Current status of the US Aquaculture Industry

In 2002, seafood imports were the second largest contributors to the United States (US) trade deficit (non-combustible), reflecting a long-term trend of increased demands without a sufficient increase in supply (Harvey 2003). In recent years, the US Department of Agriculture, or USDA, has responded to the seafood trade deficit by increasing federal funding opportunities to aquaculture researchers and producers alike. Problems affecting the industry have been outlined by both technical and academic experts, and the funding has since been directed into work being conducted in those areas of recognized need. One such area is that of genetic improvement of aquatic animal growth (NRI 2004).

Growth in aquatic animals, or more specifically fish, is a relatively simple idea: a fish consumes feed, the feed is converted into tissue, and as the tissue accumulates the fish grows (Sumpter 1992). As straightforward as the idea appears, there are countless differences among fish in their ability to carry out the growth processes. Genetic studies have indicated that there is room for improvement of fish growth through selection efforts (Gjedrem 2000), and as a result, several breeding programs have been initiated by the USDA throughout the country. For example, the USDA National Center for Cool and Cold Water Aquaculture in Leetown, West Virginia is currently developing several genetic stocks of rainbow trout with improved growth performance. It is intended that the improved stocks will be made available to the US industry in the coming years. However, a valid question arises as to whether the improved stocks will still show improved growth when they are moved to a new environment (i.e. eggs shipped to a different state for culture). Specifically, how will the phenotypes (i.e. growth rate) of the improved genetic stocks (or genotypic population) respond when the environment changes? The answer ultimately rests in our past and current understanding of genotype-environment interactions, our ability to measure the interactions, and finally on how such interactions might surface in the phenotype (i.e. fish growth) of interest.

A historical perspective of genotypes and environments

As stated by Gupta and Lewontin (1982), our modern understanding of a genotype and the environment is that, “the phenotype of an individual is a unique consequence of its genotype interacting with the environments encountered…” However, in the late 1940’s and early 1950’s, a consensus for the influence of the environment on a phenotype had not yet been attained. The influence of the environment on a phenotype was undergoing heated debates throughout the field of quantitative genetics, especially in regards to animal selection and breeding. One of the earliest theses on the subject was presented by Hammond (1947) and addressed the question of whether the environment was an important influence in determining a phenotype, or character, especially when the phenotype was undergoing selection. Hammond’s extensive thesis predicted that, “genes determining the expression of the character selected (were) mainly the same in both good and bad environments,” (Falconer and Latyszewski 1952) or simply put, the environment doesn’t interact with the genotype. According to the thesis, a superior genotype selected for in one environment would remain superior regardless of the environment in which it was found, relative to another genotype.

Scientific evidence such as that presented by Falconer and Latyszewski (1952) has since negated the ‘no interaction’ theory presented by Hammond. For example, Falconer and Lateyszewski (1952) conducted a study whereby mice were selected on the basis of improved growth in two environments (full or restricted rations). Their results demonstrated that selection for improved growth of mice under full rations was not maintained under restricted 50% rations, as it should have been if Hammond’s thesis were true. Instead, the results showed that the improved growth responses of selected mice did not carry over when the mice were moved to a different environment. Such results clearly demonstrated that the phenotype (i.e. improved growth) was dependent upon both the genotype (mouse family) and environment (ration size) for phenotype.

In the 1940’s, another thesis was presented to the scientific community by Schmalhausen and was later coined the norm of reaction (Gupta and Lewontin 1982). As described by Gupta and Lewontin (1982), the norm of reaction theorizes that within a given environment, the different genotypes present display a range of phenotypes in response to the environment. Unlike Hammond’s thesis, scientific evidence supported the ‘norm of reaction’ theory (Rawson and Hilbish 1991, Herbert et al. 1998).

Influence and variation of genotypes and environments

The relative importance of the genotype and environment in determining a phenotype is highly variable (Dobzhansky 1950). For example, if two strains of mice were given the same amount of food (identical nutritional environments), the importance of the environment would decrease in determining the weights of the two mice. However, if one mouse was given unlimited food while the other was given only ½ as much, then the nutritional environment would become very important in determining the weights of the two mice. Thus, the phenotype (P) of an individual is a theoretical summation of the genotypic (G) and environmental influences (E) and can be expressed as (Falconer and Mackay 1996):

P = G + E (Equation A)

In a given population, there is often a phenotype that is most common; let’s say a cobalt-colored fish for this example. Quantitatively, the most common phenotype is referred to as the phenotypic mean (Falconer and Mackay 1996). In addition to the phenotypic mean, there is often a range of different and less common phenotypic versions (i.e. light and dark blue color patterns). These different versions reflect variation in the scale color phenotype of the fish, and are quantified as the squared differences from the mean, or variances (Doolittle 1987). These are also the normal phenotypic responses, or reaction norms, for the population. Because the variance can be estimated and has countless quantitative applications (i.e. heritability), equation A from above is often presented in terms of variances (Falconer and Mackay 1996):

V_P = V_G + V_E (Equation B)

Equation B shows that the variation seen (or measured) in a population’s phenotype is the result of different genotypic and environmental variances. Thus, if one wanted to quantify the variance due to scale color, one might be able to measure the variance contributed by the genotype for cobalt-colored scales (V_G) plus the effects of the diet (a component of environment variation, V_E).

