28 
Fishery Bulletin 1 12(1) 
Table 3 
Multiplex group (panel), locus, fluorescent label (dye), repeat motif, and primer concentrations (pM) for poly- 
merase chain reactions for 10 Cobia-specific microsatellite loci (modified from Renshaw et al., 2006). Summary 
statistics for each locus were based on the total combined project data set. ra=sample size, Na= number of al- 
leles, A=allelic size range. 
Panel 
Locus 
WellRED dye 
Repeat motif 
[Primer] 
n 
Na 
A 
1 
Rcal-HIO 
D2 
CA 
0.10 
514 
11 
116-138 
Rcal-A04 
D4 
(CA) 9 (CACT) 4 
0.05 
516 
11 
180-210 
RcalB-EQ2 
D4 
CT 
0.15 
516 
7 
301-317 
2 
Real-All 
D4 
GT 
0.05 
513 
22 
166-212 
RcalB-H09 
D2 
GATA 
0.09 
516 
17 
169-233 
RcalB-E08A 
D3 
CA 
0.11 
516 
12 
201-247 
RcalB-C06 
D4 
GATA 
0.05 
514 
22 
329-417 
3 
RcalB-DlO 
D3 
CTAT 
0.13 
515 
28 
132-260 
Real-Ell 
D2 
CA 
0.04 
518 
6 
167-181 
Rcal-C04 
D4 
GT 
0.13 
516 
17 
221-261 
those from offshore areas (defined as captured seaward 
of the barrier islands, mostly near wrecks or reefs); 
samples without sufficient collection details for assign- 
ment were excluded (Table 2). Samples collected from 
the Port Royal Sound and St. Helena Sound estuaries 
were pooled for the SC inshore location because tagged 
fish recaptured after being at liberty for only a short 
period indicated substantial movement betv/een these 
adjacent and connected estuaries. 
An exact G-test with Markov Chain permutations, 
as implemented in Genepop, vers. 4.0.10 (Raymond 
and Rousset, 1995), was used to test for pairwise 
differences in genotypic distributions among collec- 
tion locations during each collection year. Markov 
chain parameters included 10,000 dememorizations, 
100 batches, and 5000 iterations per batch. Part of 
the analyses of genetic structure used i?sT statistics, 
which incorporate a stepwise mutation model to es- 
timate population differentiation and are analogous 
to Fgx statistics (Slatkin, 1995). Pairwise, hierarchi- 
cal RgT statistics were calculated and an analysis of 
molecular variance (AMOVA) was conducted — as im- 
plemented in Arlequin with 10,000 iterations to de- 
termine the degree of genetic structuring occurring 
among states. Structure, vers. 2.3 (Pritchard et ah, 
2000) also was used to infer subpopulations through 
a clustering-based algorithm. The Admixture with 
LocPrior model was implemented, including runs of 
£=1-4 for the 2008 collection and £=1-3 for the 2009 
collection. All models incorporated 10,000 iterations 
and a 10,000 burn-in period and were replicated in 
triplicate. Pairwise comparisons of genotypic distri- 
butions and i?sT calculations between 2008 and 2009 
collection locations were conducted as described pre- 
viously to determine the degree of temporal genetic 
stability of Cobia populations along the southeast- 
ern U.S. Atlantic coast as well as to validate tempo- 
ral pooling of samples for genetic characterization of 
detected populations. Populations were characterized 
genetically by calculation of the average number of 
alleles per locus (N a ), observed heterozygosity (Hq), 
genetic diversity (Hq; Nei, 1987), and inbreeding coef- 
ficients (Fis) by using Arlequin and Genepop. 
Results 
For this project, 764 individuals were genotyped. For 
both collection years, high proportions of loci were able 
to be scored unambiguously with high repeatability 
(>99%), resulting in low levels of missing data (2008: 
1.6%; 2009: 0%). Genetic data from 519 samples were 
used for all loci testing (Table 2). All loci at all collec- 
tion locations were found to be in HWE (P>0.05), with 
no evidence of null alleles (frequency<0.06) and no 
indication of linkage disequilibrium between any loci 
(critical P-value after Bonferonni=0.001). All 10 loci 
were polymorphic with allelic richness ranging from 6 
to 28 (mean: 15.4 alleles per locus) (Table 3). The x 2 
test for comparing hatchery broodstock and offspring 
indicated that all loci are inherited in a Mendelian 
fashion (Table 4). 
The loci suite provides an average nonexclusion 
parent-pair probability of ’..3x10 7 and average non- 
exclusion identity probability of 5.8xl0 -12 , signifying 
that the possibility of misassignment of parentage in 
the parentage analysis is substantially less than 0.01% 
and individuals can be identified confidently. Therefore, 
on the basis of initial tests, our suite of microsatel- 
lite markers is valuable for characterization of popula- 
tion genetic diversity and genetic structure, as well as 
for parentage analysis because the loci are genetically 
