均衡独立样本设计的研究

均衡独立样本设计是一种把独立样本设计与差异显著性检验相结合的试验设计．它基本上可以保证参试的优劣个体在所分样本（或组）间的分布相对均匀一致,使所分样本既具有独立性又具有均衡性,以增强对处理效果反应的灵敏度,提高试验的准确度．     

The Use of Multiple Outcomes Analysis in Neurotoxicity: An Example from the New Zealand Study of Offspring of Mothers Exposed to Methyl Mercury

Neurotoxicity has become an important area in the study of the risk to children of noncarcinogens. Many of the tests used to evaluate neurotoxic effects in neurotoxic studies are psychometric tests to assess cognitive, motor, and perceptual performance in individuals in order to determine the presence of psychological problems, impaired educational achievement, or neurological abnormalities. The underlying assumption of these procedures are that a test performance deficit is indicative of a neurological disability, and that such disability derives from the prenatal exposure to an offending substance. The usual analytic paradigm for discerning environmental impacts is to analyze the scores of a particular subtest in a battery taken separately as the dependent variable with regression analysis, which adjusts for cultural background, educational level, and socioeconomic status. The impact of this analysis is to determine the correlation of each of the psychometric test results to exposure. We show a statistical method to simultaneously analyze the multiple outcomes of many tests using General Estimating Equations (GEE) to determine if a correlation exists between the global measure of all these tests and exposure adjusted for covariates of interest and the correlation among the dependent psychometric variables. We show the application of this method to the results of children whose mothers were exposed to mercury during pregnancy.     

Variation in sclerophylly among Iberian populations of Quercus coccifera L. is associated with genetic differentiation across contrasting environments

Evergreen oaks are an emblematic element of the Mediterranean vegetation and have a leaf phenotype that seems to have remained unchanged since the Miocene. We hypothesise that variation of the sclerophyll phenotype among Iberian populations of Quercus coccifera is partly due to an ulterior process of ecotypic differentiation. We analysed the genetic structure of nine Iberian populations using ISSR fingerprints, and their leaf phenotypes using mean and intracanopy plasticity values of eight morphological (leaf angle, area, spinescence, lobation and specific area) and biochemical traits (VAZ pool, chlorophyll and β-carotene content). Climate and soil were also characterised at the population sites. Significant genetic and phenotypic differences were found among populations and between NE Iberia and the rest of the populations of the peninsula. Mean phenotypes showed a strong and independent correlation with both genetic and geographic distances. Northeastern plants were smaller, less plastic, with smaller, spinier and thicker leaves, a phenotype consistent with the stressful conditions that prevailed in the steppe environments of the refugia within this geographic area during glaciations. These genetic, phenotypic, geographic and environmental patterns are consistent with previously reported palaeoecological and common evidence. Such consistency leads us to conclude that there has been a Quaternary divergence within the sclerophyllous syndrome that was at least partially driven by ecological factors.     

Phenotyping Mouse Pulmonary Function In Vivo with the Lung Diffusing Capacity

The mouse is now the primary animal used to model a variety of lung diseases. To study the mechanisms that underlie such pathologies, phenotypic methods are needed that can quantify the pathologic changes. Furthermore, to provide translational relevance to the mouse models, such measurements should be tests that can easily be done in both humans and mice. Unfortunately, in the present literature few phenotypic measurements of lung function have direct application to humans. One exception is the diffusing capacity for carbon monoxide, which is a measurement that is routinely done in humans. In the present report, we describe a means to quickly and simply measure this diffusing capacity in mice. The procedure involves brief lung inflation with tracer gases in an anesthetized mouse, followed by a 1 min gas analysis time. We have tested the ability of this method to detect several lung pathologies, including emphysema, fibrosis, acute lung injury, and influenza and fungal lung infections, as well as monitoring lung maturation in young pups. Results show significant decreases in all the lung pathologies, as well as an increase in the diffusing capacity with lung maturation. This measurement of lung diffusing capacity thus provides a pulmonary function test that has broad application with its ability to detect phenotypic structural changes with most of the existing pathologic lung models.     

