𝕊ud et al (see reference section). This review covers most of the concepts we want to cover in this class. We will learn by doing and unpack many of the concepts mentioned in the review. Hopefully, you will re-read the paper at the end of the class and realize how much you have learned. […] … Read more →

Read more →

𝕀f you work on Midway2, some of the steps have been done for you (checked squares) […] # define environmental variables for convenience USERNAME="WRITE YOUR USER NAME HERE" PRE="/project2/bios25328/Data/Lab-QGT/" … Read more →

𝕋he slides can be found [in this link] (https://www.icloud.com/keynote/0Z_tcX5NElsNKIDL3gs0llnUQ#Post-GWAS-MR) Read more →

ℙan UK Biobank set of GWAS results, position are in hg37, alternative alleles are the effect allele. suppressMessages(library(tidyverse)) webdata_pre = "/Users/haekyungim/Box/LargeFiles/imlab-data/data-Github/web-data/" work.dir = … Read more →

𝕋oday, we will follow the applied guide to creating and validating polygenic risk scores (PRS) by Mills, Barban, and Tropf. Mills, Melinda C.; Barban, Nicola; Tropf, Felix C.. An Introduction to Statistical Genetic Data Analysis (p. 243). MIT Press. Kindle Edition. […] Recall that a polygenic … Read more →

𝔾oal -[ ] download Personal Genome Project data (http://repgp.teamerlich.org/) -[ ] analyze content, explore structure -[ ] predict height and other phenotypes with PRS and compare to reported values library(tidyverse) ## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.0 … Read more →

ℂreated by Max Winston & Brandon Pierce; modifications by Charles Washington III, Ankeeta Shah, and Yanyu Liang. Genome-wide Complex Trait Analysis (GCTA, see documentation here) was originally designed to estimate the heritability of complex traits using genome-wide SNPs, but has now been … Read more →

𝕀n this exercise, you will use LD score regression method to calculate the chip heritability of two GWAS phenotypes from the UK Biobank data and look for evidence of population stratification. […] To get started on using the software for this problem, you may find this document helpful … Read more →

𝔽ind the class notes here […] Notice that, under the null hypothesis, the estimated effect size divided by the standard error behaves like a normal(0,1) random variable when the sample size is large enough: […] \[ Z = \frac{\hat\beta}{\text{se}(\hat\beta)}\] \[Z \approx N(0,1) ~~~~~~ … Read more →