Multivariate Concepts

Overview

  • Originally presented: June 9, 2026
  • Lead: Andrew Grotzinger
  • Topics: genetic correlations, LD-score regression, Genomic SEM
Summary

This lecture series focuses on advanced methods in genome-wide association studies (GWAS) and statistical genetics, with an emphasis on analyzing and modeling complex traits using summary-level genomic data. Students are introduced to practical workflows for preparing and formatting GWAS data, including quality control steps and tools such as munging functions. The series then explores key analytical techniques like linkage disequilibrium score regression (LDSC), multivariate GWAS, and GWAS-by-subtraction, highlighting how genetic relationships between traits can be estimated. It also introduces structural equation modeling (SEM) and its application in Genomic SEM to model the shared genetic architecture of multiple traits.

Lectures

This series of lectures can be viewed as a YouTube playlist.

Genomic SEM Introduction

Slides


Lavaan syntax and SEM introduction

Slides


Explaining how S and V are estimated and what they are


The munge function

Slides


Estimating LDSC

Slides


Running usermodel

Slides


Formatting GWAS data using sumstats

Slides


Estimating multivariate GWAS with userGWAS

Slides


GWAS-by-subtraction

Slides

 

Practicals

The practical uses these slides.