The scope for improving the design of laboratory animal experiments.

The factors which need to be taken into account in designing a 'good' experiment are reviewed. Such an experiment should be unbiased, have high precision, a wide range of applicability, it should be simple, and there should be a means of quantifying uncertainty (Cox 1958). The relative precision due to the use of randomized block designs was found to range from 96% to 543% in 5 experiments involving 30 variables. However, a survey of 78 papers published in two toxicology journals showed that such designs were hardly used. Similarly, designs in which more than one factor was varied simultaneously ('factorial designs') were only used in 9% of studies, though interactions between variables such as dose and strain of animal may be common, so that single factor experiments could be misleading. The consequences of increased within-group variability due to infection and genetic segregation were quantified using data published by G?rtner (1990). Both substantially reduced precision, but toxicologists continue to use non-isogenic laboratory animals, leading to experiments with a lower level of precision than is necessary. It is concluded that there is scope for improving the design of animal experiments, which could lead to a reduction in animal use. People using animals should be required to take formal training courses which include sessions on experimental design in order to minimize animal use and to increase experimental efficiency.     

Sample Power and ExpDesign: tools for improving design of animal experiments

Proper experimental design, involving the correct number of animals, should be a basic skill for any scientist working with animals. The authors describe a university-developed and freely available tutorial program and an interactive computer-assisted learning program, both of which guide students through the steps necessary for designing animal experiments and estimating optimal sample sizes.     

Rodent quarantine programs: purpose, principles, and practice

In animal research, validity and reproducibility of data are critically influenced by the microbial status of the experimental animals. One of the most crucial aspects of assuring quality in animal research is providing research personnel with confidence that experimental results will not be invalidated due to interference caused by infectious disease. An effective quarantine program is essential to providing this assurance. Quarantine programs are generally instituted to prevent the introduction of rodent pathogens into established specific-pathogen-free colonies in a facility. Therefore, programs should be designed to isolate newly acquired rodents until their health status can be determined and to maximize the probability that microorganisms of interest will be detected before the animals are introduced into (and thus, could potentially contaminate) established colonies. Important principles that are critical to designing an effective quarantine program will be discussed here, as will the practical implementation of these principles. Although quarantine programs may be costly in terms of time and effort, these costs must be balanced against the potential costs of disease outbreaks that could invalidate long-term studies, alter normal biological baselines, and cause the loss or necessitate re-derivation of rare or valuable strains of rodents. Reducing the incidence of quarantine failures through appropriate program design and implementation helps to maintain the confidence of research personnel in the value of quarantine programs and in our competence as specialists in laboratory animal management and as partners in the research process.     

Guidelines for the design and statistical analysis of experiments using laboratory animals

For ethical and economic reasons, it is important to design animal experiments well, to analyze the data correctly, and to use the minimum number of animals necessary to achieve the scientific objectives---but not so few as to miss biologically important effects or require unnecessary repetition of experiments. Investigators are urged to consult a statistician at the design stage and are reminded that no experiment should ever be started without a clear idea of how the resulting data are to be analyzed. These guidelines are provided to help biomedical research workers perform their experiments efficiently and analyze their results so that they can extract all useful information from the resulting data. Among the topics discussed are the varying purposes of experiments (e.g., exploratory vs. confirmatory); the experimental unit; the necessity of recording full experimental details (e.g., species, sex, age, microbiological status, strain and source of animals, and husbandry conditions); assigning experimental units to treatments using randomization; other aspects of the experiment (e.g., timing of measurements); using formal experimental designs (e.g., completely randomized and randomized block); estimating the size of the experiment using power and sample size calculations; screening raw data for obvious errors; using the t-test or analysis of variance for parametric analysis; and effective design of graphical data.     

Determining Parameters of the Numerical Response

The numerical response, the change in specific growth rate with food concentration, is a fundamental component of many aquatic microbial studies. Accurately and precisely determining the parameters of this response is essential to obtain useful data for both aut- and synecological studies. In this work we emphasize four points that are often ignored in designing numerical response experiments: (1) the inclusion of subthreshold concentrations (i.e., where growth rate is negative) in the experimental design; (2) an appropriate allocation of effort, i.e., the superiority of choosing more individual prey concentrations rather than replicating fewer; (3) the potential superiority of replicating experiments rather than simply replicating treatment in a single experiment; and (4) the placement of most measurements near the lower end of the concentration gradient, well below the asymptote, possibly following a geometric progression. We illustrate the first point by examining a small subset of published data on planktonic oligotrich ciliates and then, using a Monte Carlo simulation, rigorously evaluate the experimental design, supporting the remaining points.     

