Genome-Wide Association Study (GWAS) is a hypothesis-free method for identifying associations between genetic regions and traits. GWAS analysis are usually used to identify genes involved in human disease. By applying GWAS analysis to variant data we will be able to identify a given variant (or a set of variants) involved in a given phenotype or disorder.Based on a statistical test, GWAS analysis will provide a level of significance (or p-value) for each variant. OpenCGA implements GWAS analysis based on the statistical tests: chi-square and Fisher.
Implementation
OpenCGA GWAS analysis extends Oskar GWAS implementation. GWAS is implemented using Hadoop MapReduce over HBase.
Input Parameters
OpenCGA support different input parameters:
Variant data with sample genotypes
Two list of samples (case-control study)
Statistical test: chi square or fisher.
Instead of providing two lists of samples, users can provide:
A phenotype, and
A pedigree for those variant samples
Configuration
OpenCGA implementation supports different configuration variables. This can be setup in the OpenCGA installation folder or specified during execution.
Output
A text file that consists of a header line (starting with #), and then one line per variant with the following 12-13 columns:
chromosome
Chromosome code
start
Start base-pair coordinate
end
End base-pair coordinate
strand
Strand
reference
Reference allele
alternate
Alternate allele
dbSNP
Variant identifier
gene
Gene name
biotype
Bioytpe
conseq. types
List of consequence types
chi square
Allelic test chi-square statistic. Not present with 'fisher' test.