SNP FSA is a pipeline for applying statistical analysis tools to identify interactions between SNPs and their effects on phenotypic expression.
This pipeline allows users to access and apply new statistical analysis tools to indentify possible phenotypic outcomes as a result of epistatic interactions between Single Nucleotide Polymorphisms (SNPs). This pipeline is designed to take a group of population genotype data as either a VCF or NCBI GEO files and a two column file explaning case-phenotype relationships to produce many sets of combination of SNPs related to the phenotype. This pilot version has been tested on case-controlled data obtained from NCBI GEO database. In order to rapidly produce SNP interaction sets, SNP data is separated into smaller subsets. Default subsets are created based on clinvar phenotype data and based on the coexpression of genes to partitioned SNPs. A user may also define their own subsets for the phenotypes.
This pipeline uses a set of Python script converters to generate the input files (from the original user input) for the R package, rFSA, tools to be used.
Dependencies for our pipeline include up to date versions for Python and R. The user must also download rFSA R utils package, and pandas package for python.
https://www.python.org/downloads/
http://pandas.pydata.org/pandas-build/dev/
https://github.com/joshuawlambert/rFSA
python SNPInteractPipeline.py --filename GSE66903_series_matrix.txt --snp-id-label SNP_ID --snp-mapping-file GPL6801-4019 --grouping-method all --restrict-subset-size 100 10000000 --fsa-method FSA_nogroups.R
The runtime of this pipeline is based on the number of SNP's, subsets, and samples, as well as the size of the subsets. Thus for the time being more testing is needed before an acurate runtime can be established.