WeiYa's Work Yard

A dog, who fell into the ocean of statistics, tries to write down his ideas and notes to save himself.

Systems Genetic Approach

Posted on 0 Comments
Tags: Bioinformatics

There are my notes when I read the paper called System Genetic Approach.


  • the causal modeling algorithms NEO
  • co-expression network algorithm, wMICA


  • GWAS only modest success

    • complex, heterogeneous nature of the disease
  • minimize these complexities in genetic studies of model organisms such as mice

    • classical QTL linkage analyses in mice have identified a number of novel HF-related genes
  • previous work association analyses identified both known and novel genes contributing to hypertrophy

  • an extension of this study though the modeling of biological networks

    • an improved version of the Maximal Information Component Analysis(MICA) algorithms

    • several modules that showed significant association to HF-related phenotypes were identified

    • NEO algorithm to develop a directed network with predicted casual interactions among the module genes.

    • using an in vitro model we validated several of these casual links



Gene Network Analysis Using Weighted MICA

  • HDMP: Hybrid Mouse Diversity Panel
  • prior research using HMDP to generate mRNA co-expression networks.
  • MICA: an unbiased gene network construction algorithm

    • several conceptual improvement over traditional co-expression methods.
    • captures both linear and nonlinear interactions within the data
    • allow genes to be spread proportionally across multiple modules
  • horvath: weighted network construction algorithms, in which all edges are included in the analysis, have greater versatility and power than unweighted algorithms, in which edges are included or exclude based on a hard threshold.
  • improve upon origin algorithm and develop a modified, weighted, form of MICA, called wMICA.
    • the application of wMICA to the analysis of HF, using gene expression data across inbred strains of mice from the HMDP HF study.
  • Filter probes for transcripts that were significantly expressed in at least 25% of samples and had a coefficient of variation of at least 5%. Final a set of 8126 probes, representing 31.6% of the total probes on the array.
  • Three gene networks, 20 modules each.
    • one based only on transcripts from the untreated hearts,
    • one based only on the treated
    • a third based on the change in gene expression between these two conditions two measures.
  • Two measures were used for the preliminary analyses of these networks.
    • calculate significant GO enrichments within each of these modules at several module membership cutoffs.
    • use principal component analysis(PCA) to identify the first principal component of each module.

New words















Published in categories Note