gaow/cnv-gene-mapping

Develop methods that leverage copy number variations (CNVs) for gene mapping

Fine-mapping of genes using Copy Number Variation data

Develop methods that leverage copy number variations (CNVs) for gene mapping in association studies

https://gaow.github.io/cnv-gene-mapping

Write-up documents

• Slides for 03/23/17 CNV collaboration meeting
• Manuscript
• Other write-ups

Background and literature

• A background on the SuSiE model, Wang et al 2020
• A background on VB and MCMC
• Schizophrenia GWAS
• A background on VB-based variable selection model, Carbonetto and Stephens 2012
• An illustration to pymc3 implementation of Bayesian variable selection via sparse logistic regression
  • with a pymc4 implementation

Please check out the reference folder for more literature references.

Preliminary analysis

• Conventional CNV association study in SCZ data
  • SCZ data:
    • sample size: ~1300
    • .bed file
    • case-control data
    • no sex chromosomes
  • Reference gene panel
  • Enrichment analysis via Fisher Exact Test
• Conventional CNV association study in simulated data as a diagnosis analysis
  • Plot distribution of p-value from simulated data
• Simulation study scheme, a prototype
• Assessment of inclusion of covariates (work in progress)

Method prototypes

• NUTS sampler for MCMC
  • Give background of MCMC using spike-and-slab prior
  • Examples of using pymc3
• Single effect model Bayesian logistic regression
  • Background of method 'SIER'
  • Code example of self-defined 'SIER'

Simulation studies

• Workflow process for running simulation, genome partition and various numerical methods
• Summary and visualization of simulation results
  • PIP calibration analysis and results
  • ROC curve and calibrated mean-pip