Reduction through education: the insight of a trainer

One of the articles contained within European Council Directive 86/609/EEC states that "Persons who carry out experiments or take part in them, and persons who take care of animals used for experiments, including duties of a supervisory nature, shall have appropriate training". In effect, this article stipulates that only competent individuals are allowed to work with laboratory animals. At least three groups of individuals can be identified with different responsibilities toward experimental animals: animal technicians, scientists, and veterinarians/animal welfare officers. The responsibilities and duties of the individuals within each of these categories differ. This paper focuses on the training of scientists. The scientist designs, and often also performs, animal experiments. Therefore, scientists must be educated to develop an attitude of respect toward laboratory animals, and must be trained so that, if an experiment must be performed with animals, it is designed according to the highest possible scientific and ethical standards. In The Netherlands, the law stipulates that scientists intending to work with animals must have completed a course in laboratory animal science. This compulsory course started in 1986. The Department of Laboratory Animal Science at Utrecht University is responsible for the national coordination of this course. Participants must have an academic degree (at the level of MSc) in one of the biomedical sciences, such as biology, medicine or veterinary medicine. Although the course is an intensive 3-week, 120-hour long course, which covers both technical and ethical aspects of laboratory animal experimentation, it cannot provide full competence. It is designed to provide sufficient basic training and knowledge to enable students to design animal experiments, and to develop an attitude that will be conducive to the implementation of the Three Rs. However, full competence will always require further training that can only be acquired as a result of practical experience gained while working in the field of laboratory animal research. Evaluations subsequent to the course have revealed that more than 98% of the students regard the course as indispensable for all scientists working in a research area where animal experiments are performed. They agree that the course not only contributes to the quality of experiments and to the welfare of animals, but also to a decrease in the number of animals used in experiments.     

The role of the statistician in the Scottish Agricultural and Biological Research Institutes

Several of the Scottish Agricultural and Biological Research Institutes carry out research on domestic animal health and welfare. Statistical services are provided by Biomathematics & Statistics Scotland, a sister research organisation. At one of these institutes, a statistician has been an integral member of the animal experiments and ethics committee for over 10 years, and each animal experiment is examined by the committee statistician as part of the review process. This paper will describe this review process, and then discuss those areas in which statistical advice has had most impact in the reduction of animal numbers. It is suggested that most benefit does not come from simple sample-size calculations, but rather from the application of the principles of good experimental design and close collaboration between the scientist and the statistician in the design and analysis of experiments. The final conclusion is that scientists welcome constructive, long-term statistical input, although budgetary issues can prove to be a barrier.     

A rat model for hyperkalemia

It is often necessary to have a small animal model for hyperkalemia for use in electrolyte and acid base experiments. In reviewing the literature, we found a paucity of such animal models, especially for acute hyperkalemia. We have had difficulty in inducing acute hyperkalemia in rats using potassium chloride alone either intravenously or intraperitoneally and felt the need for an easily reproducible small animal model for hyperkalemia. We gave experimental animals a combination of intraperitoneal amiloride 3 mg/kg and potassium chloride 2 meq/kg in two divided doses while control animals received only the potassium chloride. Initial serum potassiums were similar but at 2 hr, the experimental group had significantly higher serum potassium levels which were sustained throughout the 8 hr of the experiment. Arterial blood gas revealed no significant difference in blood pH values at all time points during the experiment. We conclude that the combination of amiloride and potassium chloride is useful to produce acute hyperkalemia in rats and that this hyperkalemia is sustained beyond 6 hr. This model is convenient for use in metabolic experiments requiring the use of acutely hyperkalemic rats.     

