epistasis {happy}R Documentation

Analysis of Epistasis between Markers

Description

epistasis() will test for a statistical interaction between two sets of markers within the happy framework. The markers should be sufficiently far apart that they are unlinked (in practice 10cM for a 30 generation HS is sufficient). A partial F-test is performed to test if a model allowing for interactions fits better than a model in which each marker's contribution is additive between loci. Note that the effect of each marker within a locus can be either additive or full. Merging of strain is permitted.

epistasispair() is the same as epistasis() except that only one pair of markers is tested.

Usage

epistasis( h, markers1, markers2, merge1=NULL, merge2=NULL,
model='additive', verbose=FALSE, family='gaussian' )
epistasispair( h, marker1, marker2, merge1=NULL, merge2=NULL,
model='additive', verbose=FALSE, d1=NULL, d2=NULL, main1=0, main2=0, family='gaussian' )

Arguments

h an object returned by a previous call to happy()
markers1 an array of marker names or indices
markers2 an array of marker names or indices
marker1 a single marker name or index
marker2 a single marker name or index
merge1 an optional merge object (returned by mergematrices()) determining how the strains should be merged together for the markers listed in marker1
merge2 an optional merge object (returned by mergematrices()) determining how the strains should be merged together for the markers listed in marker2
model the type of model fitted at each locus. Either 'additive' or 'full'
verbose switch controlling output to screen
d1 optional design matrix for the main effect of the first marker (saves computation time)
main1 optional log-P-value for the main effect of the first marker. NOTE: If d1 is not NULL then main1 must be set
d2 optional design matrix for the main effect of the second marker (saves computation time).
main2 optional log-P-value for the main effect of the second marker. NOTE: If d2 is not NULL then main2 must be set
family The distribution of errors in the data. The default is 'gaussian'. This variable controls the type of model fitting. In the gauusian case a standard linear model is fitted using lm(). Otherwise the data are fitted as a generalised linear model using glm(), when the value of family must be one of the distributions hangled by glm(), such as 'binomial', 'gamma'. See family() for the full range of models.

Value

epistasis() returns a matrix with columns named 'marker1', 'marker2', 'main1', 'main2', 'main1+main2', 'main1*main2', 'main1.main2'. marker1 and marker2 are the names of the markers being compared in a given row, the remaining values are the ANOVA log-P-values of the main effects (main1 and main2), the combined additive effect (main1+main2), the addtive plus interaction (main1*main2) and the partial F of the interaction (main1.main2) after allowing for main1+main2. epistasispair() returns a list with the same fields.

Author(s)

Richard Mott


[Package happy version 2.1 Index]