跳至正文
首页 » Covariates, fixed and environmental random effects

Covariates, fixed and environmental random effects

For covariates, fixed effect and environmental random effect of a trait, please specify its position at phenotype file to flag --qcovar, --dcovar, --rand respectively.

For single trait, please use comma as separator, for example:
--qcovar 4,5: take “day” and weight” as covariates
--dcovar 2,3: take sex” and season” as fixed effects
--rand 6: take loc” as environmental random effect

For multiple traits, if there are some environmental factors needed to be fitted as a model effect, all traits should be specified at this model effect, and please use comma as separator within a trait, and use space among traits, use 0 for a trait if those factors are not considered for it, take fixed effect for an example:

./hiblup ... --dcovar 2,3 0 2,3,6 ...

which means there are 3 traits in analysis, the first trait has two fixed effects: “sex” and “season“; the second trait has no fixed effect; the third trait has three fixed effects, which are “sex“, “season” and “loc“.

IMPORTANT

--qcovar is used to specify the environmental effects as quantitative covariates, the records for the effects should be in digits, any character is not allowed.

--dcovar is used to specify the environmental effects as discrete covariates, the records for the effects could be in either digits or characters, both are acceptable.

--rand is used to specify the environmental random effects, which come from the recorded environmental factors, the genetic random effects are from either pedigree or genomic information, if pedigree or genotype file is provided, HIBLUP will fit the genetic random effect in the model automatically, users DO NOT need to add the column of individuals id as random effect here.