Hard-to-find Papers
November 2006
Main Index
January 2003
Feb-Mar 2003
Apr-May 2003
June-Aug 2003
Sept-Oct 2003
Nov-Dec 2003
Jan-Feb 2004
Mar-Apr 2004
May-June 2004
July 2006
June 2006
August 2006
Sept 2006
Oct 2006
Nov 2006
Dec 2006

These aquaculture- and conservation-oriented commentaries are not abstracts written by the original authors.  They reflect the opinions of someone else -- usually Roger Doyle.  Direct quotations from the papers or abstracts are marked with inverted commas.

564.  Projected tilapia and carp invasion of the Americas
         Invasive potential of common carp (Cyprinus carpio) and Nile tilapia (Oreochromis niloticus) in American freshwater systems. 2006. Zambrano, L., E. Martínez-Meyer, N. Menezes and A. T. Peterson. Canadian Journal of Fisheries and Aquatic Sciences 63:1903-1910.
         A Geographic Information System (GIS) approach is used to predict the potential -- or inevitable -- distribution of common carp and O. niloticus in the Americas. Ecological and environmental information from the native distributions of these species in Asia and Africa is used as "training data" for the computer program, which then proceeds to identify candidate locations in the Americas where there appears to be niche space available (by the Hutchinsonian definition of niche).
         Tilapia should be able to invade the southeastern United States and the coastal lowlands of Mexico and Central America. In South America the possible range of O. niloticus includes much of Brazil, Argentina, Venezuela and Guyana. The predicted carp distributions are more to the north and south in the lowlands, and at higher elevations in the tropics.
         Of course these areas are not currently free of fish; the invaders muscle into a niche through successful competition and predation directed at the locals. zambrano@ibiologia.unam.mx

563.  Does competition among individuals change their breeding values?
         Incorporation of competitive effects in forest tree or animal breeding programs. 2005. Muir, W. M. Genetics 170:1247-1259.
         Competition strongly affects growth rate in fish. Competition may in fact be the most important source of size variation among individuals in a community tank. Tilapia and carp are notoriously prone to stunting their growth in the presence of larger fish.
         Stunting often involves a behavioural (or endocrinological) feedback loop which magnifies initial size variation even when plenty of food is available to everyone. See Dec 2000 #148, Oct 2003 #427, Feb 2004 #456. At the genetic level, a gene which has a direct positive effect on the growth of an individual will have a negative effect on the growth of others in the group. All of this is ignored in standard genetic models of covariation, which may explain why the results of BLUP-based index selection for growth have often been disappointing in aquaculture, often conspicuously so.
         This paper develops a model in which "mixed-model methodology (BLUP) is utilized to separate effects on the phenotype due to the individuals' own genes (direct effects) and those from competing individuals (associative effects) and thereby to allow an optimum index selection on those effects". The model was tried out in a selection experiment on an ill-tempered and cannibalistic population of quail.
         Size-at-age selection using standard BLUP selection made no progress but selection on the index including competitive effects worked much better. "The differences in response show that competitive effects can be included in breeding programs, without measuring new traits, so that costs of the breeding program need not increase."
         The details of the model are well explained both mathematically and intuitively. The author plainly wants to be helpful and, in an appendix, provides SAS code for solving one of the examples in the paper. bmuir@purdue.edu . (Also see another paper which provides BLUPF90 script for applying the same model to larger problems: J. Anim. Sci. 2005. 83:1241-1246; arangoj@uga.edu.) bmuir@purdue.edu

562. Where did Vietnamese WSSV come from?
         Molecular epidemiology of white spot syndrome virus within Vietnam. 2004. Dieu, B. T., H. Marks, J. J. Siebenga, R. W. Goldbach, D. Zuidema, T. P. Duong and J. M. Vlak. Journal of General Virology 85:3607-3618.
         The WSSV virus shows relatively little variation, either in its genome or its protein make-up, which suggests that it has emerged recently or recently experienced a severe bottleneck. RFLP markers are used in this paper to develop a hypothesis for the biogeography of the spread of WSSV through Vietnam. The authors propose that WSSV entered Vietnam by multiple introductions from a common ancestor on one side or the other of the Taiwan Strait. From there, it probably spread outwards to Thailand, North Vietnam, China (Hainan) and Cambodia, mutating slightly as it went. See Aug 2001 #222.  just.vlak@wur.nl 

