模型选择中交叉验证量的改进

对模型选择中交叉验证量CV进行改进,得到新的验证模型是否合适的准则RCV,RCV包含了CV的信息,并包含了拟合程度,模型中的待估参数个数和样本容量等等,比起AIC,BIC和CV具有更好的稳定性和分辨功能.     

Information-Theoretic Model Selection and Model Averaging for Closed-Population Capture-Recapture Studies

Specification of an appropriate model is critical to valid statistical inference. Given the “true model” for the data is unknown, the goal of model selection is to select a plausible approximating model that balances model bias and sampling variance. Model selection based on information criteria such as AIC or its variant AIC_c, or criteria like CAIC, has proven useful in a variety of contexts including the analysis of open-population capture-recapture data. These criteria have not been intensively evaluated for closed-population capture-recapture models, which are integer parameter models used to estimate population size (N), and there is concern that they will not perform well. To address this concern, we evaluated AIC, AIC_c, and CAIC model selection for closed-population capture-recapture models by empirically assessing the quality of inference for the population size parameter N. We found that AIC-, AIC_c-, and CAIC-selected models had smaller relative mean squared errors than randomly selected models, but that confidence interval coverage on N was poor unless unconditional variance estimates (which incorporate model uncertainty) were used to compute confidence intervals. Overall, AIC and AIC_c outperformed CAIC, and are preferred to CAIC for selection among the closed-population capture-recapture models we investigated. A model averaging approach to estimation, using AIC, AIC_c, or CAIC to estimate weights, was also investigated and proved superior to estimation using AIC-, AIC_c-, or CAIC-selected models. Our results suggested that, for model averaging, AIC or AIC_c should be favored over CAIC for estimating weights.     

Building Hierarchical Models of Avian Distributions for the State of Georgia

Abstract To predict the distributions of breeding birds in the state of Georgia, USA, we built hierarchical models consisting of 4 levels of nested mapping units of decreasing area: 90,000 ha, 3,600 ha, 144 ha, and 5.76 ha. We used the Partners in Flight database of point counts to generate presence and absence data at locations across the state of Georgia for 9 avian species: Acadian flycatcher (Empidonax virescens), brown-headed nuthatch (Sitta pusilla), Carolina wren (Thryothorus ludovicianus), indigo bunting (Passerina cyanea), northern cardinal (Cardinalis cardinalis), prairie warbler (Dendroica discolor), yellow-billed cuckoo (Coccyzus americanus), white-eyed vireo (Vireo griseus), and wood thrush (Hylocichla mustelina). At each location, we estimated hierarchical-level-specific habitat measurements using the Georgia GAP Analysis18 class land cover and other Geographic Information System sources. We created candidate, species-specific occupancy models based on previously reported relationships, and fit these using Markov chain Monte Carlo procedures implemented in OpenBugs. We then created a confidence model set for each species based on Akaike's Information Criterion. We found hierarchical habitat relationships for all species. Three-fold cross-validation estimates of model accuracy indicated an average overall correct classification rate of 60.5%. Comparisons with existing Georgia GAP Analysis models indicated that our models were more accurate overall. Our results provide guidance to wildlife scientists and managers seeking predict avian occurrence as a function of local and landscape-level habitat attributes.     

Size and heterogeneity rather than landscape context determine plant species richness in semi‐natural grasslands

Question: Is plant diversity in fragmented semi‐natural grasslands related to present and historical landscape context? Location: Southern Sweden. Methods: Plant diversity was described at 30 semi‐natural grassland sites in terms of total and specialist plant species richness at the site and species density at different scales (0.5–10 m²). These measures are commonly used to assess conservation value of semi‐natural grasslands. Landscape context was measured as contemporary connectivity to other semi‐natural grasslands, historical connectivity 50 years ago, amount of linear elements potentially suitable for dispersal (road verges, power line clearings), and amount of forest (inverse of the openness of the landscape). Results: The diversity measures were generally correlated with each other, implying that species richness in a subset of the grassland can predict the total richness. Plant species density at three scales (0.5 m², 10 m² and total) was related to the landscape context using an information theoretic approach. Results showed that total species richness increased with increased size of grasslands, contrary to earlier diversity studies in semi‐natural grasslands. Larger grasslands were more heterogeneous than smaller grasslands, and this is a likely reason for the species‐area relationship. Heterogeneity was also of high importance at the smaller scales (0.5 m², 10 m²). With increased amount of forest, total species richness increased but species density on 10 m² decreased. There was no influence of connectivity in either the contemporary or the historical landscape, contrary to previous studies. Conclusions: Grassland size and heterogeneity are of greater importance for plant diversity in semi‐natural grassland, than grassland connectivity in the landscape.     

