Anderson and Karel: Limited genetic structure of Brevoortia patronus revealed by microsatellite markers 
79 
Table 4 
Statistics from the Whichloci analysis of 14 microsatellite loci in the genus Brevoortia. The average rank, 
locus score, relative score (the power of an individual locus for species discrimination, relative to the power 
of all 14 loci), and cumulative score were calculated from locus scores across 100 replicated simulations of 
species assignment for Gulf Menhaden (Brevoortia patronus), Finescale Menhaden (B. gunteri), or Yellowfin 
Menhaden ( B . smithi). The LOD (log odds) columns indicate whether inclusion of a locus was necessary 
(“1”) or not necessary (“0”) for correct assignment at the given level of stringency. 
Locus 
Avg. rank 
Score 
Relative score (%) 
Cumulative score 
LOD=2 
LOD=3 
LOD: 
AF039657 
1 
0.938 
18.45 
18.45 
1 
1 
1 
Bp500 
2 
0.900 
17.70 
36.15 
1 
1 
1 
Bp473 
3 
0.763 
15.00 
51.14 
1 
1 
1 
Bp531 
4 
0.701 
13.79 
64.93 
0 
1 
1 
AF049462 
5 
0.384 
7.54 
72.48 
0 
1 
1 
Bp489 
6 
0.324 
6.38 
78.85 
0 
1 
1 
Bp003 
7 
0.252 
4.95 
83.80 
0 
0 
1 
Bp039 
8 
0.238 
4.67 
88.48 
0 
0 
1 
Bpl21 
9 
0.209 
4.11 
92.59 
0 
0 
1 
Bp221 
10 
0.183 
3.60 
96.18 
0 
0 
1 
Bp239 
11 
0.117 
2.30 
98.48 
0 
0 
1 
Bp275 
12 
0.071 
1.39 
99.88 
0 
0 
1 
AF039658 
13 
0.006 
0.12 
100.00 
0 
0 
1 
Bp230 
14 
0.000 
0.00 
100.00 
0 
0 
1 
derson and Karel, 2007) and potentially could be the 
case with Gulf Menhaden and Finescale Menhaden 
(Anderson and McDonald, 2007). In the case of hybrids 
of Gulf Menhaden and Yellowfin Menhaden, FI indi- 
viduals usually carry the mtDNA haplotype of Yellow- 
fin Menhaden. Assuming that hybridization is common 
in the eastern Gulf of Mexico, mtDNA species identifi- 
cation potentially would lead to overestimation of the 
presence of Yellowfin Menhaden in catch data if hybrid 
populations are encountered because hybrids would be 
identified incorrectly as pure Yellowfin Menhaden. Hy- 
brids were not encountered in our current study; how- 
ever, to obtain individuals of unequivocal taxonomic 
classification, sampling was directed toward locales 
where hybridization had not previously been observed. 
The uncertainty in mtDNA monophyly among 
Brevoortia species led Lynch et al. (2010) to question 
the taxonomic validity of Gulf Menhaden and Atlantic 
Menhaden. Given this uncertainty, species identifica- 
tion in these species by DNA analysis would be accom- 
plished most reliably with multiple co-dominant loci. 
The analysis of Anderson and Karel (2007) indicated 
that species identification in the presence of hybrids 
could be achieved with as few as 5 microsatellites, al- 
though there was some uncertainty in estimation of 
levels of introgression in hybrid individuals. 
In our study, the 14 loci used in simulated species- 
identification analysis (Pqwsim analysis) included not 
only 3 of the loci used in Anderson and Karel (2007) 
but also an additional 11 loci, some of which performed 
at least as well in species identification as the original 
loci. The Whichloci analysis indicated that the 6 most 
informative loci in this group could be used to distin- 
guish between Gulf Menhaden, Yellowfin Menhaden, 
and Finescale Menhaden with 100% accuracy, and with 
a statistical error rate of a=0.001. Accuracy in species 
identification can be achieved with a statistical error 
rate of a=0.0001 given use of all 14 loci. These loci, 
therefore, represent a powerful tool in species identifi- 
cation, and can be used to reliably distinguish all Gulf 
of Mexico species (and hybrids) for the purpose of fu- 
ture stock enhancements. 
To our knowledge, only a single previous study has 
attempted to explicitly examine the genetic structure 
of Gulf Menhaden in the Gulf of Mexico with intensive 
sampling. 3 This earlier study was conducted with 5 mi- 
crosatellite loci and a single mtDNA locus; the Powsim 
analysis used here indicates that this level of genetic 
sampling is inadequate to answer questions about fine- 
scale population structure in this species. Success in 
assignment of individuals to populations is a function 
of both the number of loci used and the polymorphism 
(number of alleles) at each locus (Bernatchez and 
Duchesne, 2000). The loci described here improve both 
of those aspects of population assignment and repre- 
sent a significant improvement in the tools available 
3 Anderson, J. D. 2006. Conservation genetics of Gulf Men- 
haden (Brevoortia patronus): implications for the manage- 
ment of a critical forage component for Texas coastal game- 
fish ecology. Federal Aid in Sportfish Restoration Act Tech- 
nical Series, F-144-R, 34 pp. 