561.  Marine protected areas are good for genetic diversity as well as for species diversity
         Effects of fishing protection on the genetic structure of fish populations. 2006. Pérez-Ruzafa, Á., M. González-Wangüemert, P. Lenfant, C. Marcos and J. A. García-Charton. Biological Conservation 129:244-255.
         Marine protected areas (MPA) are proposed or already established in many parts of the world, but while their usefulness for conservation at the population level has been demonstrated, we don't know much about their value at the genetic level.
         "This paper analyses the effects of fishery protection on the structure of populations of Diplodus sargus [White sea bream], a target species, in protected and non-protected areas of the western Mediterranean. ... Protected areas have significantly higher allelic richness. The lower levels of heterozygosis and higher heterozygote deficit showed by islands compared with coastal areas makes clear the importance of considering the connectivity processes when designing a [MPA]."
         Allelic richness was the most statistically significant measure of the effectiveness of reserves, as might be expected because of its known sensitivity to genetic drift (See Aug 2006 #524, Oct 2006 #543 for other papers on allele richness in conservation.) The authors emphasize that because a lot of the genetic diversity is found between populations, "the design of marine protected areas must take into account the spatial heterogeneity in the genetic structure of populations and the connectivity between protected and non-protected populations as well as between MPA network constituents".
         In other words, a marine protected area which maintains a population at a viable level will not necessarily retain or even capture all the relevant genetic diversity. angelpr@um.es

560.  Identification of wild trout families for optimal conservation
         Sibship within samples of brown trout (Salmo trutta) and implications for supportive breeding. 2005. Hansen, M. M. and L. F. Jensen. Conservation Genetics 6:297-305.
         Small natural populations of trout must, by definition, consist of a limited number of families. But what does "limited" mean in terms of our technical ability to detect family groups? Can we actually find them with commonly available technology?
         In this study, brown trout resident in two small Danish rivers were compared with a large anadromous trout population and with simulated control populations. Microsatellite markers at 8 loci did reveal the family structure of the resident trout samples; the program SPAGEDI found meaningful pairwise similarities and the program COLONY (Wang) found groups of full-sibs nested within half-sibs.  "The expected increase of inbreeding coefficient in the two [small river] samples due to family structure was 0.026 and 0.030 respectively."  This is indeed meaningful information in the context of supportive breeding, where it is advantageous to avoid mating within sibs and to capture as much family diversity as possible in the founder generation (Jan 2002 #283). See also Feb 2004 #464 for trout metapopulations and #556, below, for schooling sibs. mmh@dfu.min.dk 

559.  Salmon maturation depends more on growth rate than body size
         Does size matter most? The effect of growth history on probabilistic reaction norm for salmon maturation. 2006. Morita, K. and M. Fukuwaka. Evolution 60:1516-1521.
         Early or "precocious" maturation is a serious problem for aquaculture, and an important component of the reproductive strategy of wild salmon. It is under strong genetic control, shows marked local adaptation in the wild, and is environmentally manipulated in aquaculture through the use of lights and special feeding regimes. Something similar happens with tilapia under crowded conditions. No one knows for sure exactly how it works in either species.
         Reaction norm is a convenient technical phrase denoting the distribution of phenotypes expressed by a particular genotype (or strain, or population) under a variety of environmental conditions. The reaction norm describes the outcome of a genotype's interaction with its environment; for example, variation in the light regime. Early maturation in an aquacultural broodstock can thus be described in terms of the reaction norms of the available genotypes in the available environments.
         For salmon, it is known that one environmental factor is body weight, often recorded as the weight at which 50% of the animals are mature. However, there are an infinite number of growth trajectories different animals can follow to reach -- eventually -- a given body weight. So growth trajectories are also the outcomes of genotypes interacting with environments.  "Therefore, to understand the evolution of the maturation schedule, it is necessary to comprehend the relationships among body size, growth history, and maturation schedule." Thorpe, Bromage and many others have been making this point for decades.
         The authors of this paper have found something interesting. "Previous growth history was found to be more closely linked to maturation probability than body size. The most recent growth condition was the most important factor affecting whether a fish matured during the subsequent breeding season." See Sept 2006 #530.
         Growth rings on scales were used to infer growth, size and maturation status. (This also works with tilapia scales, by the way.) Logistic regression with "mature/immature" as the response variable was used to gauge the relative importance of body size vs. annual growth increment. The latter was much more important, in general.
         Note that while this is not a genetic paper per se it provides a good measurement tool for genetic comparison of strains or for selection within strains. moritak@affrc.go.jp 

