Validation of computational methods in genomics

High-throughput technologies for genomics provide tens of thousands of genetic measurements, for instance, gene-expression measurements on microarrays, and the availability of these measurements has motivated the use of machine learning (inference) methods for classification, clustering, and gene networks. Generally, a design method will yield a model that satisfies some model constraints and fits the data in some manner. On the other hand, a scientific theory consists of two parts: (1) a mathematical model to characterize relations between variables, and (2) a set of relations between model variables and observables that are used to validate the model via predictive experiments. Although machine learning algorithms are constructed to hopefully produce valid scientific models, they do not ipso facto do so. In some cases, such as classifier estimation, there is a well-developed error theory that relates to model validity according to various statistical theorems, but in others such as clustering, there is a lack of understanding of the relationship between the learning algorithms and validation. The issue of validation is especially problematic in situations where the sample size is small in comparison with the dimensionality (number of variables), which is commonplace in genomics, because the convergence theory of learning algorithms is typically asymptotic and the algorithms often perform in counter-intuitive ways when used with samples that are small in relation to the number of variables. For translational genomics, validation is perhaps the most critical issue, because it is imperative that we understand the performance of a diagnostic or therapeutic procedure to be used in the clinic, and this performance relates directly to the validity of the model behind the procedure. This paper treats the validation issue as it appears in two classes of inference algorithms relating to genomics - classification and clustering. It formulates the problem and reviews salient results.     

Understanding the nutrigenomic definitions and concepts at the food-genome junction

The marked differences in individual response to dietary factors have led to major controversies in nutrition and puzzled nutrition scientists over the last century. The emerging field of nutrigenomics helps us to understand the basis for some of these differences and also promises us the ability to tailor diet based on individual genetic makeup. Great advances in Human Genome Project, documentation of single nucleotide polymorphisms (SNPs) in candidate genes and their association with metabolic imbalances have gradually added new tests to the nutrigenomic panel. Studies based on ethnopharmacology and phytotherapy concepts showed that nutrients and botanicals can interact with the genome causing marked changes in gene expression. This has led to the commercial development of nutraceuticals and functional foods that can modify negative health effects of individual genetic profile bringing the field to the "food/genome" junction. Despite the promise of nutrigenomics to personalize diet, there is skepticism whether it can truly bring about meaningful modification of the risk factors connected to chronic diseases, due to the lack of large scale nutrition intervention studies. Several intervention studies currently underway in the United States and abroad (Israel, Spain, and France) will further help validate nutrigenomic concepts. France has already introduced a National Nutrition and Health Program to assess nutritional status and risk of major metabolic diseases. As the field(s) related to nutritional genomics advance in their scope, it is essential that: (a) strict guidelines be followed in the nomenclature and definition of the subdisciplines; and (b) the state/federal regulatory guidelines be updated for diagnostic laboratories, especially for those offering tests directly to the public (without a physician's request) to help protect the consumer.     

营养基因组学的研究进展

伴随着基因组学、生物信息学等的迅猛发展及其在生命科学领域的应用,营养基因组学应运而生,并迅速成为营养学研究的新前沿。营养基因组学主要研究营养素和植物化学物质对人体基因的转录、翻译表达以及代谢机制,其可能的应用范围包括营养素作用的分子机制、营养素的人体需要量、个体食谱的制定以及食品安全等。本文重点介绍营养基因组学的研究内容与现状,并对今后的研究趋势作了展望。     

NutriTRAILomics in prostate cancer: time to have two strings to one’s bow

The astonishing development of broad genomics and proteomics tools have catalyzed a new era in both therapeutic interventions and nutrition in prostate cancer. The terms pharmacogenomics and nutrigenomics have been derived out of their genetic forbears as large-scale genomics technologies have been established in the last decade. It is unquestionable that rationale of both disciplines is to individualize or personalize medicine and food and nutrition, and eventually health, by tailoring the drug or the food to the individual genotype. The purpose of this review is to significantly inspect results from current research concerning the mechanisms of action of phytonutrients and potential effects on prostate cancer. Substantial emerging data supports the synergistic adiministration of nutraceuticals with TRAIL in prostate cancer progression to circumvent TRAIL refractoriness. Nonetheless, developing novel scientific methods for discovery, validation, characterization and standardization of these multicomponent phyto-therapeutics is vital to their recognition into mainstream medicine. The key to interpret a personalized response is a greater comprehension of nutrigenomics, proteomics and metabolomics.     

