Identify inferred partitions in mapping QTL to a phylogenetic tree
Identify the inferred partitions for a chromosome from the results of scanPhyloQTL.
inferredpartitions(output, chr, lodthreshold, probthreshold=0.9)
output |
An object output by the function
|
chr |
A character string indicating the chromosome to consider. (It can also be a number, but it's then converted to a character string.) |
lodthreshold |
LOD threshold; if maximum LOD score is less than this, the null model is considered. |
probthreshold |
Threshold on posterior probabilities. See Details below. |
We consider a single chromosome, and take the maximum LOD score for
each partition on that chromosome. The presence of a QTL is inferred
if at least one partition has LOD score greater than
lodthreshold. In this case, we then convert the LOD scores for
the partitions to approximate posterior probabilities by taking
10^LOD and then rescaling them to sum to 1.
These are sorted from largest to smallest, and we
then take as the inferred partitions the smallest set whose posterior
probabilities cumulatively add up to at least probthreshold.
A vector of character strings. If the null model (no QTL) is
inferred, the output is "null". Otherwise, it is the set of
inferred partitions.
Karl W Broman, broman@wisc.edu
Broman, K. W., Kim, S., An\'e, C. and Payseur, B. A. Mapping quantitative trait loci to a phylogenetic tree. In preparation.
# example map; drop X chromosome
data(map10)
map10 <- map10[1:19]
# simulate data
x <- simPhyloQTL(4, partition="AB|CD", crosses=c("AB", "AC", "AD"),
map=map10, n.ind=150,
model=c(1, 50, 0.5, 0))
# run calc.genoprob on each cross
## Not run: x <- lapply(x, calc.genoprob, step=2)
# scan genome, at each position trying all possible partitions
out <- scanPhyloQTL(x, method="hk")
# inferred partitions
inferredpartitions(out, chr=3, lodthreshold=3)
# inferred partitions with prob'y threshold = 0.95
inferredpartitions(out, chr=3, lodthreshold=3, probthreshold=0.95)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.