558.  How antimicrobial diversity is generated in Litopenaeus vannamei
         Genomic structure and transcriptional regulation of the penaeidin gene family from Litopenaeus vannamei. 2006. O'Leary, N. A. and P. S. Gross. Gene 371:75-83.
         Shrimps produce peptides called penaeidins which are effective antimicrobial and antifungal agents (see Dec 2003 #453 and references therein). Like the peptide receptors of the vertebrate MHC immune system, penaeidins are highly diverse both within and between individual shrimp. This study helps explain how the genetic diversity of penaeidins is generated.
         There are three classes of penaeidins which are coded by different genes. The genes are so different that the variety of peptide products is unlikely to result from alternative splicing of RNA messengers. There appears to be no post-transcriptional cutting-and-splicing as in the vertebrate MHC.
         Instead, "genomic DNA sequence analysis indicates that each penaeidin class is encoded by its own unique gene with full independent coding potential, indicating that post-transcriptional mechanisms are not responsible for ... diversity. Rather, genomic content [i.e. DNA nucleotide variation in the peptide coding region of the genes] accounts for all expressed penaeidin classes and isoforms found in multiple individual shrimp."
          "Quantitative real-time PCR was used to demonstrate that the penaeidin genes are expressed at dramatically different levels". The authors conclude that there must be variation in the regulatory portions of the genes as well. See Oct 2006 #549. grossp@musc.edu

557.  Can we estimate breeding values of non-identified animals grown together?
         Predicting breeding values and accuracies from group in comparison to individual observations. 2006. Olson, K. M., D. J. Garrick and R. M. Enns. Journal of Animal Science 84:88-92.
         A lot of useful aquacultural information is measured on groups of animals, not individuals. Examples are feed intake or food conversion ratio in a tank, percent survival in a challenge test, fry production from a batch of spawners.
         It is often easy to arrange matters so that each tank contains a known set of families. Is quantitative genetic analysis of pooled data possible, e.g. when every individual in the tank is assigned the same average value? How accurate are estimates of breeding value derived from pooled data?
         This theoretical and simulation study of certain analytical designs (sire model, maternal grandsire model) shows that the loss of accuracy with pooled data in not as serious as one might expect, especially when the number of individuals in each tank is small.  Furthermore, designs in which each tank contains animals from only one (sire) family are not especially sensitive to tank effects, even when the effects cannot be estimated. (Tank effects can be estimated if every family occupies more than one tank and every tank contains more than one family.)
         It seems, then, that collecting aquacultural data on pooled individuals might be considered as a possibly cost-effective tradeoff between genetic accuracy (= more rapid progress) and the overall expense of a breeding program. The paper (which is actually about cows, by the way) clearly explains the simple adjustments to the mixed model which are needed to handle pooled data. Surprisingly, the intuitive approach which simply assigns each individual in a pool the same (average) value doesn't work too badly. Competition is not included in this model (#563, above). dorian.garrick@colostate.edu 

