Statistical Genetics II: Quantitative Genetics

Biostatistics 551
Statistics 551

Autumn 2014


Class Meetings: Tuesdays/Thursdays, Health Sciences Building T474, 9:00-10:20 AM
Office Hour: Mondays, 2:00-3:00 PM

Instructor: Timothy Thornton, Ph.D.
Assistant Professor, Department of Biostatistics
Email: tathornt@u.washington.edu
Office: Health Science Building F652
Telephone: 543-8004

Course Information

Syllabus


Announcements

  • (12/11/14)
    Homework 4 and homework 5 solutions have been posted.


  • (12/8/14)
    For question 3, the MLM-Transferrin.txt file originally posted on the course website had the wrong p-values. Please use the "MLM-Transferrin-CORRECTED.txt" file below for question 3 as this file contains the correct association results and p-values for the mixed linear model associaiton analysis of transferrin serum.


  • (12/5/14)
    The final exam has now been posted. It is due at noon on Friday, December 12, 2014. There is no collaboration for the final exam. Contact the instructor for questions about the exam. Click on the link below to download the exam:


    Below are two additional files for the final exam:



  • (12/4/14)
    Below are a few files that will be using in class today



  • (12/3/14)
    In class on Thursday we will do a demonstration with the PLINK software package that will be used for the final exam for QTL association mapping. Please bring your laptops to class with the PLINK software installed if you would like to follow along on your computer. Slides for the PLINK software have been posted in the "Lecture Notes" section. The PLINK software can be downloaded using the following link to the PLINK software webpage:


    Click on the "Download" link near the top of the PLINK website to obtain the software for various platforms (Apple Mac, MS DOS, Linux) Slides for the PLINK software have been posted in the "Lecture Notes" section. We will also be using the R software package. Please make sure that you have R installed on your laptop. Also, please install the R software packages "GWASTools" and "Genabel". You can use the following commands in R to install GWASTools:

    source("http://bioconductor.org/biocLite.R")
    biocLite("GWASTools")

    You can use the following commands in R to install GenABEL:

    install.packages("GenABEL")

  • (11/26/14)
    Homework 5 has been posted. It is due on Thursday, December 4, 2014.


  • (11/26/14)
    Homework 3 solutions have been posted.


  • (11/21/14)
    In class on Tuesday, we will discuss Nature Genetic papers entitled "Common SNPs explain a large proportion of the heritability for human height" by Yang, Visscher et al. (2010). Please read the article for Tuesday's class, any pay careful attention to the "Online Methods" section at the end. The article can be found in the Journal Articles section of the course website or you can download the paper directly from the Nature Genetics Website using the following link:



  • (11/19/14)
    In class tomorrow we will use the QTDT software package for variance components estimation. Four data files for the QTDT software that we will use in class today can be found in the "Data Sets" section of the course webpage. Please bring your laptops to class with the QTDT software installed if you would like to follow along on your computer. Before class tomorrow, follow the QTDT quick tour on the QTDT website and follow the intructions to get familiar with the software. The QTDT software (and the QTDT tutorial) can be downloaded using the following link to the QTDT software webpage:



  • (11/11/14)
    Homework 4 has been posted. It is due on Thursday, November 20, 2014.


  • (11/6/14)
    Homework 2 solutions have been posted.


  • (11/6/14)
    For the PEDSTATS example in class today, all heritability estimates for the different relative pair types should be calculated using the correlation (r), not the correlation squared (r^2). For example, the correlation of trait values for full-siblings is r_sibs=COV(Sib1 Trait,Sib2 Trait)/Var(Trait), and we obtain heritabilty estimates using r_sibs, not (r_sibs)^2. Assuming no dominance effects (and no shared environmental effects), an estimate for the heritability of a trait using the correlation of full sibling trait values would be 2*r_sibs, not 2*(r_sibs)^2. Be careful not to confuse narrow sense heritability, which is often written as h^2, with the squared correlation of trait values, r^2, for different classes of relatives. For examples, when considering the correlaiton of trait values for MZ twins and assuming no dominance or shared environmental effects, heritability would be equal to the correlation (not the correlation squared), i.e., h^2=r_MZ.


  • (11/5/14)
    In class tomorrow (Thursday) we will use the PEDSTATS software package. Please bring your laptops to class with the PEDSTAT software installed if you would like to follow along on your computer. Data files for the PEDSTATS software that we will use in class tomorrow can be found in the "Data Sets" section of the course webpage. Introductory slides for Pedstat software have been posted. The PEDSTATS software can be downloaded using the following link to the PEDSTATS software webpage:



  • (10/28/14)
    Tuesday, November 11 is Veterans Day, so there is no class. Please email your homework 3 solutions to the Instructor by November 11, 2015. There also will be no class on Thursday, November 13, 2015.


  • (10/27/14)
    Homework 3 has been posted. It is due on November 11, 2015.


  • (10/22/14)
    Homework 1 solutions have been posted.


  • (10/14/14)
    Homework 2 has been posted. It is due on Thursday, October 23, 2014.




Lecture Notes and Handouts




Homework Assignments




Journal Articles



Data Sets