The role of genomics in prevention or reducing the impact of congenital malformations

Congenital malformations (CMs) are permanent changes produced by abnormality of development in a body structure during prenatal life. Population based studies place the incidence of major malformations at about 2-3% of all live births. The etiology is mostly due multifactorial inheritance or unknown (50-80%). The continuum and gradual shift from genetics to genomics will offer new possibilities for diagnosis, treatment, prediction and prevention of congenital malformations. Genomics has many tools including pathogenomics, pharmacogenomics, nutrigenomics and bioinformatics. Pathogenomics will help to discover new genes or susceptibility genes and genetic variants with a role in the pathogenesis of CMs. Pharmacogenomics will identify genetic variants affecting the response to drugs and it should be applied to study drug induced birth defects. Nutrigenomics will determine the impact of diet on genome stability and how genotype determines nutritional requirements. Bioinformatics then will collect, store obtained data, which will facilitate analysis of systems biology questions involving relationships between genes, their variants and biological functions. This knowledge should be translated into more sensitive and specific genetic tests.     

Nutrigenomics and nutrigenetics

PURPOSE OF REVIEW: Nutritional genomics has tremendous potential to change the future of dietary guidelines and personal recommendations. Nutritional genomics covers nutrigenomics, which explores the effects of nutrients on the genome, proteome and metabolome, and nutrigenetics, the major goal of which is to elucidate the effect of genetic variation on the interaction between diet and disease. Nutrigenetics has been used for decades in certain rare monogenic diseases such as phenylketonuria, and it has the potential to provide a basis for personalized dietary recommendations based on the individual's genetic makeup in order to prevent common multifactorial disorders decades before their clinical manifestation. RECENT FINDINGS: Preliminary results regarding gene-diet interactions in cardiovascular diseases are for the most part inconclusive because of the limitations of current experimental designs. Success in this area will require the integration of various disciplines, and will require investigators to work on large population studies that are designed to investigate gene-environment interactions. SUMMARY: Based on the current knowledge, we anticipate that in the future we will be able to harness the information contained in our genomes to achieve successful aging using behavioral changes, with nutrition being the cornerstone of this endeavor.     

Harnessing Nutrigenomics: Development of web-based communication, databases, resources, and tools

Nutrient - gene interactions are responsible for maintaining health and preventing or delaying disease. Unbalanced diets for a given genotype lead to chronic diseases such as obesity, diabetes, cardiovascular, and are likely to contribute to increased severity and/or early-onset of many age-related diseases. Many nutrition and many genetic studies still fail to properly include both variables in the design, execution, and analyses of human, laboratory animal, or cell culture experiments. The complexity ofnutrient-gene interactions has led to the realization that strategic international alliances are needed to improve the completeness of nutrigenomic studies - a task beyond the capabilities of a single laboratory team. Eighty-eight researchers from 22 countries recently outlined the issues and challenges for harnessing the nutritional genomics for public and personal health. The next step in the process of forming productive international alliances is the development of a virtual center for organizing collaborations and communications that foster resources sharing, best practices improvements, and creation of databases. We describe here plans and initial efforts of creating the Nutrigenomics Information Portal, a web-based resource for the international nutrigenomics society. This portal aims at becoming the prime source ofinformation and interaction for nutrigenomics scientists through a collaborative effort.     

Power in pairs: assessing the statistical value of paired samples in tests for differential expression

Background

When genomics researchers design a high-throughput study to test for differential expression, some biological systems and research questions provide opportunities to use paired samples from subjects, and researchers can plan for a certain proportion of subjects to have paired samples. We consider the effect of this paired samples proportion on the statistical power of the study, using characteristics of both count (RNA-Seq) and continuous (microarray) expression data from a colorectal cancer study.

Results

We demonstrate that a higher proportion of subjects with paired samples yields higher statistical power, for various total numbers of samples, and for various strengths of subject-level confounding factors. In the design scenarios considered, the statistical power in a fully-paired design is substantially (and in many cases several times) greater than in an unpaired design.

Conclusions

For the many biological systems and research questions where paired samples are feasible and relevant, substantial statistical power gains can be achieved at the study design stage when genomics researchers plan on using paired samples from the largest possible proportion of subjects. Any cost savings in a study design with unpaired samples are likely accompanied by underpowered and possibly biased results.

     

Machine learning for profile prediction in genomics

A recent deluge of publicly available multi-omics data has fueled the development of machine learning methods aimed at investigating important questions in genomics. Although the motivations for these methods vary, a task that is commonly adopted is that of profile prediction, where predictions are made for one or more forms of biochemical activity along the genome, for example, histone modification, chromatin accessibility, or protein binding. In this review, we give an overview of the research works performing profile prediction, define two broad categories of profile prediction tasks, and discuss the types of scientific questions that can be answered in each.     