Genotype by environment interaction

In many situations, a third term, V_GE, becomes important when measuring a population’s phenotypic variation. The V_GE term represents the variation phenotypes due to an interaction between the genotype and the environment (Falconer and Mackay 1996). For example, suppose the weights of two populations of fish are measured. In one environment feed is restricted, and in the second environment unlimited feed is available. Suppose the two groups of fish represent fast and slow growing genotypes. If there were no genotype by environment interaction, one would expect both groups of fish to grow well in the unlimited feed treatment, with the fast fish growing at a faster rate than the slow fish. A similar result would be expected for the limited feed treatment, except that both groups of fish would grow proportionally slower. However, the fast fish would still grow faster than the slow fish (Figure 1, first graph), regardless of the limited nutritional environment. If a genotype by environment interaction were present, then the fast fish might grow faster than the slow fish in one environment, but not in another. In other words, the phenotype depends not only on the genotype and environment, but it depends specifically upon which genotype and environment are under consideration. Although the above example used fish, the example actually reflects the work of Falconer (1960) who demonstrated the importance of genotype by environment interactions when selecting mice for improved growth rates.

Figure 1. The graphs above provide examples of how genotype and environment influences can affect a phenotype. Growth rates of two genotypic populations (fast and slow growing) are shown within two nutritional environments, 100 or 50 % rations. In the first graph, both genotypic populations respond similarly to the different nutrition environments, indicating there is no genotype by environment interaction. In other words, the fast population always grows faster than the slow population by the same amount, regardless of the environmental effect on the actual growth rate. In the second graph, the fast population grows faster than the slow population only in the 100 % ration environment. This indicates that the performance of the fast genotypic population depends on (interacts with) the nutritional environment, or that it does not respond similarly in each environment. Thus, there is a genotype by environment interaction in the second graph.

Quantifying genotype by environment interactions

When the theses presented by Hammond and Schmalhausen were first presented to the scientific community, there were no satisfactory methods available to actually quantify genotype by environment interactions, and so no way to support or negate either thesis. Genetic tools that are commonplace today were not yet available, so there was no direct way to determine subtle genetic differences. Instead, only changes in phenotypic responses could be measured. Also, while some descriptive research had been conducted, a solid, unbiased statistical model for quantifying and comparing the responses remained elusive. It wasn’t until1952 that a quantitative method was developed to estimate the interaction. In 1952, Falconer presented a theoretical model for quantifying the genotype by environment interaction component of a phenotypic response based upon basic correlation statistics (Falconer 1952). Falconer hypothesized that the genotype by environment influence on a phenotype (i.e. growth rate) going through several generations of selection could be quantified by 1.) measuring the phenotype when exposed to either maximal genotype influences (unlimited food) or maximal environmental influences (limited food supply), then 2.) changing the dominate phenotypic influence that the respective groups were exposed to (i.e restricting feed in the offspring of animals previously given unlimited feed). The magnitude of difference, or correlation, between the subsequent phenotypes of the two groups (max genotype group vs. max environmental influence group) would reflect the genotype by environment interaction. The model was applicable to a single genotype and two environmental treatments. Application, verification, expansion, and subsequent specialization of Falconer’s model quickly ensued in the years following its introduction (Clayton et al. 1956, Falconer 1960, Gillespie and Turelli 1989, Volis et al. 2002).

Today, new tools in biotechnology have allowed for analyses previously unavailable. Genotypes can be identified through the use of genotyping, and so it is now possible to separate environmental and genotypic influences more specifically. In addition to the new analytical tools, new statistical models have also been developed. A recent model proposed by Wong et al (2003) uses simple linear regression analysis to quantify the genotype-environmental interaction of a continuous variable (i.e. blood pressure), a measurement that is not possible using Falconer-based models.

Genotype by environmental interaction in aquatic organisms

We know today that genotype by environmental interactions occur, and that they can be an important contributor to an organism’s phenotype. However, our understanding of the genotype by environmental interactions in aquatic organisms has only just begun. The intimate relationship between aquatic organisms such as fish has long been recognized. For example, growth rates of fish are very sensitive to environmental temperature since fish body temperatures fluctuate with the environment (Sumpter 1992). In terrestrial mammals, however, the influence of environmental temperature is minimal, since their body temperatures are maintained within a very narrow range. Despite this knowledge, few studies have investigated the genotype by environmental interactions in fish. One recent study demonstrated the importance of temperature in sex determination of Atlantic silversides (Menidia menidia), showing that at conception the sex ratio was approximately 1:1 (Conover and Kynard 1981). After exposure to warm summer temperatures, the sex ratios shifted in favor of females (up to 80 %). While the silverside study clearly showed the importance of the genotype by environment in determining a fish’s phenotype, relatively few studies have followed since its publication in 1981, and most of those studies have been conducted outside of the US (Imsland and Jonassen 2001).

