In DavisBrian/gwasr: Provide a framework to make GWAS easier.

Basic Todo list of ideas for the analysis pipeline.

Basic Pipeline procedure

• create a phenotype object (basic error checking)
• create a snpinfo class (basic error checking)
• create a genotype class (basic error checking)
• create a gwas_list (reduce pheno/geno/snpinfo into a common minimal data set)
• create a standard output structure (not part of the package)
• run arbitrary test
  • seqMeta
  • create a prepScores object
  • run T1/T5/Skat/etc
  • gwaslm
  • firth
• create reports
  • results tables
  • qq plots
  • manhattan plots
  • compare tests
• ideas
  • for a gwas_list run an arbitrary function on each "gene" (do??)
  • tidyr / dplyr concepts
• functions
  • gwaslm
  • maf
  • impute(x, type="mean")
  • cov
  • score
  • SEE seqMeta function

phenotype

• create a phenotype class
  • drop unwanted / unneeded columns
• improved error checking
• summary function
• print (similar to data_frame)
• data_frame support

genotype

• create a genotype class
• improved error checking
• summary function
• "split" by gene (similar to dplyr with attributes)
• print (similar to data_frame)
• option to store genotype in sparse matrix format

snpinfo

• create a genotype class
  • drop unwanted / unneeded columns
• improved error checking
• summary function
• print (similar to data_frame)
• data_frame support

gwas_list

• create gwas_list
  • reduce down to common data
  • warning / stop if nothing in common
• summary function
• print (similar to data_frame)
  • number of subjects (first x)
  • number of snps (first x)
  • number of "genes" (first x)
  • number of "chromosomes" (first x)
  • print(pheno)
  • print(geno)
  • print(snpinfo)
• subsetting (see multitable - Extract.R, mefa, mefa4)
  • [subjects, snps, silevel]
  • `[[ subject/snp/aggregateBy/chr/snpName ==
• get_phenotype
• get_snpinfo
• get_genotype
• get_kinship
• get_phenotype_attributes
• get_snpinfo_attributes
• get_genotype_attributes
• get_kinship_attributes
• option to store genotype in sparse matrix format
• str.gwas.list <- function(object, give.attr = FALSE, hide.internals = TRUE, ...){

Apply function for gwas_dataset

• function(X, MARGIN, FUN, ...) {

reporting

• compare p-values of test (facet for > 2 test ???)
• manhattan plot of p-values
• summary of inputs / run
• results to xlsx

misc

single object functions

• summary
• print
• subsetting

multi object functions

• intersect phenotype, genotype, snpinfo into a data set to run tests on.
• gene lookup (return just pheno, geno, snpinfo of that gene)
• seqMeta parts (score, cov, maf, etc )
• test <- function(geno, snpinfo, pheno, silevel, fun=, args=)???

test

• gwasrlm with weights
• burden test (create a wrapper function fo Tx, MB)
• skat
• permutation based results
• Firth (speed up???)

DavisBrian/gwasr documentation built on May 6, 2019, 1:55 p.m.