close: Measures closeness of genetic architectures to target.
In byandell/qtlbim: QTL Bayesian Interval Mapping

Description Usage Arguments Details Author(s) References See Also Examples

Boxplots and summaries of how close MCMC samples of genetic architectures are to target architecture.

qb.close(qbObject, target = NULL, epistasis = TRUE, signed = FALSE,
  score.type = c("sq.atten","attenuation","variance","recombination",
  "distance"), ...)
## S3 method for class 'qb.close'
plot(x, category = c("pattern", "nqtl"), xlab,
  cutoff, sort.pattern = c("percent","score"), ...)
## S3 method for class 'qb.close'
summary(object, cutoff, digits = 0,
  show = "score", ... )

`qbObject`	Object of class `qb`.
`x,object`	Object of class `qb.close`.
`target`	Target architure as data frame with columns `chrom` and `locus`. Extracted from `summary` if `target` is a `qb.scanone` object. If `target` = `NULL`, `score.type` is set to `"variance"` to be compared with null model.
`epistasis`	Include epistasis in patterns if `TRUE`.
`signed`	Sign score. Most useful to examine single chromosome.
`score.type`	Type of score to use as distance.
`category`	Boxplots indexed by `nqtl` or `pattern`.
`xlab`	Label for X axis (default taken from `x` object).
`cutoff`	Percent cutoff for pattern inclusion in model selection. Default is 0.25 (0.5) if `epistasis` is `TRUE` (`FALSE`).
`sort.pattern`	If `type` = `"pattern"`, sort by `percent` posterior or by median of `score`.
`digits`	Number of digits displayed for locus.
`show`	Character string with name from `object` to show.
`...`	Parameters to methods.

Closeness for each loci is measured as 1-2r, with r the recombination rate. Thus unlinked loci have measure 0. Loci between a MCMC sample architecture and the target architecture on the same chromosome are matched by closest distance in cM between subsets of the the same length (if target has 2 QTL on chr 3 and sample has 3, consider all pairs from sample to find closest pair in 2-D). Measure per sample is sum across all loci. A quick way to generate a target is to use qb.BestPattern.

The score.type is "recombination" = r, the recombination rate; "attenuation" = 1-2r; "sq.atten" = squared attenuation, "distance" in cM, or genetic "variance".

Brian S. Yandell, yandell@stat.wisc.edu

http://www.qtlbim.org

qb.BestPattern, qb.hpdone

data(qbExample)

## Here target is known for simulated data.
target <- data.frame(chrom = c(1,1,2,3),
  locus = c(15,45,12,15))
temp <- qb.close(qbExample, target)
summary(temp)
plot(temp)
plot(temp, category = "nqtl")

## Or use target from qb.BestPattern
target <- summary(qb.BestPattern(qbExample))$best
qb.close(qbExample, target)