qtl (version 1.42-8)

summary.scanone: Summarize the results of a genome scans

Description

Print the rows of the output from scanone that correspond to the maximum LOD for each chromosome, provided that they exceed some specified thresholds.

Usage

# S3 method for scanone
summary(object, threshold,
        format=c("onepheno", "allpheno", "allpeaks", "tabByCol", "tabByChr"),
        perms, alpha, lodcolumn=1, pvalues=FALSE,
        ci.function=c("lodint", "bayesint"), …)

Arguments

object

An object output by the function scanone.

threshold

LOD score thresholds. Only peaks with LOD score above this value will be returned. This could be a single number or (for formats other than "onepheno") a threshold for each LOD score column. If alpha is specified, threshold should not be.

format

Format for the output. See Details, below.

perms

Optional permutation results used to derive thresholds or to calculate genome-scan-adjusted p-values. This must be consistent with the object input, in that it must have the same number of LOD score columns, though it can have just one column of permutation results, in which case they are reused for all LOD score columns in the scanone output, object. (These can also be permutation results from scantwo, which permutations for a one-dimensional scan.)

alpha

If perms are included, this is the significance level used to calculate thresholds for determining which peaks to pull out. If threshold is specified, alpha should not be.

lodcolumn

If format="onepheno", this indicates the LOD score column to focus on. This should be a single number between 1 and the number of LOD columns in the object input.

pvalues

If TRUE, include columns with genome-scan-adjusted p-values in the results. This requires that perms be provided.

ci.function

For formats "tabByCol" and "tabByChr", indicates the function to use to get approximate confidence intervals for QTL location.

…

For formats "tabByCol" and "tabByChr", additional arguments are passed to the function indicated by ci.function (for example, drop for lodint or prob for bayesint, or expandtomarkers for either).

Value

An object of class summary.scanone, to be printed by print.summary.scanone.

Details

This function is used to report loci deemed interesting from a one-QTL genome scan (by scanone).

For format="onepheno", we focus on a single LOD score column, indicated by lodcolumn. The single largest LOD score peak on each chromosome is extracted. If threshold is specified, only those peaks with LOD meeting the threshold will be returned. If perms and alpha are specified, a threshold is calculated based on the permutation results in perms for the significance level alpha. If neither threshold nor alpha are specified, the peak on each chromosome is returned. Again note that with this format, only the LOD score column indicated by lodcolumn is considered in deciding which chromosomes to return, but the LOD scores from other columns, at the position with maximum LOD score in the lodcolumn column, are also returned.

For format="allpheno", we consider all LOD score columns, and pull out the position, on each chromosome, showing the largest LOD score. The output thus may contain multiple rows for a chromosome. Here threshold may be a vector of LOD score thresholds, one for each LOD score column, in which case only those positions for which a LOD score column exceeded its threshold are given. If threshold is a single number, it is applied to all of the LOD score columns. If alpha is specified, it must be a single significance level, applied for all LOD score columns, and again perms must be specified, and these are used to calculate the LOD score threshold for the significance level alpha.

For format="allpeaks", the output will contain, for each chromosome, the maximum LOD score for each LOD score column, at the position at which it achieved its maximum. Thus, the output will contain no more than one row per chromosome, but will contain the position and maximum LOD score for each of the LOD score columns. The arguments threshold and alpha may be specified as for the "allpheno" format. The results for a chromosome are returned if at least one of the LOD score columns exceeded its threshold.

For format="tabByCol", there will be a separate table for each LOD score column, with a single peak per chromosome. Included are columns indicating chromosome, peak position, lower and upper limits of the confidence interval calculated via lodint or bayesint, and lod score.

The output for format="tabByChr", is similar to that of format="tabByCol", but with results organized by chromosome rather than by LOD score column.

If pvalues=TRUE, and perms is specified, genome-scan-adjusted p-values are calculated for each LOD score column, and there are additional columns in the output containing these p-values.

In the case that X-chromosome specific permutations were performed (with perm.Xsp=TRUE in scanone), autosome- and X-chromosome specific thresholds and p-values are calculated by the method in Broman et al. (2006).

References

Broman, K. W., Sen, , Owens, S. E., Manichaikul, A., Southard-Smith, E. M. and Churchill G. A. (2006) The X chromosome in quantitative trait locus mapping. Genetics, 174, 2151--2158.

See Also

scanone, plot.scanone, max.scanone, subset.scanone, c.scanone, summary.scanoneperm c.scanoneperm

Examples

# NOT RUN {
data(fake.bc)
# }
# NOT RUN {
fake.bc <- calc.genoprob(fake.bc, step=5)

# genome scan by Haley-Knott regression
out <- scanone(fake.bc, method="hk")

# permutation tests
# }
# NOT RUN {
operm <- scanone(fake.bc, method="hk", n.perm=1000)
# }
# NOT RUN {
# peaks for all chromosomes
summary(out)

# results with LOD >= 3
summary(out, threshold=3)

# the same, but also showing the p-values
summary(out, threshold=3, perms=operm, pvalues=TRUE)

# results with LOD meeting the 0.05 threshold from the permutation results
summary(out, perms=operm, alpha=0.05)

# the same, also showing the p-values
summary(out, perms=operm, alpha=0.05, pvalues=TRUE)

##### summary with multiple phenotype results
out2 <- scanone(fake.bc, pheno.col=1:2, method="hk")

# permutations
# }
# NOT RUN {
operm2 <- scanone(fake.bc, pheno.col=1:2, method="hk", n.perm=1000)
# }
# NOT RUN {
# results with LOD >= 2 for the 1st phenotype and >= 1 for the 2nd phenotype
#     using format="allpheno"
summary(out2, thr=c(2, 1), format="allpheno")

# The same with format="allpeaks"
summary(out2, thr=c(2, 1), format="allpeaks")

# The same with p-values
summary(out2, thr=c(2, 1), format="allpeaks", perms=operm2, pvalues=TRUE)

# results with LOD meeting the 0.05 significance level by the permutations
#     using format="allpheno"
summary(out2, format="allpheno", perms=operm2, alpha=0.05)

# The same with p-values
summary(out2, format="allpheno", perms=operm2, alpha=0.05, pvalues=TRUE)

# The same with format="allpeaks"
summary(out2, format="allpeaks", perms=operm2, alpha=0.05, pvalues=TRUE)

# format="tabByCol"
summary(out2, format="tabByCol", perms=operm2, alpha=0.05, pvalues=TRUE)

# format="tabByChr", but using bayes intervals
summary(out2, format="tabByChr", perms=operm2, alpha=0.05, pvalues=TRUE,
        ci.function="bayesint")

# format="tabByChr", but using 99% bayes intervals
summary(out2, format="tabByChr", perms=operm2, alpha=0.05, pvalues=TRUE,
        ci.function="bayesint", prob=0.99)
# }