#Impute F2 genotypes based on parents' genotypes

This is currently only for F2 populations.

 (1) each chromosome is imputed separately
 (2) If parents are the same for that marker, marker is imputed as the parent gentoype
 (3) if there is only one marker in the chromsome, it is imputed as the mode genotype
 (4) If one parent is NA, the genotype is imputed as the mode genotype
 (5) for (3) and (4), if the mode genotype is NA, marker is not imputed, and the marker is removed from the final imputed data.
 (6) All other situations are imputed based on conditional distribution of flanking markers

##Install F2imputation is not available on CRAN yet. However, it can be installed directly from GitHub using the devtools package.

1. Install devtools package: install.packages('devtools')
2. Load devtools package: library(devtools)
3. Install BGData package from GitHub: install_github('lian0090/F2imputation')
4. Load SKAT2 package: library(F2imputation)

##Usage imputation(progeno,pargeno,map,nColSkip)

• Arguments

  • progeno and pargeno: typed markers of F2 progeny genotypes (progeno) and parents genotypes (pargeno). Each row is an individual and there should be only two rows for pargeno for the two parents. The 1:nColSkip columns can be pedigree information. Marker names must be in column names. Genotypes must be coded in -1,0,1. For example, if progeny are only genotyped for M1, M3, and parents are genotyped for M1, M2, M3, M4, M5. If ID is included in the genotype, then, progeno will have three columns with column names ID, M1, M3. pargeno will have six columns with column names ID, M1, M2, M3, M4, M5.
  • map: each row of map describes a single marker and must contain 3 columns: chromosome, SNPid, Genetic distance (cM)
  • nColSkip: number of columns to skip before the first marker genotype
  • IDcol: column number for LINE ID, if specified, the LINE IDs will be added to the dimnames of genotype probability array.

• Return Values

  • pimputedgeno: a dataframe of imputed genotypes
  • genoprob: an array of genotype probabilities

##Example

data(map)
data(progeno)
data(pargeno)
out=imputation(progeno,pargeno,map,nColSkip=1)

##Compatability with old version imputation.cmbgeno(combinedgeno)

• arguments - combinedgeno: a matrix or data.frame combined the map information, parents genotype and progeny genotype. Entries of combinedgeno are character types so that all the information can be put together into a single matrix. Marker genotypes must be coded as -1, 0, 1. - column 1 is LINE names, first two rows of column 1 are NA. - row 1 is chrosomes number - row 2 is genetic distance in centimorgan (cM) - row 3 and row 4 are marker genotype for the two parents - row 5 onwards are the progeny genotypes

• Return Values

  • pimputedgeno: a dataframe of imputed genotypes
  • genoprob: an array of genotype probabilities

• R code comparing the old version and new version

data(combinedgeno)
outcmb=imputation.cmbgeno(combinedgeno)
outcmb.geno=outcmb$pimputedgeno[-c(1:4),]
outcmb.prob=outcmb$genoprob
data(progeno)
data(pargeno)
data(map)
out=imputation(progeno,pargeno,map,nColSkip=1,IDcol=NULL)
out.geno=out$pimputedgeno
out.prob=out$genoprob
all(out.geno==outcmb.geno)
all(out.prob==outcmb.prob)