Segmented correspondence curve regression for quantifying covariate effects on the reproducibility of high-throughput experiments

High-throughput biological experiments are essential tools for identifying biologically interesting candidates in large-scale omics studies. The results of a high-throughput biological experiment rely heavily on the operational factors chosen in its experimental and data-analytic procedures. Understanding how these operational factors influence the reproducibility of the experimental outcome is critical for selecting the optimal parameter settings and designing reliable high-throughput workflows. However, the influence of an operational factor may differ between strong and weak candidates in a high-throughput experiment, complicating the selection of parameter settings. To address this issue, we propose a novel segmented regression model, called segmented correspondence curve regression, to assess the influence of operational factors on the reproducibility of high-throughput experiments. Our model dissects the heterogeneous effects of operational factors on strong and weak candidates, providing a principled way to select operational parameters. Based on this framework, we also develop a sup-likelihood ratio test for the existence of heterogeneity. Simulation studies show that our estimation and testing procedures yield well-calibrated type I errors and are substantially more powerful in detecting and locating the differences in reproducibility across workflows than the existing method. Using this model, we investigated an important design question for ChIP-seq experiments: How many reads should one sequence to obtain reliable results in a cost-effective way? Our results reveal new insights into the impact of sequencing depth on the binding-site identification reproducibility, helping biologists determine the most cost-effective sequencing depth to achieve sufficient reproducibility for their study goals.     

国际前沿

高伟坚博士,英籍华人,1986年出生于英国。18岁以全优成绩考取曼彻斯特大学生命科学院。2008年以优等生荣誉毕业,获学士学位。2012年,获得该校博士学位。2004～2012年,先后获得“默沙东杰出学术成就奖”、英联邦帕金森氏病学会的“杰出研究和展示奖”、“曼彻斯特领袖项目”金奖、双重博士奖学金。他从小热爱生物医学,经过系统求学过程及严格的科研训练,在科研课题设计和科技论文撰写方面,表现出不菲的成绩。毕业后仅仅两年,便发表有影响的科技论文4篇。并参与Springer 组织的‘L-DOPA-induced dyskinesia in Parkinson ’ s disease ’一书的编撰工作。同时,还组织申请国际合作课题两项,参研课题四项。高博士现供职于法国波尔多第二大学,在华开展联合研发工作期间,多次表达了“自己作为华人,愿意为祖国生物医学发展贡献自己力量的想法”。也深知华人科学家因为语言障碍,而在科技论文发表和科研课题申请上遇到的重重困难。经我刊编委推荐及编辑部与部分专家讨论商议,特聘请高博士为两刊特约通讯员,开设专栏,希望高博士从信息获取、论文撰写、课题设计等方面开展工作,并定期向我刊介绍业内国际前沿动态,为我刊读者扩大视角。 本期推出神经科学研究中的动物选择和模型制作,欢迎读者就相关内容展开互动。     

"Statistics" is not a sausage machine: a statistician's viewpoint and some comments on experimental design

Any experiment involving the use of animals which is not well-planned, meticulously carried out, and scrupulously analysed, is unethical. Planning, or good experimental design, followed by analysis appropriate for the design, will help to ensure the optimal use of animals. Thus, collaboration between biologist and statistician, especially at the planning and analysis stages, is one of the best ways of achieving an ethical and successful experiment. However, genuine communication is necessary for any collaboration, and this requires time and patience, on the part of both biologist and statistician. Although the three fundamental principles of experimental design, replication, randomisation and local control, are straightforward in theory, there is substantial scope for misunderstanding and misinterpretation in practice. Each experiment presents unique and interacting biological and statistical problems, and both the right design and the correct analysis should be decided on a case-by-case basis.     

Role of ancillary variables in the design,analysis, and interpretation of animal experiments

During the course of an experiment using animals, many variables (e.g., age, body weight at several times, food and water consumption, hematology, and clinical biochemistry) and other characteristics are often recorded in addition to the primary response variable(s) specified by the experimenter. These additional variables have an important role in the design and interpretation of the experiment. They may be formally incorporated into the design and/or analysis and thus increase precision and power. However, even if these variables are not incorporated into the primary statistical design or into the formal analysis of the experiment, they may nevertheless be used in an ancillary or exploratory way to provide valuable information about the experiment, as shown by various examples. Used in this way, ancillary variables may improve analysis and interpretation by providing an assessment of the randomization process and an approach to the identification of outliers, lead to the generation of new hypotheses, and increase generality of results or account for differences in results when compared across different experiments. Thus, appropriate use of additional variables may lead to reduction in the number of animals required to achieve the aims of the experiment and may provide additional scientific information as an extra benefit. Unfortunately, this type of information is sometimes effectively discarded because its potential value is not recognized. Guidelines for use of animals include, in addition to the obligation to follow humane procedures, the obligation to use no more animals than necessary. Ethical experimental practice thus requires that all information be properly used and reported.     