Improved Logrank-Type Tests for Survival Data Using Adaptive Weights

Summary .  For testing for treatment effects with time-to-event data, the logrank test is the most popular choice and has some optimality properties under proportional hazards alternatives. It may also be combined with other tests when a range of nonproportional alternatives are entertained. We introduce some versatile tests that use adaptively weighted logrank statistics. The adaptive weights utilize the hazard ratio obtained by fitting the model of Yang and Prentice (2005,  Biometrika   92 , 1–17). Extensive numerical studies have been performed under proportional and nonproportional alternatives, with a wide range of hazard ratios patterns. These studies show that these new tests typically improve the tests they are designed to modify. In particular, the adaptively weighted logrank test maintains optimality at the proportional alternatives, while improving the power over a wide range of nonproportional alternatives. The new tests are illustrated in several real data examples.     

Multiple McNemar Tests

Summary Methods for performing multiple tests of paired proportions are described. A broadly applicable method using McNemar's exact test and the exact distributions of all test statistics is developed; the method controls the familywise error rate in the strong sense under minimal assumptions. A closed form (not simulation‐based) algorithm for carrying out the method is provided. A bootstrap alternative is developed to account for correlation structures. Operating characteristics of these and other methods are evaluated via a simulation study. Applications to multiple comparisons of predictive models for disease classification and to postmarket surveillance of adverse events are given.     

Patterns of peroxidative ethane emission from submerged rice seedlings indicate that damage from reactive oxygen species takes place during submergence and is not necessarily a post-anoxic phenomenon

Using ethane as a marker for peroxidative damage to membranes by reactive oxygen species (ROS) we examined the injury of rice seedlings during submergence in the dark. It is often expressed that membrane injury from ROS is a post-submergence phenomenon occurring when oxygen is re-introduced after submergence-induced anoxia. We found that ethane production, from rice seedlings submerged for 24–72 h, was stimulated to 4–37 nl gFW⁻¹, indicating underwater membrane peroxidation. When examined a week later the seedlings were damaged or had died. On de-submergence in air, ethane production rates rose sharply, but fell back to less than 0.1 nl gFW⁻¹ h⁻¹ after 2 h. We compared submergence-susceptible and submergence-tolerant cultivars, submergence starting in the morning (more damage) and in the afternoon (less damage) and investigated different submergence durations. The seedlings showed extensive fatality whenever total ethane emission exceeded about 15 nl gFW⁻¹. Smaller amounts of ethane emission were linked to less extensive injury to leaves. Partial oxygen shortage (O₂ levels <1%) imposed for 2 h in gas phase mixtures also stimulated ethane production. In contrast, seedlings under anaerobic gas phase conditions produced no ethane until re-aerated: then a small peak was observed followed by a low, steady ethane production. We conclude that damage during submergence is not associated with extensive anoxia. Instead, injury is linked to membrane peroxidation in seedlings that are partially oxygen deficient while submerged. On return to air, further peroxidation is suppressed within about 2 h indicating effective control of ROS production not evident during submergence itself.     

Gene flow, effective population size and selection at major histocompatibility complex genes: brown trout in the Hardanger Fjord, Norway

Brown trout populations in the Hardanger Fjord, Norway, have declined drastically due to increased exposure to salmon lice from salmonid aquaculture. We studied contemporary samples from seven populations and historical samples (1972 and 1983) from the two largest populations, one of which has declined drastically whereas the other remains stable. We analysed 11 microsatellite loci, including one tightly linked to the UBA gene of the major histocompatibility class I complex (MHC) and another locus linked to the TAP2A gene, also associated with MHC. The results revealed asymmetric gene flow from the two largest populations to the other, smaller populations. This has important conservation implications, and we predict that possible future population recoveries will be mediated primarily by the remaining large population. Tests for selection suggested diversifying selection at UBA, whereas evidence was inconclusive for TAP2A. There was no evidence for temporally fluctuating selection. We assessed the distribution of adaptive divergence among populations. The results showed the most pronounced footprints of selection between the two largest populations subject to the least immigration. We suggest that asymmetric gene flow has an important influence on adaptive divergence and constrains local adaptive responses in the smaller populations. Even though UBA alleles may not affect salmon louse resistance, the results bear evidence of adaptive divergence among populations at immune system genes. This suggests that similar genetic differences could exist at salmon louse resistance loci, thus rendering it a realistic scenario that differential population declines could reflect differences in adaptive variation.