Imputation Algorithm for Genetic Data
R programs for imputation of microarray gene expression data
- Simulation program This program simulates Affymetrix GeneChips using the model of Stevens and Doerge (2005).
- Analysis program
This program analyzes differential expression for Affymetrix GeneChip data using single and multiple imputation. Portions of the R source code are based on the random variance model (RVM; Wright and Simon, 2003, http://linus.nci.nih.gov/BRB-ArrayTools.html) and the COMBCHI SAS macro (http://www.ssc.upenn.edu/~allison/combchi.sas).
- ImputationSimAnalyze.r – for simulated data
- ImputationMAQC.r – for MAQC data
- Combine Results
This program combines results of a batch of analyses from the programs above.
- CombineImputationSim.r – for simulated data
- CombineImputationMAQC.r – for MAQC data
R programs for imputation of pedigree genetic data
- Mig Package
The ‘Mig’ package stands for multiple imputation in genetics. The main focus was to develop methods for multiple imputation of pedigree data both genotype and covariate imputation. This package consists of two functions
- miCovariate – To impute covariate data. The details of the input, output and method are available at (MI_Covariate.pdf)
- miGenotype – To impute genotype for a single marker. The details of the input, output and method are available at (MI_GenotypeImputation.pdf)
This package has man pages that describe the above two functions and small R demo that will run a small example of these functions. We provide the package as Windows binary and tar.gz (The file name is “mig_0.1.0.tar_.gz”. Please rename it to “mig_0.1.0.tar.gz” – remove “_” in the file name.) that can be built separately on any Linux machine.