ncase: Function for estimating effective size (Nc) and sex rate (s)

In ssoubeyrand/CloNcaSe: Estimation of Effective Size and Sex Rate of a Partially Clonal Population

Description

This function computes point estimates of Nc and s for a partially clonal population, confidence intervals (using parametric bootstrap) and the coverage rates of the intervals. The parametric bootstrap is also used to propose a bias correction of the point estimates and their confidence intervals.

Usage

ncase(counts, nb1ClonalCycles, nb2ClonalCycles, start.param, maxit = 500, 
	nbBootstrap = 0, conf.level = 0.95, nbCoverage = 0, trace = TRUE)

Arguments

counts	2-column matrix whose 1st (resp. 2nd) column gives the counts of strains in sample 1 (resp. 2).
nb1ClonalCycles	Non-negative integer indicating the number of clonal cycles separating the 1st sample and the mixed (i.e. clonal-sexual) cycle.
nb2ClonalCycles	Non-negative integer indicating the number of clonal cycles separating the mixed (i.e. clonal-sexual) cycle and the 2nd sample.
start.param	2-column matrix whose rows give initial values for the effective size (Nc) and the sex rate (s) that are used in the optimization of the contrast.
maxit	Maximum number (non-negative integer) of iterations in the optimization of the contrast. Default to 500.
nbBootstrap	Number (non-negative integer) of bootstrap replications to assess confidence intervals (based on a normal approximation) and to carry out a bias correction. Default to 0 (no bootstrap procedure).
conf.level	Confidence level (between 0 and 1) of the marginal confidence intervals for Nc and s. Default to 0.95. Used if nbBootstrap>0.
nbCoverage	Number of (non-negative integer) iterations to assess the coverage rates of the marginal confidence intervals. Default to 0 (no coverage assessment). Used if nbBootstrap>0.
trace	Logical indicating if tracing information are displayed (TRUE) or not displayed (FALSE). Default to TRUE. Tracing information consist of a message every 100 iterations in the bootstrap procedure and in the assessment of the coverage rates of the confidence intervals.

Details

The optimization method that is used is the Nelder and Mead (1965) method implemented in the optim function (stats package).

Value

A list whose items are:

bootstrap.estim	(if nbBootstrap>0) Estimates of Nc (1st col.) and s (2nd col.) obtained for bootstrap replications.
cover	(if nbCoverage>0) Coverage rates of the marginal intervals for Nc and s when no bias correction is applied (1st col.) and when bias correction is applied (2nd col.).
optim.output	A list consisted of (i) the output of the optim function (optim is applied to the logarithm of Nc and the logit of s) and (ii) the optimal values of Nc and s (transformed back to the original units of Nc and s).
estimates	A numeric vector providing the estimates of Nc and s (if nbBootstrap=0) or a matrix providing the estimates of Nc and s, their bias-corrected versions, and the associated confidence intervals (if nbBootstrap>0).

Author(s)

Samuel Soubeyrand

References

Ali S, Soubeyrand S, Gladieux P, Giraud T, Leconte M, Gautier A, Mboup M, de Vallavieille-Pope C and Enjalbert J (2015). CloNcaSe: Estimation of sex frequency and effective population size by clonemate re-sampling in partially clonal organisms. Research Report.

See Also

simul, ncase4simul.

Examples

## Yellow rust data
data(yellowrust)

## Point estimates
ncase(yellowrust,nb1ClonalCycles=17,nb2ClonalCycles=7,
	start=expand.grid(c(100,1000),c(0.2,0.8)))

## Point estimates, Parametric bootstrap, Bias correction 
## and confidence intervals

ncase(yellowrust,nb1ClonalCycles=17,nb2ClonalCycles=7,
	start=expand.grid(c(100,1000),c(0.2,0.8)),
	nbBootstrap=5000,conf.level=0.95)


## Point estimates, Parametric bootstrap, Bias correction
## and confidence intervals and their coverage

ncase(yellowrust,nb1ClonalCycles=17,nb2ClonalCycles=7,
	start=expand.grid(c(100,1000),c(0.2,0.8)),
	nbBootstrap=5000,conf.level=0.95,nbCoverage=100)

ssoubeyrand/CloNcaSe documentation built on May 30, 2019, 8:43 a.m.