Introduction
Statistical genetics is the field of science concerned with developing and applying statistical methods to draw inferences from genetic data. The research involved in statistical genetics usually comprises developing theory or methodology to support research in various fields such as population genetics, genetic epidemiology, and quantitative genetics. When phenotype and genotype data from a sufficient number of individuals affected by a suspected genetic disorder are collected, numerous statistical approaches can be made to quantify the risk and significance of disease-causing variants. However, many factors determine whether the approaches are sufficiently powered, out of which some can only be estimated by nontrivial uncertainty during the research study design and analysis. The most relevant factors in the analytical study design of any rare disease are its prevalence and the mode of inheritance. These factors determine whether affected individuals are likely to aggregate in families or appear sporadically. It certainly impacts the number of affected individuals available for clinical evaluation and can be included in genetics studies.
Why Is Statistical Genetics Important?
In the early days, statistics largely used statistical genetics for experimental studies in plants and animals. Over the past years, the methodological basis of statistical genetics has changed constantly to keep pace with the numerous genetic data available that technology has provided. Most recently, new technologies have arisen from the Human Genome Project and the HapMap Project, which generated various methodological developments addressing the unsolved problems in human genetics. The development of statistical methods and models to understand how genes influence traits are still progressing in plant, animal, and human genetics. Advanced statistical methods used in statistical genetics are necessary for interpreting and making modern and complex biomolecular, epidemiological, or environmental data more understandable. Modern biostatistical methods can help understand the genetic structure of any disease, disease progression, and their pathways, and also study the impact of environmental exposures on various chronic diseases.
What Is the Role of a Statistical Geneticist?
The duties of a statistical geneticist involve developing novel statistical methods to model and interpret the data confined to human genetic material. The responsibilities of a statistical geneticist depend on the branch of statistical genetics in which they work. The models created by a geneticist must be built in a way that helps in specific types of research, like population genetics, epidemiology, or quantitative genetics, and also help other scientists analyze their experiments and results. Geneticists work closely with researchers in other fields, such as computational biologies, like clinicians, physicians, molecular biologists, and bioinformatics researchers.
What Is the Application of Statistical Genetics in Medical Studies?
The applications of statistical genetics were presented in numerous publications, and some of them are as follows-
-
Applications of Bayesian Gene Selection and Classification With Mixtures of Generalized Singular G-Priors - The proposed approach in this article was applied to a colon and leukemia cancer study. In this study, compared with other existing methods, the authors found that the classification accuracy of the proposed model is higher with a smaller set of selected genes, resulting in replicated findings in several previous studies and a stronger association with posterior probabilities.
-
Modified Logistic Regression Models Using Gene Coexpression and Clinical Features to Predict Prostate Cancer Progression - A new logistic regression model was proposed for predicting prostate cancer progression. The authors incorporated coexpressed gene profiles into the logistic model based on clinical features to improve the inference accuracy. Then they used the top-scoring pair method to select the genes associated with significant disease. This study suggested that the proposed method works better than a common one, and a top-scoring pair method is a useful tool for feature selection used in predictive models.
-
Power and Stability Properties of Resampling-Based Multiple Testing Procedures With Applications to Gene Oncology Studies - The authors investigated the power and stability of seven commonly used resampling-based testing procedures used in high-throughput data analysis for small sample size data. This study suggested that the bootstrap single-step minP procedure and the bootstrap step-down minP procedure (the bootstrap value is the proportion of replicated phylogenies that recovered a particular clade from the original phylogeny built using an original alignment) work the best when the sample size is as small as three in each group. Either familywise error rate or false discovery rate control is desired. When the sample size increases to twelve and false discovery rate control is desired, the permutation maxT procedure and the permutation minP procedure function the best.
What Are the Challenges in Statistical Genetics?
With numerous benefits, the field of statistical genetics is also faced with certain challenges, such as
-
Methodologies for New Data Types - Better statistical analysis methods that different carrier sources of error during the processing stages of genomic assay data are needed as technologies are constantly improved, extended, and replaced with new ones.
-
Approaches to Integration of Different Data Types - Drawing appropriate inferences from combined sets of assays is not insignificant, as it requires not only a familiarity with methods and models for handling sources of error associated with each assay but also an ability to model the relevant system as a ‘one’ above its assayable parts. Such modeling can pose a complex problem for statistical geneticists.
-
Modeling Population Level Phenomenon - Population genetics is a vast area of biomedical science that is not just confined to the ecological or genetic epidemiological sciences. Hence, statistical methodologies for assembling and studying the networks, the flow and transmission of information from generation to generation, the hierarchical functioning of gene regulatory circuits, and other related phenomena will become important and increase constantly.
-
Predictive Modeling - The availability of genetic assay results related to phenotypic categories creates a need to develop predictive models and classifiers for the phenotypic categories. Applications-oriented researchers in numerous settings may use such models. Still, they can become complicated by various statistical issues, like having more predictors or variables than units of observations, the need to accommodate potential confounding by covariates, dealing with mixed longitudinal data, and many more.
-
Design of Studies Pertinent to Genetics - The availability of cost-effective genomic technologies can hinder scientists.
Conclusion
Statistical genetics is a vast field of science that helps develop statistical methods to understand the genetic basis of human diseases and their traits. The methods involve large-scale data sets from candidate-gene, genome-wide, and resequencing studies that use related and unrelated individuals. Though faced with some challenges, statistical genetics is still developing for better innovations.