Other important environmental influences of fish growth include diet composition and availability, water quality, and photoperiod (Sumpter 1992). For example, Imsland and Jonassen (2001) conducted a study using Atlantic halibut (Scophthalmus maximus Rafinesque) and found that fish growth was sensitive to the length of the light period during the day, with longer light periods showing enhanced growth. What was not expected, however, was that growth was enhanced at lower temperatures and longer day lengths. Usually, warmer temperatures enhance growth. Thus, the authors concluded a strong genotype by environmental interaction must be present (Imsland and Jonassen 2001). Many other interactions likely occur that have not yet been unveiled.

Directions for future research

The importance genotype by environmental interactions in determining a phenotype has been fully recognized. In aquaculture, considerable research is underway to determine the extent of the influences, and to help predict what phenotypic changes can be expected when progeny from selected broodstock are transferred to new locations. For example, sibling progeny from the rainbow trout stocks undergoing selection at the USDA Leetown lab have been transferred to several alternative locations (i.e. North Carolina State University) in an effort to evaluate potential environment by genotype interactions (Jeff Silverstein, personal communication, USDA-NCCCWA, Leetown, WV). In time, it is hoped that a comprehensive understanding of the genotype by environmental interactions will aid in the genetic improvement of aquatic animal growth, and ultimately, help alleviate the US seafood trade deficit.

References

Conover, DO, and BE Kynard. 1981. Environmental sex determination: interaction of

temperature and genotype in a fish. Science 213:577-579.

Dobzhansky, T. 1950. Heredity, Environment, and Evolution. Science 111:161-166.

Doolitle, DP. 1987. Environmental Variance. Pages 197-198 in E. Bresler, GW Thomas,

and LD VanVleck, Editors. Population Genetics, Basic Principles. Springer-Verlag, New York.

Falconer, DS. 1952. The problem of environment and selection. The American

Naturalist 86:293-298.

Falconer, DS. 1960. Selection of mice for growth on high and low planes of nutrition.

Genetical Research 1:91-113.

Falconer, DS, and M Latyszewski. 1952. The environment in relation to selection for

size in mice. Journal of Genetics 51:67-80.

Falconer, DS and Mackay, TFC. 1996. Introduction to Quantitative Genetics. 4^th

Edition. Pearson Education Ltd., Essex, England.

Ferguson, MM, AP Liskauskas, and RG Danzmann. 1995. Genetic and environemtnal

correlates of variation in body weight of brook trout (Salvelinus fontinalis).

Canadian Journal of Fisheries and Aquatic Sciences 52:307-314.

Gjedrem, T. 2000. Genetic improvement of cold-water fish species. Aquaculture

Research 31:25-33.

Gupta, AP and Lewontin, RC. 1982. A study of reaction norms in natural populations of

Drosophila pseudoobscura. Evolution 36(5):934-948.

Hammond, J. 1947. Animal breeding in relation to nutrition and environmental

conditions. Biological Reviews 22:195-213.

Harvey, D. 2003. Aquaculture outlook. Economic Research Service Report No.

LDP-AQS-17. United States Department of Agriculture, Washington, DC.

Hebert, KP, PL Goddard, WW Smoker, and AJ Gharrett. 1998. Quantitative genetic

variation and genotype by environment interaction of embryo development rate in

pink salmon (Oncorhynchus gorbuscha). Canadian Journal of Fisheries and Aquatic Sciences 55:2048-2057.

Imsland, A, and TM Jonassen. 2001. Regulation of growth in turbot (Scophthalmus

maximus Rafinesque) and Atlantic halibut (Hippoglossus hippoglossus L.): aspects of environment x genotype interactions. Reviews in Fish Biology and Fisheries 11:71-90.

NRI (National Research Initiative). 2004. Competitive Grants Program No. 42.0:

Animal growth and nutrient utilization. US Department of Agriculture, Washington, DC.

Rawson, PD, and TJ Hilbish. 1991. Gentoype-environment interaction for juvenile

growth in the hard clam mercenaria mercenaria (L). Evolution 45:1924-1935.

Shearer, KD, JT Silverstein, and WW Dickhoff. 1997. Control of growth and adiposity

of juvenile Chinook salmon (Oncorhynchus mykiss). Aquaculture 157:311-323.

Sumpter, JP. 1992. Control of growth of rainbow trout (Oncorhynchus mykiss).

Aquaculture 100:299-320.

Volis, S, S Mendlinger, and D Ward. 2002. Differentiation in populations of Hordeum

spontaneum Koch along a gradient of environmental productivity and

predictability: plasticity in response to water and nutrient stress. Biological Journal of the Linnean Society 75:301-312.

Wong, MY, DE Day, JA Luan, KP Chan, and NJ Wareham. 2003. The detection of

gene-environment interaction for continuous traits: should we deal with measurement error by bigger studies or better measurement? International Journal of Epidemiology 32:51-57.