Create the second generation of fast-growing lines of YP for the aquaculture industry via simple marker-assisted cohort selection.

Yellow perch (YP) is a particularly important aquacultural and ecological species in the Midwest. The demand for YP has remained very high in GLR since they are the traditional fish species used in local restaurants, social organizations, and the Friday night fish fry dinners that are a staple in many Great Lakes states. Although there are several mature aquaculture industries, such as catfish, trout and salmon in this country, YP has its unique and niche market in GLR.  For example, wholesale prices for YP fillets reached to a peak of $6.50-9.00/lb, and retail prices to $9.00-15.00/lb. Currently, Ohio ranks first in pounds of yellow perch sold in the nation. Despite this opportunity, rapid expansion of the YP aquaculture industry has not occurred in Ohio and GLR. One reason in particular hindering expansion has been relatively slow growth of currently cultured populations of this species. Using current YP strains, only 60% of the fish cultured in aquaculture operations reach market size in a normal growth cycle (16 months), with the rest being below market size. This is an inefficient use of resources, feed, and operational costs, and leads to marginal profits at best. Therefore, improving and promoting YP growth and aquaculture using new technology will significantly improve the profitability of fish farmers.

Historically, the supply of YP largely relied on capture fisheries in the Great Lakes. Wild harvests had declined to 11-18 million pounds per year during the 1980s and 1990s, and are currently limited to less than 6 million pounds per year. Except for Lake Erie (and Green Bay), commercial fishing of YP has been closed in the Great Lakes due to overfishing, and quotas for sport fishing have also been greatly reduced. New virus such as viral hemorrhagic septicemia will further threaten wild YP populations. Increasing perch aquaculture production will reduce the pressure on the natural resource, therefore, sustaining and improving the ecological environment and natural resource in the Great Lakes.

Genetic improvement of aquaculture species offers a substantial opportunity for increasing production efficiency, health, production quality and, ultimately, profitability in aquaculture industries. The potential of these gains has long been recognized as a significant impetus for aquaculture. Gains in profit resulting from genetic improvement have been realized in terrestrial domesticated livestock species, agricultural, horticultural and ornamental plant, forest trees and aquaculture species, such as salmonids, tilapia and catfish. As Sorgeloos (1999) recently suggested, the real challenge for the next decade is to get the aquaculture industries to introduce effective genetic improvement program using selective breeding.

The impact of this proposed project will be primarily via the delivery of superior YP strains to farmers in Ohio and NCR. The greatest return on investment for this project is the ultimate reduction in production costs due to increased growth rate and reduced feed costs. If success in this project is similar to that achieved for striped bass, trout, and catfish, this type of marker-assisted selection should improve growth and/or food conversion rate by 20-25% per generation and have tremendous impact on the GLR and Ohio YP aquaculture industry.

Preliminary Data

We have being conducted selective breeding of YP using marker-assisted cohort selection (MACS) for three years. Eight strains of YP were obtained from eight states and stock evaluations in genetic variation and growth have been completed. Approximately 2,000 genetically superior broodfish were selected as the base breeding population for the long-term selective breeding program. To date, eight improved lines of perch have been produced in the first round of selection, and distributed to a research facility for initial on-station tests. Current research data shows that the improved lines grow 18.3%-39.8% faster than unimproved fish (Wang at el., 2007). Forty-five new microsatellite markers have been developed (Li and Wang et al. 2006), and a method for parentage analysis of YP has been established for MACS in our new established Aquaculture Genetics and Breeding Laboratory (AGBL) at Piketon (Li and Wang et al. 2007). We are proposing to create the second generation of fast-growing lines of YP for the aquaculture industry.  

 

The foundation of our improvement program is to simultaneously create intense selection while maintaining genetic variation in our broodstock.  This requires that at least three specific practices be conducted in the selective breeding component of the program.  First, single-pair mating is necessary to avoid the possibility that milt from only a proportion of the males accounts for the majority of fertilization (Brown et al. 2000) and to ensure accurate retroactive pedigree assignment.  Second, communal rearing is necessary to reduce environmental effects and to conserve pond/tank resources on station.  Third, molecular fingerprinting is necessary to reconstruct the pedigrees of communally reared individuals so that crossing between related fish is minimized each generation (Doyle and Herbinger 1994).  We have employed this integrated strategy because it has been demonstrated successful in providing for high selection intensities, low inbreeding, and extreme economy (Herbinger et al. 1995, Naish and Skibinshi 1998).

Broodstock selection

We will use part of the previously developed 1^st generation of improved lines to create the 2^nd generation of improved lines. When a majority of improved lines created in 2006 has reached harvest size, three hundred best fish (top 5-10% female and male) will be selected for selection line based on their body weight and breeding value. Then they will be tagged with passive integrated transponder (PIT) tags and genotyped at the Aquaculture Genetics and Breeding Laboratory at Piketon using microsetellite marker by non-destructively sampling finclip of each parent. Molecular genetic pedigrees will be determined and a genetic relatedness chart will be constructed. Among the 300 fish, at least 50 pairs of the least related, with highest breeding value will be selected and divided into five cohorts, where each cohort will have 10 females and 10 males. If the constraint on the rate of inbreeding cannot be achieved, another batch of fish will be genotyped and included in the total number of candidates. A selection line will be created by pair-mating at least 10 pairs within each cohort to found the second generation of improved line. The individuals of average weight will be chosen for the control line.

