Christoph Rau- Multiple Phenotypes

Project Member

Christoph Rau

The goal of this quarter

Develop a method and a computer program for analysing how component phenotypes can be analysed using QTL analysis to further enhance knowledge of the end-state disease.

The Schedule for the quarter


Project Description

Traditional genetic analysis, such as quantitative trait locus (QTL) analysis, attempts to determine specific genes that regulate the phenotype of a complex trait. These methods have had some success, but suffer from a problem of resolution. In many cases, QTL analysis can only narrow the region of interest to about a 20-40 centimorgan window, which contains a large number of genes, although recent progress has helped narrow that window. (Farber)
Gene network analysis approaches the resolution issue from a different direction. Instead of trying to pinpoint individual genes, gene network analysis looks for similar expression patterns across multiple genes, linking those genes together into gene modules and then examining how these modules behave when subjected to perturbation. This is especially important for complex genetic traits, since even the totality of identified genes for a given complex trait typically only account for 20-30% of the total variability (Jake Lusis, private conversation). Most of the disease variation lies in the hundreds of genes that contribute very little individually (less than 1% each). QTL analysis is limited to finding the few genes that stand out, percentage-wise, from the rest of the trait-related genes, while gene network analysis is able to sum together many small-effect genes to produce a more significant block of genes, allowing for a greater level of identification.
The Watershed model of complex diseases involves approaching a complex disease as a large network of genes that combine with one another to eventually create the overall phenotype. The intermediate steps between genes and the disease are gene modules, or small groups of genes that work in a related way, and intermediate phenotypes, which are groups of gene modules that together display an identifiable trait. A good example of this model is schizophrenia. The disease is characterized by, among other things, the co-incidence of depressive and social anxiety symptoms. These symptoms are recognizable on their own and are intermediate phenotypes.
The Project
This project attempts to use gene modules and intermediate phenotypes to improve the identification of specific genes related to a complex disease. In brief, gene arrays will be determined for each intermediate phenotype, then compared to one another to remove unimportant genes and rank the remaining genes in order of likelihood.

Related Papers

Zhang B. Horvath S. A general framework for weighted gene co-expression network analysis. Statistical Applications in Genetics and Molecular Biology Vol 4. no 1. 2005.
Langfelder, P. Zhang, B. Horvath, S. Defining clusters from a hierarchical cluster tree: Dynamic Tree Cut library for R. Bioinformatics. Nov. 2007

April 23rd

  1. Last week I created this page and started thinking about my project
  2. This coming week I plan to learn how to work with qtl and start writing code
  3. I did everything I planned to do last week! hooray!
  4. I deserve an A+++.

May 2nd

  1. Last Week I figured out how Rqtl works (stupid formatting) and came up with how I was going to deal with the multiple phenotypes. I didnt start coding, but I think I got a lot of the important planning out of the way, so… Im okay with this.
  2. This coming week, its time to code, and code a lot. I'd like to finish the qtl parts, and maybe start on the comparision program.
  3. I got most of what I planned to do last week done…
  4. I deserve an A-

May 8th

  1. Last week I continued reading about comparison programs, but got a bit stuck on how to compare QTLs to one another. I managed to hammer out a basic perl program to do the analysis up to that last point for me, however, upon talking to Dr. Lusis, Im going to redo this program to look at gene modules as opposed to individual gene QTLs. Lots of work to be done :)
  2. This coming week… more coding, and changing the code to work with the gene modules, which I ha ve a more intuitive understanding of how to compare between phenotypes (sorta like glorified haplotypes, really…)
  3. Well, I kinda changed my focus completely on the last day of the week, so… I think that a lot of what I did last week is mostly moot… but I did put in the effort!
  4. I deserve an… A.

May 19th

  1. Last week was a less than productive week for me, for personal reasons. However, I was able to write a significant amount of code, and, aside from a few snags, have most of my program figured out… I think.
  2. This week… finish coding, start debugging, develop some sample data sets.
  3. I got less done than I had hoped to this past week, but Im okay with that.
  4. I deserve a B+(ish)

May 30th

  1. Well, I missed a week due to several grants being due that ate my soul. In good news, though… I've finished coding, started making the presentation and… got mired in debugging.
  2. This week… DEBUG like mad, finish presentation
  3. Well, two weeks ago I got nothing done, but I dont think thats a problem. THis past week I've been very productive
  4. I deserve a B+(ish)

June 10th

  1. Well, coding is done and Im about to turn in the project. Im quite pleased with what I came up with :)
  2. No plans for next week :)
  3. I presented and finished up, so… I did what I wanted to
  4. A.
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License