556.  Schoolmates?
         Migratory charr schools exhibit population and kin associations beyond juvenile stages. 2005. Fraser, D. J., P. Duchesne and L. Bernatchez. Molecular Ecology 14:3133-3146.
         The authors sampled individuals from a large number of schools of  brook charr (or brook trout; Salvelinus fontinalis) from two populations and, using microsatellite markers, determined that the number of siblings found in a school was greater than expected based on assumed random mixing within the population.
         In other words, charr siblings tend to be found in the same school, just like people siblings. And for similar reasons: "We discuss the hypothesis that the stable kin groups, rather than arising from kin selection, may instead be a by-product of familiarity based on individual selection for the maintenance of local adaptations related to migration (natal and feeding area philopatry)."
         "Our results are noteworthy because they suggest that there is some degree of permanence in the composition of wild fish schools. Additionally, they support the hypothesis that schools can be hierarchically structured (from population members down to family groups) and are thus nonrandom genetic entities." See #560, above. dylan.fraser@dal.ca

555.  Tilapia: a better procedure for comparing growth rates of different strains
         Growth response of Nile tilapia fry to salinity stress in the presence of an 'internal reference' fish. 2005. Basiao, Z. U., R. V. Eguia and R. W. Doyle. Aquaculture Research 36:712-720.
         Comparing the growth rates of different strains can be difficult in aquaculture, especially when facilities for replication are limited. The problem can be alleviated by rearing the strains together in the same tank, but then you have the problem of telling them apart -- by the time they are big enough to tag they may already have developed different sizes or growth rates for environmental reasons. Molecular markers can solve this problem but are still expensive.
         An alternative approach is to use a visually-distinguishable third strain as a standard reference, or "common denominator" in a series of pairwise tests. (See for example Feb 2002 #294.) Each of the two test strains is grown from birth in the presence of reference fish, and by comparing the test strains to the common reference they can be compared to each other.
          A statistical power analysis conducted on this experiment in the Philippines with three strains of tilapia showed the procedure to be effective.  "Two-way analysis of variance (ANOVA) revealed no significant strain differences (P=0.081; r2=0.106). However, analysis of covariance with the internal reference strain used as a covariate showed significant (P=0.049; r2=0.638) strain effects on specific growth (based on standard length measurements). ... The use of internal reference strain as a covariate improved the r2 from 0.106 to 0.638 and increased the efficiency of the test in detecting a true difference." zbasiao@compass.com.ph 

554.  How to identify families when parents have not been well sampled
         Pasos (parental allocation of singles in open systems): a computer program for individual parental allocation with missing parents. 2005. URL Duchesne, P., T. Castric and L. Bernatchez. Molecular Ecology Notes 5:701-704.
         This parental assignment program (using co-dominant markers) is designed to detect missing parents when not all parents have been sampled. It would be useful, for instance,  in shrimp breeding where a number of males usually die after reproducing but before they can be sampled.
         Assumptions must of course be made to achieve this, in particular random mating and binomially-distributed reproductive success. (The latter assumption implies an equal probability of producing offspring, but not a strictly equal numbers of offspring.) The authors say some violation of these assumptions can be tolerated and the program looks to be useful.
         The interface resembles the familiar PAPA program, by the same authors, where sampling is assumed to be complete. Both programs can be downloaded from  http://www.bio.ulaval.ca/louisbernatchez/links.htm#soft_parent_anal  .

553.  Estimating effective population size from a single sample
         A bias correction for estimates of effective population size based on linkage disequilibrium at unlinked gene loci. 2006. Waples, R. S. Conservation Genetics 7:167-184.
         Effective population size (Ne) is an important parameter in applied conservation genetics because it is the starting point for speculation & meditation about the rate of inbreeding and loss of genetic diversity in small populations. The value of Ne in a wild population can be estimated in several ways from sample data on neutral markers such as microsatellites.
         One such procedure is especially useful in that it requires only one sample from the population rather than two or more samples taken at different times.  The value of Ne is inferred from the disequilibrium (lack of random Mendelian assortment) among alleles at different marker loci within genotypes. Disequilibrium is a function of Ne.
         A companion paper in this issue of Conservation Genetics (England et al., 7(2), pp. 303-308, phillip.england@csiro.au)  reports that the single-sample estimate of Ne can be severely biased downwards if  the number of individuals in the sample is smaller than Ne itself.
         The paper noted here (Waples) offers several modifications to the standard calculation, which "effectively eliminate the bias in ...  most cases. The modified method also performs well in estimating Ne in non-ideal populations with skewed sex ratio or non-random variance in reproductive success". robin.waples@noaa.gov