Behavioral phenotyping enhanced--beyond (environmental) standardization

It is basic biology that the phenotype of an animal is the product of a complex and dynamic interplay between nature (genotype) and nurture (environment). It is far less clear, however, how this might translate into experimental design and the interpretation of animal experiments. Animal experiments are a compromise between modelling real world phenomena with maximal validity (complexity) and designing practicable research projects (abstraction). Textbooks on laboratory animal science generally favour abstraction over complexity. Depending on the area of research, however, abstraction can seriously compromise information gain, with respect to the real world phenomena an experiment is designed to model. Behavioral phenotyping of mouse mutants often deals with particularly complex manifestations of life, such as learning, memory or anxiety, that are strongly modulated by environmental factors. A growing body of evidence indicates that current approaches to behavioral phenotyping might often produce results that are idiosyncratic to the study in which they were obtained, because the interactive nature of genotype-environment relationships underlying behavioral phenotypes was not taken into account. This paper argues that systematic variation of genetic and environmental backgrounds, instead of excessive standardization, is needed to control the robustness of the results and to detect biologically relevant interactions between the mutation and the genetic and environmental background of the animals.     

Cross-contamination with tamoxifen induces transgene expression in non-exposed inducible transgenic mice

Inducible transgenic mouse models that impose a constraint on both temporal and spatial expression of a given transgene are invaluable. These animals facilitate experiments that can address the role of a specific cell or group of cells within an animal or in a particular window of time. A common approach to achieve inducibility involves the site-specific recombinase 'Cre', which is linked to a modified version of one of various steroid hormone-binding domains. Thus, the expression of Cre is regulated such that a functional nuclear transgene product can only be generated with the addition of an exogenous ligand. However, critical requirements of this system are that the nuclear localization of the transgene product be tightly regulated, that the dosage of the inducing agent remains consistent among experimental animals and that the transgene cassette cannot express in the absence of the inducing agent. We used the Cre ER(T2) cassette, which is regulated by the addition of the estrogen antagonist tamoxifen to determine whether cross-contamination of tamoxifen between animals housed together can be a significant source of spurious results. We found that cross-contamination of exogenous tamoxifen does occur. It occurred in all animals tested. We suggest that the mechanism of contamination is through exposure to tamoxifen in the general environment and/or to coprophagous behavior. These results have important implications for the interpretation and design of experiments that use 'inducible' transgenic animals.     

Use of factorial designs to optimize animal experiments and reduce animal use

Optimization of experiments, such as those used in drug discovery, can lead to useful savings of scientific resources. Factors such as sex, strain, and age of the animals and protocol-specific factors such as timing and methods of administering treatments can have an important influence on the response of animals to experimental treatments. Factorial experimental designs can be used to explore which factors and what levels of these factors will maximize the difference between a vehicle control and a known positive control treatment. This information can then be used to design more efficient experiments, either by reducing the numbers of animals used or by increasing the sensitivity so that smaller biological effects can be detected. A factorial experimental design approach is more effective and efficient than the older approach of varying one factor at a time. Two examples of real factorial experiments reveal how using this approach can potentially lead to a reduction in animal use and savings in financial and scientific resources without loss of scientific validity.     

Stimulus design for model selection and validation in cell signaling

Mechanism-based chemical kinetic models are increasingly being used to describe biological signaling. Such models serve to encapsulate current understanding of pathways and to enable insight into complex biological processes. One challenge in model development is that, with limited experimental data, multiple models can be consistent with known mechanisms and existing data. Here, we address the problem of model ambiguity by providing a method for designing dynamic stimuli that, in stimulus–response experiments, distinguish among parameterized models with different topologies, i.e., reaction mechanisms, in which only some of the species can be measured. We develop the approach by presenting two formulations of a model-based controller that is used to design the dynamic stimulus. In both formulations, an input signal is designed for each candidate model and parameterization so as to drive the model outputs through a target trajectory. The quality of a model is then assessed by the ability of the corresponding controller, informed by that model, to drive the experimental system. We evaluated our method on models of antibody–ligand binding, mitogen-activated protein kinase (MAPK) phosphorylation and de-phosphorylation, and larger models of the epidermal growth factor receptor (EGFR) pathway. For each of these systems, the controller informed by the correct model is the most successful at designing a stimulus to produce the desired behavior. Using these stimuli we were able to distinguish between models with subtle mechanistic differences or where input and outputs were multiple reactions removed from the model differences. An advantage of this method of model discrimination is that it does not require novel reagents, or altered measurement techniques; the only change to the experiment is the time course of stimulation. Taken together, these results provide a strong basis for using designed input stimuli as a tool for the development of cell signaling models.     

