confound: Examine confounding of covariate with pseudomarkers.

In byandell/qtlbim: QTL Bayesian Interval Mapping

Description

Covariates used in gene mapping may be correlated with covariates. These routines examine the pattern of confounding.

Usage

qb.confound(qbObject, covar = 1)
## S3 method for class 'qb.confound'
plot(x, ylim, main, ...)
## S3 method for class 'qb.confound'
print(x, ...)
## S3 method for class 'qb.confound'
summary(object, ...)

Arguments

qbObject	Object of class qb.
x	Object of class qb.confound.
object	Object of class qb.confound.
covar	Index to covariate
ylim	Limits for y (vertical) plotting axis.
main	Title for plot.
...	Additional parameters passed alone.

Details

This examines possible confounding between a covariate and pseudomarkers across the genome. Confouding, evidenced by large correlation with a marker, would raise suspicions about mapping in a genomic region, unless of course the covariate is a marker in that region. Blue curves are correlation with additive effect; red curves are correlation with dominance effect. Dashed lines at 5 percent significance limits.

Value

qb.confound returns a matrix with columns for:

coradd	Correlation with additive pseudomarker effect.
cordom	Correlation with dominance pseudomarker effect (if F2).
chr	Chromosome identifier.

The object inherits from scanone objects.

Author(s)

Brian S. Yandell

References

http://www.qtlbim.org

See Also

qb.mcmc

Examples

data(qbExample)

temp <- qb.confound(qbExample)
plot(temp)

byandell/qtlbim documentation built on Dec. 19, 2021, 12:47 p.m.