Moss and vascular plant indices in Ohio wetlands have similar environmental predictors

Mosses and vascular plants have been shown to be reliable indicators of wetland habitat delineation and environmental quality. Knowledge of the best ecological predictors of the quality of wetland moss and vascular plant communities may determine if similar management practices would simultaneously enhance both populations. We used Akaike's Information Criterion to identify models predicting a moss quality assessment index (MQAI) and a vascular plant index of biological integrity based on floristic quality (VIBI-FQ) from 27 emergent and 13 forested wetlands in Ohio, USA. The set of predictors included the six metrics from a wetlands disturbance index (ORAM) and two landscape development intensity indices (LDIs). The best single predictor of MQAI and one of the predictors of VIBI-FQ was an ORAM metric that assesses habitat alteration and disturbance within the wetland, such as mowing, grazing, and agricultural practices. However, the best single predictor of VIBI-FQ was an ORAM metric that assessed wetland vascular plant communities, interspersion, and microtopography. LDIs better predicted MQAI than VIBI-FQ, suggesting that mosses may either respond more rapidly to, or recover more slowly from, anthropogenic disturbance in the surrounding landscape than vascular plants. These results supported previous predictive studies on amphibian indices and metrics and a separate vegetation index, indicating that similar wetland management practices may result in qualitatively the same ecological response for three vastly different wetland biological communities (amphibians, vascular plants, and mosses).     

Independent effects of song quality and experience with photostimulation on expression of the immediate,early gene ZENK (EGR‐1) in the auditory telencephalon of female European starlings

Age influences behavioral decisions such as reproductive timing and effort. In photoperiodic species, such age effects may be mediated, in part, by the individual's age‐accrued experience with photostimulation. In female European starlings (Sturnus vulgaris) that do not differ in age, experimental manipulation of photostimulation experience (photoexperience) affects hypothalamic, pituitary, and gonadal activity associated with reproductive development. Does photoexperience also affect activity in forebrain regions involved in processing a social cue, the song of males, which can influence mate choice and reproductive timing in females? Female starlings prefer long songs over short songs in a mate‐choice context, and, like that in other songbird species, their auditory telencephalon plays a major role in processing these signals. We manipulated the photoexperience of female starlings, photostimulated them, briefly exposed them to either long or short songs, and quantified the expression of the immediate‐early gene ZENK (EGR‐1) in the caudomedial nidopallium as a measure of activity in the auditory telencephalon. Using an information theoretic approach, we found higher ZENK immunoreactivity in females with prior photostimulation experience than in females experiencing photostimulation for the first time. We also found that long songs elicited greater ZENK immunoreactivity than short songs did. We did not find an effect of the interaction between photoexperience and song length, suggesting that photoexperience does not affect forebrain ZENK‐responsiveness to song quality. Thus, photoexperience affects activity in an area of the forebrain that processes social signals, an effect that we hypothesize mediates, in part, the effects of age on reproductive decisions in photoperiodic songbirds. © 2009 Wiley Periodicals, Inc. Develop Neurobiol 2009     

Model-free Analysis of Protein Dynamics: Assessment of Accuracy and Model Selection Protocols Based on Molecular Dynamics Simulation

The popular model-free approach to analyze NMR relaxation measurements has been examined using artificial amide (15)N relaxation data sets generated from a 10 nanosecond molecular dynamics trajectory of a dihydrofolate reductase ternary complex in explicit water. With access to a detailed picture of the underlying internal motions, the efficacy of model-free analysis and impact of model selection protocols on the interpretation of NMR data can be studied. In the limit of uncorrelated global tumbling and internal motions, fitting the relaxation data to the model-free models can recover a significant amount of quantitative information on the internal dynamics. Despite a slight overestimation, the generalized order parameter is quite accurately determined. However, the model-free analysis appears to be insensitive to the presence of nanosecond time scale motions with relatively small magnitude. For such cases, the effective correlation time can be significantly underestimated. As a result, proteins appear to be more rigid than they really are. The model selection protocols have a major impact on the information one can reliably obtain. The commonly employed protocol based on step-up hypothesis testing has severe drawbacks of oversimplification and underfitting. The consequences are that the order parameter is more severely overestimated and the correlation time more severely underestimated. Instead, model selection based on Bayesian Information Criteria (BIC), recently introduced to the model-free analysis by d'Auvergne and Gooley (2003), provides a better balance between bias and variance. More appropriate models can be selected, leading to improved estimate of both the order parameter and correlation time. In addition, the computational cost is significantly reduced and subjective parameters such as the significance level are unnecessary.     

