secr (version 3.2.1)

stoatDNA: Stoat DNA Data

Description

Data of A. E. Byrom from a study of stoats (Mustela erminea) in New Zealand. Individuals were identified from DNA in hair samples.

Usage

stoatDNA

Details

The data are from a pilot study of stoats in red beech (Nothofagus fusca) forest in the Matakitaki Valley, South Island, New Zealand. Sticky hair-sampling tubes (n = 94) were placed on a 3-km x 3-km grid with 500-m spacing between lines and 250-m spacing along lines. Tubes were baited with rabbit meat and checked daily for 7 days, starting on 15 December 2001. Stoat hair samples were identified to individual using DNA microsatellites amplified by PCR from follicular tissue (Gleeson et al. 2010). Six loci were amplified and the mean number of alleles was 7.3 per locus. Not all loci could be amplified in 27% of samples. A total of 40 hair samples were collected (Gleeson et al. 2010), but only 30 appear in this dataset; the rest presumably did not yield sufficient DNA for genotyping.

The data are provided as a single-session capthist object `stoatCH'. Hair tubes are `proximity' detectors which allow an individual to be detected at multiple detectors on one occasion (day), but there are no multiple detections in this dataset and for historical reasons the data are provided as detector type `multi'. Three pre-fitted models are included: stoat.model.HN, stoat.model.HZ, and stoat.model.EX (with halfnormal, hazard-rate and negative exponential detection functions, respectively).

Object	Description
stoatCH	capthist object
stoat.model.EX	fitted secr model -- null, exponential detection function
stoat.model.HN	fitted secr model -- null, halfnormal detection function
stoat.model.HZ	fitted secr model -- null, hazard-rate detection function

References

Gleeson, D. M., Byrom, A. E. and Howitt, R. L. J. (2010) Non-invasive methods for genotyping of stoats (Mustela erminea) in New Zealand: potential for field applications. New Zealand Journal of Ecology 34, 356--359. Available on-line at http://www.newzealandecology.org.

See Also

capthist, Detection functions, secr.fit

Examples

# NOT RUN {
summary(stoatCH)

# }
# NOT RUN {
stoat.model.HN <- secr.fit(stoatCH, buffer = 1000, detectfn = 0)

# this generates an error unless we use biasLimit = NA
# to suppress the default bias check
stoat.model.HZ <- secr.fit(stoatCH, buffer = 1000, detectfn = 1,
    biasLimit = NA)

stoat.model.EX <- secr.fit(stoatCH, buffer = 1000, detectfn = 2)
confint(stoat.model.HN, "D")
## Profile likelihood interval(s)...
##         lcl        ucl
## D 0.01275125 0.04055662
# }
# NOT RUN {
## plot fitted detection functions
xv <- seq(0,800,10)
plot(stoat.model.EX, xval = xv, ylim = c(0,0.12), limits = FALSE,
    lty = 2)
plot(stoat.model.HN, xval = xv, limits = FALSE, lty = 1, add = TRUE)
plot(stoat.model.HZ, xval = xv, limits = FALSE, lty = 3, add = TRUE)

## review density estimates
collate(stoat.model.HZ, stoat.model.HN, stoat.model.EX,
    realnames = "D", perm = c(2,3,4,1))
## use secr:: in case of conflicting model.average from RMark
secr::model.average(stoat.model.HN, stoat.model.EX,
    realnames = "D")

# }