorded in 2000 inventory was 33, 48 and 58, respectively and slightly higher than the 1994 inventory. The number of stems ⩽ 10 cm dbh varied from 451 to 489 per hectare in the deciduous forest, and from 283 to 298 per hectare in the gallery forest. The basal area was much larger for species in the gallery than dry deciduous forest. Fabaceae, sub family Papilionoideae, was the most specious family in the deciduous forest while Meliaceae was the dominant family in the gallery forest. Similarity in species composition and abundance between deciduous and gallery forests was low. In terms of species diversity, the gallery forest was found more diverse than the deciduous forest using Fisher's diversity index. Both forest communities were characterized by a typical inverse J shape. Therefore, emphasis should be given to the protection of rare species, i.e. as the forests are still under continued human pressure, an immediate action should be taken to conserve the remaining flora.     

基于集成类随机森林方法的神经胶质瘤特征基因选择的研究

神经胶质瘤(glioma)是一种严重的颅内肿瘤疾病,具有高复发率、高死亡率和低治愈率等特点。利用基因微阵列数据识别与神经胶质瘤相关的特征基因,对该疾病的临床诊断和生物医学研究将起到有益的参考和借鉴作用。作者针对神经胶质瘤数据,提出了一种集成类随机森林特征基因选择方法。首先应用有监督奇异值分解对数据进行降维并粗选出基因;其次应用类随机森林特征选择方法选出特征基因。实验结果显示,该方法对分类器的适应性强;对比其他方法,分类率优势明显;更重要的是,在选出的前50个特征基因中有39个基因与神经胶质瘤或肿瘤细胞生物过程存在着密切联系,证实该方法不仅保持了较高的分类率,而且保证了选择的特征基因具有很强的生物学关联意义,具有较高的可行性和实用性。     

Mapping Above- and Below-Ground Carbon Pools in Boreal Forests: The Case for Airborne Lidar

A large and growing body of evidence has demonstrated that airborne scanning light detection and ranging (lidar) systems can be an effective tool in measuring and monitoring above-ground forest tree biomass. However, the potential of lidar as an all-round tool for assisting in assessment of carbon (C) stocks in soil and non-tree vegetation components of the forest ecosystem has been given much less attention. Here we combine the use airborne small footprint scanning lidar with fine-scale spatial C data relating to vegetation and the soil surface to describe and contrast the size and spatial distribution of C pools within and among multilayered Norway spruce (Picea abies) stands. Predictor variables from lidar derived metrics delivered precise models of above- and below-ground tree C, which comprised the largest C pool in our study stands. We also found evidence that lidar canopy data correlated well with the variation in field layer C stock, consisting mainly of ericaceous dwarf shrubs and herbaceous plants. However, lidar metrics derived directly from understory echoes did not yield significant models. Furthermore, our results indicate that the variation in both the mosses and soil organic layer C stock plots appears less influenced by differences in stand structure properties than topographical gradients. By using topographical models from lidar ground returns we were able to establish a strong correlation between lidar data and the organic layer C stock at a stand level. Increasing the topographical resolution from plot averages (~2000 m²) towards individual grid cells (1 m²) did not yield consistent models. Our study demonstrates a connection between the size and distribution of different forest C pools and models derived from airborne lidar data, providing a foundation for future research concerning the use of lidar for assessing and monitoring boreal forest C.