Ethical, legal and social issues in nutrigenomics: the challenges of regulating service delivery and building health professional capacity

Nutrigenomics, the conjunction of molecular nutrition with human genomics, is among the first publicly available applications of the human genome project. Nutrigenomics raises ethical, legal and social issues particularly with respect to how the public may access nutrigenetic tests and associated nutritional and lifestyle advice. Current regulatory controversy focuses on potential harms associated with direct-to-consumer (DTC) marketing of nutrigenetic tests and especially the need to protect consumers from unreliable tests, false claims and unproven dietary supplements. Nutrigenomics does, however, offer the potential of important health benefits for some individuals. The regulation of nutrigenomic services is slowly evolving, but there is little indication of increased professional capacity to support service delivery. Primary care physicians have minimal training in nutrition and genetics, and medical geneticists are in high demand and short supply. Dietetic practitioners are experts in nutrition science and interest in nutrigenomics is growing among members of this professional group. However, as with physicians, dietetics practitioners would require considerable training to bring nutrigenomics into their practice capacity. A downside of regulatory restrictions on direct consumer access to nutrigenomics companies is that responsible businesses may be hindered in meeting emergent public demand while health care professional groups have not yet developed capacity to provide nutrigenomics services.     

Web-based education in bioprocess engineering

The combination of web technology, knowledge of bioprocess engineering, and theories on learning and instruction might yield innovative learning material for bioprocess engineering. In this article, an overview of the characteristics of web-based learning material is given, as well as guidelines for the design of learning material from theories of learning and instruction and from the bioprocess engineering domain. A diverse body of learning material is presented, which illustrates the application of these guidelines; this material has been developed during the past six years for different courses, mostly at undergraduate level, and it illustrates how web-based learning material can enable various different approaches to learning objectives that might improve overall learning. Such learning material has been used for several years in education, it has been evaluated with positive results, and is now part of the regular learning material for bioprocess engineering at Wageningen University.     

Using a standard framework for the phenotypic analysis of Medicago truncatula: an effective method for characterizing the plant material used for functional genomics approaches

A crucial step for identifying genes of interest in legume crops is to determine gene function in Medicago truncatula. To facilitate functional genomics in this species, an ecophysiological framework of analysis was developed. Our primary aim was to establish a standard terminology for identifying each organ on the plant. A standard system for the characterization of the vegetative and the reproductive developmental stages was then proposed. Using these tools, the time course of vegetative development of nitrogen-fixing A17 plants was analysed in experiments conducted under different environmental conditions. To take into account the influence of temperature on plant development timing, an original approach was used by modelling vegetative development as a function of thermal time. Interestingly, the use of thermal time highlighted genotypic constants in plant development. Thereafter, to illustrate how this methodology can be used in explaining phenotypic alterations, the phenotype of two allelic mutants was analysed. Because the tools proposed in this paper allow the following: (1) standardization of how the plant material should be characterized to be used for functional genomics; (2) prediction of plant vegetative development; and (3) a more accurate phenotyping, the use of these tools by the M. truncatula community should provide a relevant framework for facilitating the production of reproducible functional genomics data.     

Anticipation of Personal Genomics Data Enhances Interest and Learning Environment in Genomics and Molecular Biology Undergraduate Courses

An important discussion at colleges is centered on determining more effective models for teaching undergraduates. As personalized genomics has become more common, we hypothesized it could be a valuable tool to make science education more hands on, personal, and engaging for college undergraduates. We hypothesized that providing students with personal genome testing kits would enhance the learning experience of students in two undergraduate courses at Brigham Young University: Advanced Molecular Biology and Genomics. These courses have an emphasis on personal genomics the last two weeks of the semester. Students taking these courses were given the option to receive personal genomics kits in 2014, whereas in 2015 they were not. Students sent their personal genomics samples in on their own and received the data after the course ended. We surveyed students in these courses before and after the two-week emphasis on personal genomics to collect data on whether anticipation of obtaining their own personal genomic data impacted undergraduate student learning. We also tested to see if specific personal genomic assignments improved the learning experience by analyzing the data from the undergraduate students who completed both the pre- and post-course surveys. Anticipation of personal genomic data significantly enhanced student interest and the learning environment based on the time students spent researching personal genomic material and their self-reported attitudes compared to those who did not anticipate getting their own data. Personal genomics homework assignments significantly enhanced the undergraduate student interest and learning based on the same criteria and a personal genomics quiz. We found that for the undergraduate students in both molecular biology and genomics courses, incorporation of personal genomic testing can be an effective educational tool in undergraduate science education.     