Broodstock genotyping

All the broodfish candidates will be genotyped as done for their parents using eight microsatellite loci we developed and optimized. Microsatellite amplification reactions for each experimental individual will be performed using approximately 100 ng genomic DNA derived from ethanol-preserved tissue. PCR will be performed using BioRad PTC-200 DNA engine thermal cycler to cycle according to Li and Wang (2007a).  Genotyping will be performed using ABI 3130 DNA Sequencing and Genotyping System, and genotypes automatically scored using Genemapper. Individual genotypes will be checked for accuracy and consistency.

Genetic pedigree and relatedness chart construction

Microsatellite profiles for the 8 loci will be used to identify the parents of all the broodfish candidates. Parentage assignment will be performed using the exclusion-based approach implemented in the program CERVUS 2.0 (Marshall et al., 1998). Also the likelihood-based method described in CERVUS 2.0 will be applied to obtain a probability of the most confident parent as an additional support to the exclusion-based strategy. All broodstock individuals will be included as putative candidates. To ascertain the possible influence of kinship among breeders in the deviation of theoretical and actual exclusion power of microsatellites (Marshall et al., 1998), relatedness coefficients (r) between all pairs of breeders from experimental broodstock will be obtained using the statistical package RELATEDNESS (Queller and Goodnight, 1989). Based on the parentage assignment and relatedness coefficients, a microsatellite pedigree and genetic relatedness chart will be constructed for pair-mating.

Mating, spawning and incubation

Single-pair mating based on the genetic pedigree and relatedness chart will be conducted in 50 cm-diameter tanks with flow-through water in early March when fish have reached a mature stage. We will synchronize females with one or two injections of HCG at the dosage of 200 - 600 IU/kg body weight based on their need and maturity. PIT tags will be used to identify and track parents of all families.

The fertilized egg ribbon from each pair-mating will be collected daily from spawning tanks starting 2 days post-injection. Hardened egg ribbons will be transferred to and incubated separately in 30 cm-diameter tanks with flow-through well water at a temperature of 11 C. Incubation rings will be constructed of chicken wire, and egg ribbons will gently be woven in and out of the wire to hold them in place and under water. Aeration will be increased once clumps of eggs or individual eggs are released from the ribbon to keep eggs gently moving in the water. Three days post hatch, fry will be siphoned to a 20 L bucket for stocking to the ponds for nursery.

Pond Fertilization

All nursery ponds will be fertilized with the liquid organic fertilizers ammonium nitrate (28-0-0) and phosphoric acid (0-54-0) once a week starting right after fish spawn until the time fry are ready for harvest, to stimulate phytoplankton and zooplankton blooms. The ammonium nitrate and phosphoric acid will be mixed in pond water and sprayed on the pond surface with a hand sprayer. Fertilizer amounts will be based on the concentrations recommended by Culver (1998). Briefly, 3.16 kg of ammonium nitrate (nitrogen source) and 0.13 kg of phosphoric acid (phosphorus source) will be added to each pond to achieve targeted concentrations of 31.6 kg/ha for nitrogen and 1.3 kg/ha for phosphorus.

Larval stocking and nursery

Equal numbers of 4,000 of newly hatched fry from each hatching tank or family within cohort will be stocked into one of five 0.1-ha ponds by cohort group. Each pond or cohort will ensure to have 20 families (from 25 pair-matings) and 80,000 fry.  Hatched-out fry counts will be estimated volumetrically. This method involves calculating the total volume of the incubation tank, uniformly mixing the fry in the tank, and taking 3 small volumes (100 ml) of water and fry from the tank. Fry will individually be counted in each 100 ml sample, and the 3 samples will be averaged to estimate the total number in each tank. Fry will be nursed in ponds for six weeks before harvesting for feed-training. Feed-training will be conducted in 1m-diameter tanks with flow-through well for three weeks. Fingerlings will be fed starter feed at 8% of body weight using automatic belt feeder during feed training period.

Fingerling stocking and grow-out

Each of 5 cohorts with twenty families will be stocked and communally reared in each of five 0.1-ha research ponds at a rate of 30,000 fish/ha at Piketon. Fish will be fed 3% of body weight with high protein feed (>40%) using automatic feeders. Feeding ration will be adjusted monthly based on previous data. Daily temperature, DO and mortality will also be recorded for each pond. At the end of year one, fish will be harvested and improved line will be evaluated vs. control line for the next generation of broodstock candidates.

Data analysis

Parentage will be assigned by probability tests using PROBMAX (Danzmann, 1997) and CURVES program. After experimental progeny are assigned to family, weight data will be log-transformed before analysis to stabilize heterogeneity of variances due to scale effects and data sets will be checked for homoscedasticity and normality. Differences among families will be tested, and variance components and BLUP breeding values (Henderson, 1984) will be computed for performance traits using restricted maximum likelihood (REML) method in PEST (Groeneveld and Kovac, 1990). The optimum contributions of the candidates will be calculated using the method of Meuwissen (1997). Best linear unbiased predictors for the sires will be computed.