qtlhot (version 1.2.10)

GetCandReg: Get genetic information on candidate regulators and co-mapping traits.

Description

Get chromosome (phys.chr) and physical position in cM (phys.pos), along with the LOD score (peak.lod) at the peak position (peak.pos), and the chromosome where the peak is located (peak.chr). Some candidates may map to the same chromosome where they are physically located.

Usage

GetCandReg(highobj, annot, traits)
GetCoMappingTraits(highobj, cand.reg)

Value

`GetCoMappingTraits` returns a list with each element being the names of co-mapping traits for a particular name in `traits`. `GetCandReg` returns a data frame while `GetCisCandReg` returns a list with a similar candidate regulator data frame as the element `cis.reg`, and the index of trait names as the element `cis.index`. The elements of the candidate regulator data frame are as follows (`peak.pos.lower` and `peak.pos.upper` only for `GetCisCandReg`):

gene

name of trait, which might be a gene name

phys.chr

chromosome on which gene physically resides

phys.pos

physical position (in cM)

peak.chr

chromosome where peak LOD is located

peak.pos

position of peak (in cM)

peak.lod

LOD value at peak

peak.pos.lower,peak.pos.upper

lower and upper bounds of the 1.5-LOD support interval around `peak.pos`

Arguments

highobj

data frame from `highlod`, which is sparse summary of high LODs in large scanone object

annot

data frame with annotation information; must have first column as unique identifier, third column as chromosome, and fifth column as position in cM; typically column 2 has gene name, and column 4 has position in Mb

traits

names of traits to examine as candidate regulators; names must correspond to phenotypes in `cross` object

cand.reg

data frame with candidate regulator; see value section below

Author

Elias Chaibub Neto

Details

Traits that map to positions close to their physical locations are said to map in cis (local linkages). Traits that map to positions away from their physical locations are said to map in trans (distal linkages). There is no unambiguous way to determine how close a trait needs to map to its physical location in order to be classified as cis. Our choice is to classify a trait as cis if the 1.5-LOD support interval (Manichaikul et al. 2006) around the LOD peak contains the trait's physical location, and if the LOD score at its physical location is higher the the LOD threshold. The function `GetCisCandReg` determines which of the candidate regulators map in cis. The function `GetCoMappingTraits` returns a list with the putative targets of each trait. A trait is included in the putative target list of a trait when its LOD peak is greater than `lod.thr` and the `drop` LOD support interval around the peak contains the location of the trait's QTL. The function `JoinTestOutputs` currently relies on external files that contain results of `FitAllTests`. It needs to be rewritten to save space.

References

Manichaikul et al. (2006) Genetics

See Also

`highlod`, `FitAllTests`, scanone

Examples

if (FALSE) {
# Create CMSTCross object
example(SimCrossCausal)
# data(CMSTCross) loaded lazily
CMSTscan <- scanone(CMSTCross, pheno.col = 1:3, method = "hk")
CMSThigh <- highlod(CMSTscan)
traits <- names(CMSTCross$pheno)
annot <- data.frame(name = traits, traits = traits, chr = rep(1, 3),
 Mb.pos = c(55,10,100))
annot$cM.pos <- annot$Mb.pos
cand.reg <- GetCandReg(CMSThigh, annot, traits)
cis.cand.reg <- GetCisCandReg(CMSThigh, cand.reg)
comap.targets <- GetCoMappingTraits(CMSThigh, cand.reg)
}