A new pooling strategy for high-throughput screening: the Shifted Transversal Design

Background  

In binary high-throughput screening projects where the goal is the identification of low-frequency events, beyond the obvious issue of efficiency, false positives and false negatives are a major concern. Pooling constitutes a natural solution: it reduces the number of tests, while providing critical duplication of the individual experiments, thereby correcting for experimental noise. The main difficulty consists in designing the pools in a manner that is both efficient and robust: few pools should be necessary to correct the errors and identify the positives, yet the experiment should not be too vulnerable to biological shakiness. For example, some information should still be obtained even if there are slightly more positives or errors than expected. This is known as the group testing problem, or pooling problem.     

Balancing animal research with animal well-being: establishment of goals and harmonization of approaches

A resource is provided for the creation of an institutional program that balances the scientific mission of an institution with the well-being of the animals used in support of the research. The concept of harmonizing scientific goals with animal well-being was first suggested in the early part of the twentieth century and later revitalized in the literature of the 1950s. Harmonization can best be achieved through the promotion of a team initiative. The team should include, at a minimum, the scientist, veterinarian, institutional animal care and use committee, and animal care staff. It is the responsibility of this animal research team to promote and balance the generation of scientifically valid data with animal well-being. The team must strive to minimize or eliminate non-protocol variables that could adversely affect the validity and repeatability of the experimental data. Good experimental design coupled with excellent communication between team members can often minimize or eliminate many variables and result in both better science and animal well-being. To ensure the scientific validity of experimental data, scientists must be aware of the complex nature of the environment in which their animals are maintained. To ensure repeatablity of an experiment, scientists must document and publish both the inanimate and social environments in which their animals are housed. Better documentation of environmental variables and their correlation with experimental results will promote critical knowledge about the relationships between an animal's environment, its well-being, and science.     

Anesthesia and other considerations for in vivo imaging of small animals

The use of small animal imaging is increasing in biomedical research thanks to its ability to localize altered biochemical and physiological processes in the living animal and to follow these processes longitudinally and noninvasively. In contrast to human studies, however, imaging of small animals generally requires anesthesia, and anesthetic agents can have unintended effects on animal physiology that may confound the results of the imaging studies. In addition, repeated anesthesia, animal preparation for imaging, exposure to ionizing radiation, and the administration of contrast agents may affect the processes under study. We discuss this interplay of factors for small animal imaging in the context of four common imaging modalities for small animals: positron emission tomography (PET) and single photon emission computed tomography (SPECT), computed tomography (CT), magnetic resonance imaging (MRI), and optical imaging. We discuss animal preparation for imaging, including choice of animal strain and gender, the role of fasting and diet, and the circadian cycle. We review common anesthesias used in small animal imaging, such as pentobarbital, ketamine/xylazine, and isoflurane, and describe techniques for monitoring the respiration and circulation of anesthetized animals that are being imaged as well as developments for imaging conscious animals. We present current imaging literature exemplifying how anesthesia and animal handling can influence the biodistribution of PET tracers. Finally, we discuss how longitudinal imaging studies may affect animals due to repeated injections of radioactivity or other substrates and the general effect of stress on the animals. In conclusion, there are many animal handling issues to consider when designing an imaging experiment. Reproducible experimental conditions require clear, consistent reporting, in the study design and throughout the experiment, of the animal strain and gender, fasting, anesthesia, and how often individual animals were imaged.     

VDA, a method of choosing a better algorithm with fewer validations

The multitude of bioinformatics algorithms designed for performing a particular computational task presents end-users with the problem of selecting the most appropriate computational tool for analyzing their biological data. The choice of the best available method is often based on expensive experimental validation of the results. We propose an approach to design validation sets for method comparison and performance assessment that are effective in terms of cost and discrimination power.Validation Discriminant Analysis (VDA) is a method for designing a minimal validation dataset to allow reliable comparisons between the performances of different algorithms. Implementation of our VDA approach achieves this reduction by selecting predictions that maximize the minimum Hamming distance between algorithmic predictions in the validation set. We show that VDA can be used to correctly rank algorithms according to their performances. These results are further supported by simulations and by realistic algorithmic comparisons in silico.VDA is a novel, cost-efficient method for minimizing the number of validation experiments necessary for reliable performance estimation and fair comparison between algorithms.Our VDA software is available at http://sourceforge.net/projects/klugerlab/files/VDA/