Molecular Ecology Notes (2005)
© 2005 Blackwell Publishing Ltd
Blackwel Publishing, Ltd.
: a program to identify problem loci and samples for
noninvasive genetic samples in a capture-mark-recapture
K. S. M
KELVEY and M. K. SCHWARTZ
USDA Forest Service, Rocky Mountain Research Station, 800 E. Beckwith, Missoula, MT 59801, USA
Genotyping error, often associated with low-quantity/quality DNA samples, is an important
issue when using genetic tags to estimate abundance using capture-mark-recapture (CMR).
, an MS-Windows program, identifies both loci and samples that likely contain
errors affecting CMR estimates.
uses a ‘bimodal test’, that enumerates the number
of loci different between each pair of samples, and a ‘difference in capture history test’ (DCH)
to determine those loci producing the most errors. Importantly, the DCH test allows one to
determine that a data set is error-free.
(2004) and is now available online.
: allelic dropout,
, genotyping error, mark recapture, molecular tagging, noninvasive
Received 09 March 2005; revision accepted 14 April 2005
The fields of ecology, wildlife management and conservation
biology have embraced identification of species and
individuals using noninvasively collected genetic samples.
Projects using noninvasive genetic samples can provide a
wealth of data on rare and elusive species once thought
impossible to accurately count. Unfortunately, using DNA
from noninvasive samples can also lead to genotyping
errors which can bias estimates unless carefully controlled
(Waits & Leberg 2000; Creel
Genotyping errors often occur when DNA is collected from
low quality samples (hair, feathers, faeces, etc.; Taberlet
. 1996; Morin
. 2001). The most common genotyp-
ing errors are allelic dropout, the preferential amplification
of one of two alleles, false alleles, amplification products
that mimic true alleles, and various laboratory errors such
as misreading bands or transcription errors. Molecular
ecologists have recognized the importance of genotyping
error and arrived at multiple solutions. Programs
(Valiere 2002) and
. 2002) are useful
for evaluating errors when samples have been multitubed
. 1996). Program
2000) is useful when pedigree information is available.
compares randomly constructed genotypes to observed
genotypes, locating errors due to stutter and short allele
dominance. Other approaches include quantifying the
amount of extracted DNA and avoiding analysis of samples
with low yield (Morin
Here we provide details on a program,