Integrating anticipated nutrigenomics bioscience applications with ethical aspects

Nutrigenomics is a subspecialty of nutrition science which aims to understand how gene-diet interactions influence individuals' response to food, disease susceptibility, and population health. Yet ethical enquiry into this field is being outpaced by nutrigenomics bioscience. The ethical issues surrounding nutrigenomics face the challenges of a rapidly evolving field which bring forward the additional dimension of crossdisciplinary integrative research between social and biomedical sciences. This article outlines the emerging nutrigenomics definitions and concepts and analyzes the existing ethics literature concerning personalized nutrition and presents "points to consider" over ethical issues regarding future nutrigenomics applications. The interest in nutrigenomics coincides with a shift in emphasis in medicine and biosciences toward prevention of future disease susceptibilities rather than treatment of already established disease. Hence, unique ethical issues emerge concerning the extent to which nutrigenomics can alter our relation to food, boundaries between health and disease, and the folklore of medical practice. Nutrigenomics can result in new social values, norms, and responsibilities for both individuals and societies. Nutrigenomics is not only another new application of "-omics" technologies in the context of gene-diet interactions. Nutrigenomics may fundamentally change the way we perceive human illness while shifting the focus and broadening the scope of health interventions from patients to healthy individuals. In resource- and time-limited healthcare settings, this creates unique ethical dilemmas and distributive justice issues. Ethical aspects of nutrigenomics applications should be addressed proactively, as this new science develops and increasingly coalesces with other applications of genomics in medicine and public health.     

Student retention of course content is improved by collaborative-group testing

We recently reported that collaborative testing (i.e., group test taking) increased student performance on quizzes. It is unknown, however, whether collaborative testing improves student retention of course content. Therefore, this study was designed to test the hypotheses that collaborative-group testing improves student retention of course content. To test this hypothesis, our undergraduate exercise physiology class of 38 students was randomly divided into two groups: group A (n = 19) and group B (n = 19). During exam 1, students from both groups answered questions in the traditional format as individuals. Immediately after completing the exam as individuals, students from group A answered a randomly selected subset of questions from exam 1 in groups of two (1 group had 3 students) to test the effectiveness of collaborative-group testing on test performance and level of student retention. On the next exam (exam 2, 4 wk later), students from both groups answered questions in the traditional format as individuals and responded to the same subset of questions from exam 1. The subset of questions was analyzed to determine the level of retention of the original test material. In addition, immediately after completing the exam as individuals, students from group B answered a randomly selected subset of questions from exam 2 in groups of two (1 group had 3 students). Finally, on the next exam (exam 3, 4 wk later), students from both groups answered questions in the traditional format as individuals and responded to the same subset of questions from exam 2. This protocol followed a randomized crossover design to control for time and order effects. Student retention of course content was reduced when students completed the original examinations individually. In sharp contrast, student retention was improved (P < 0.05) when students completed the original examinations in groups. Results suggest that collaborative testing is an effective strategy to enhance learning and increase student retention of course content.     

How can we adapt education for children across different countries?

The development of educational material as part of major changes in the service of the diabetic clinic in Glostrup University Hospital turned out to be a very rewarding process for the multidisciplinary team. This material includes age-appropriate goals and guidelines for teaching and learning about diabetes, and advice on coping with diabetes in hand-outs designed for diabetic educators, parents/adults and teenagers. The paper briefly describes the developmental process, the concepts and designs of the material, and makes recommendations on how it can be applied. Part of the material was later translated into Polish and Russian in connection with educational projects in these countries, and the material as a whole was recently incorporated into the Novo Care website at www.d4pro.com - Childhood Diabetes. Growing Internet access by diabetes patients, their relatives and health care professionals creates almost overwhelming possibilities, but also brings up many new questions. The development of diabetes educational material of the more 'traditional' type is discussed with regard to content, ideas and design, and focuses on how these concepts and projects can be influenced and adapted across different countries by Internet communication.     

Protein isoelectric point as a predictor for increased crystallization screening efficiency

MOTIVATION: Increased efficiency in initial crystallization screening reduces cost and material requirements in structural genomics. Because pH is one of the few consistently reported parameters in the Protein Data Bank (PDB), the isoelectric point (pI) of a protein has been explored as a useful indirect predictor for the optimal choice of range and distribution of the pH sampling in crystallization trials. RESULTS: We have analyzed 9596 unique protein crystal forms from the August 2003 PDB and have found a significant relationship between the calculated pI of successfully crystallized proteins and the difference between pI and reported pH at which they were crystallized. These preferences provide strong prior information for the design of crystallization screening experiments with significantly increased efficiency and corresponding reduction in material requirements, leading to potential cost savings of millions of US$ for structural genomics projects involving high-throughput crystallographic structure determination. AVAILABILITY: A prototype example of a screen design and efficiency estimator program, CrysPred, is available at http://www-structure.llnl.gov/cryspred/     

Teaching simple experimental design to undergraduates: do your students understand the basics?

This article provides instructors with guidelines for teaching simple experimental design for the comparison of two treatment groups. Two designs with specific examples are discussed along with common misconceptions that undergraduate students typically bring to the experiment design process. Features of experiment design that maximize power and minimize the effects of interindividual variation, thus allowing reduction of sample sizes, are described. Classroom implementation that emphasizes student-centered learning is suggested, and thought questions, designed to help students discover and name the basic principles of simple experiment design for themselves, are included with an answer key.