Fire,landscape change and models of small mammal habitat suitability at multiple spatial scales

Fire is an important process in many ecosystems, but inappropriate fire regimes can adversely affect biodiversity. We identified a naturally flammable heathy woodland ecosystem where the use of planned fire had increased the extent of older vegetation, and quantified the abundance of two small native mammals in this landscape (silky mouse Pseudomys apodemoides and heath rat P. shortridgei). We defined four time‐since‐fire (TSF) categories representing a 2‐ to 55‐year post‐fire sequence and, on the basis of a habitat accommodation model, predicted that both species would select younger age‐classes over older ones. We also predicted that (i) much of the variance in vegetation structure would remain unexplained by TSF and (ii) statistical models of mammal abundance and occupancy including structural variables as predictors would be better than models including TSF. Pseudomys apodemoides selected 17‐ to 23‐year‐old sites, while there was no evidence that P. shortridgei selected a particular TSF category, findings that were inconsistent with our predictions. In line with our predictions, relatively large portions of the variance in vegetation structure remained unexplained by TSF (adjustedr² for four structural variables: 0.24, 0.29, 0.35 and 0.57), and in three of four cases there was strong evidence that statistical models of mammal abundance and occupancy including structural variables were better than those including TSF. At the site scale (hectares), P. shortridgei abundance was positively related to the cover of dead material at the base of Xanthorrhoea plants and at the trap scale (metres), the trapability of both species was significantly related to vegetation volume at 0–20 cm. Our findings suggest that TSF may not be a good proxy for either vegetation structure or species abundance/occupancy.     

RELATING GROUSE NEST SUCCESS AND CORVID DENSITY TO HABITAT: A MULTI-SCALE APPROACH

Abstract: Predators are the major cause of nest failure for prairie grouse, and corvids are widespread generalist predators that exploit land altered by humans where grouse are found. We studied how human-caused habitat change affected predator and prey by using habitat variables to model nest selection, corvid density, and nest success for sharp-tailed grouse (Tympanuchus phasianellus) in Alberta, Canada, 1999–2001. Habitat was quantified over a range of extents (radius of observation) from 2 to 2,265 m. We predicted that habitat features associated with corvid density at broad extents would also relate to grouse nest success, and that nesting cover and the presence of avian predator perch sites would be important at smaller extents. Corvid density was higher in landscapes with higher proportions of crop and sparse grassland (1,600-m extent). Conversely, nest success was markedly higher (≥4 times) in landscapes with <10% crop and <35% crop and sparse grassland (aggregated) at broad extents (1,600 m). Moreover, nests were 8 times more likely to succeed in landscapes with lower relative corvid densities (<3 vs. ≥3 corvids/km²). At smaller scales, nests were more likely to succeed with greater heights of concealment cover within 50-m of nests. Land managers can likely improve nest success for grouse in grassland systems by targeting concealment cover heights of at least 13 cm measured over a 50-m extent, and focusing efforts in landscapes with <10% crop and <35% crop and sparse grassland (1,600-m extent).     

丹参酮对于阿尔茨海默症大鼠颞叶iNOS，MMP-2 表达的 影响及机理研究

目的：应用中药丹参酮(tanshinone II A,Tan II A)治疗AD大鼠,观察TanⅡ A 干预前后,AD大鼠学习记忆、颞叶中诱导型一氧 化氮合成酶(iNOS)、基质金属蛋白酶（MMP-2）表达的变化。方法：采用beta- 淀粉样蛋白（A beta）定向注射法建立AD大鼠模型,并使用 Tan II A 干预,通过避暗测试、real-time PCR和Western Blot 分别观察大鼠学习记忆能力、大鼠颞叶MMP-2、iNOS 两者的mRNA 及蛋白表达的变化。应用SPSS13.0 进行统计学分析。结果：与假手术组相比,AD 组的平均潜伏期缩短（P＜0.01）,平均错误次数 增加（P＜0.01）,差异均有统计学意义。颞叶内iNOS、MMP-2 mRNA 表达均显著增高（P＜ 0.01, P＜0.01）;两蛋白的表达均显著增 高（P＜0.01, P＜0.01）。与AD组相比,Tan IIA 组的平均潜伏期延长（P＜0.01）,平均错误次数减少（P＜0.01）,差异均有统计学意 义。颞叶内iNOS、MMP-2 mRNA表达均显著下降（P＜0.05, P＜0.05）,两蛋白的表达均显著下降（P＜0.01, P＜0.01）。结论：Tan II A 干预可显著降低AD 大鼠颞叶中iNOS、MMP-2 mRNA及蛋白的表达,显著改善AD 大鼠的学习记忆能力。其作用机制可能是 通过降低A-beta诱导的iNOS及MMP-2 的表达,抑制氧化应激损伤来完成。     

