statgen-primer

These notes are written to help (some of our own) trainees who already have some grounds in basic (human) genetics concepts and undergraduate-level statistics, yet find it difficult to learn in somewhat detail various statistical methods commonly used in statistical genetics research nowadays. The goal is to bridge the knowledge gap between these existing backgrounds and the more advanced learning goal of interest.

Overview of Contents

Section	Subsection	Topic
Basic Concepts		Genotype Coding
		Minor Allele Frequency
		Hardy-Weinberg Equilibrium
Correlation	Between variants	Linkage Disequilibrium
		Linkage Disequilibrium Score
	Between individuals	Genetic Relationship Matrix
Genetic Associations	Basic model	Ordinary Least Squares
	Extend to binary outcome	Odds Ratio
	Extend to multiple variables	Marginal vs. Joint Effects
		Summary Statistics
Genetic effects		Random Effect
		Proportion of Variance Explained
		Linear Mixed Model
	Covariates	Confounder
		Collider
		Mediator
	Multiple studies	Meta Analysis Fixed Effect
		Meta Analysis Random Effect
Statistical Inference	Likelihood	Likelihood
		Maximum Likelihood Estimation
		Likelihood Ratio
		Expectation-Maximum Algorithm
	Bayesian versus Frequentist	Bayesian versus Frequentist
		Bayes Rule
		Bayes Factor
		p-value and Bayesian Hypothesis Testing
	Bayesian Model	Bayesian Normal Mean Model
		Bayesian Multivariate Normal Mean Model
	Multiple Bayesian Models	Bayesian Model Comparison
		Bayesian Mixture Model
	Latent Structures in Data	Factor Analysis
		Principal Component Analysis
		Hidden Markov Model

These notes draw inspiration from fiveMinuteStats by Matthew Stephens and statistical genetics equations by Robert Maier. Compared to Matthew’s materials, these notes are more narrowly focused on human and statistical genetics with only as much statistical details to understand the applications. Compared to Robert’s materials, these notes include a slightly stronger statistical component to serve as “primer” for readers to advance into details in advanced statistical genetics methods in practice.