腹腔注射四氯化碳辅以饮食改良制备犬慢性肝纤维化模型

探讨一种经典的、与人类慢性肝纤维化——肝硬化进程相似的大动物模型建立方法.采用15只成年健康中华田园犬:对照组5只,使用生理盐水腹腔注射及正常饲料干预;实验组10只,采用小剂量四氯化碳(CCl4)腹腔注射及高脂饮食干预.间断采血检测肝功能、血清学肝纤维化指标,B超引导下定期肝穿活检作HE染色和Masson染色.56周后所有实验动物成功诱导至不同程度的肝纤维化.丙氨酸氨基转移酶、门冬氨酸转氨酶、球蛋白、总胆红素、γ-谷氨酰基转肽酶以及透明质酸与对照组相比,差异均有统计学意义(P<0.05).联合应用小剂量CCl4和单纯高脂饮食可成功诱导犬慢性肝纤维化-肝硬化模型,该方法动物存活率和成模率均较高,可为后续的相关研究提供更好的平台.     

Interferon induced with helicase C domain 1 (IFIH1): Trends on helicase domain and type 1 diabetes onset

The Interferon-induced with helicase C domain 1 (IFIH1) gene has been hypothesized as involved in the viral etiology of type 1 diabetes (T1D) and other autoimmune disorders, since it is implicated in viral recognition. In our study we analyzed IFIH1 rs6432714 and rs10930046 SNPs in T1D patients stratified for the presence of celiac disease and autoimmune thyroid disease. No association with susceptibility to develop the diseases was found. The rs6432714, a tag-SNP that represents part of helicase domain of IF1H1 protein showed a trend of association only with T1D development (P > 0.025 after Bonferroni adjustment) in log-additive model (OR = 1.45, P = 0.0365, power = 0.99), indicating that helicase domain of IFIH1 protein could be associated with the susceptibility to T1D.     

Recruitment and diversification of an ecdysozoan family of neuropeptide hormones for black widow spider venom expression

Venoms have attracted enormous attention because of their potent physiological effects and dynamic evolution, including the convergent recruitment of homologous genes for venom expression. Here we provide novel evidence for the recruitment of genes from the Crustacean Hyperglycemic Hormone (CHH) and arthropod Ion Transport Peptide (ITP) superfamily for venom expression in black widow spiders. We characterized latrodectin peptides from venom gland cDNAs from the Western black widow spider (Latrodectus hesperus), the brown widow (Latrodectus geometricus) and cupboard spider (Steatoda grossa). Phylogenetic analyses of these sequences with homologs from other spider, scorpion and wasp venom cDNAs, as well as CHH/ITP neuropeptides, show latrodectins as derived members of the CHH/ITP superfamily. These analyses suggest that CHH/ITP homologs are more widespread in spider venoms, and were recruited for venom expression in two additional arthropod lineages. We also found that the latrodectin 2 gene and nearly all CHH/ITP genes include a phase 2 intron in the same position, supporting latrodectin's placement within the CHH/ITP superfamily. Evolutionary analyses of latrodectins suggest episodes of positive selection along some sequence lineages, and positive and purifying selection on specific codons, supporting its functional importance in widow venom. We consider how this improved understanding of latrodectin evolution informs functional hypotheses regarding its role in black widow venom as well as its potential convergent recruitment for venom expression across arthropods.     

The complete mitochondrial genome of Temminck's ground pangolin (Smutsia temminckii; Smuts, 1832) and phylogenetic position of the Pholidota (Weber, 1904)

Temmincki's ground pangolin is primarily a nocturnal mammal belonging to the order Pholidota. The body is covered in hard overlapping scales and these animals find refuge in burrows, feeding only on termites and ants. In this study, the whole mtDNA of Temmincki's ground pangolin was sequenced and the phylogenetic position of Pholidota determined within Eutheria, using whole mtDNA sequences from various representative species. The results indicate that the whole mtDNA of Temmincki's ground pangolin is 16,559 bp long and shared some similarities with the whole mtDNA of the back-bellied tree pangolin and the Chinese pangolin. Phylogenetic analysis indicate that the order Pholidota is closely related and share a recent common ancestor with the order Carnivora rather than with the ant/insect eating order Xenarthra and the group Afrotheria. A time measured phylogeny of Pholidota estimated a split from Carnivora at around 87 mya, followed by a split of the African pangolins from their Asian counterparts such as the Chinese pangolin at around 47 mya. This suggests a Laurasian origin and convergent evolution of the Pholidota with respect to Xenarthra and Afrotheria.