Selective improvement of rainbow trout: assessment of potential in UK strains.

Abstract

The research assessed the potential of developing a selective breeding programme for the UK rainbow trout industry. Levels of genetic variation at 12 microsatellite loci were first compared in seven different commercial strains. The Observed heterozygosity ranged from Ho = 48.1% in a gold rainbow trout strain (GTR) to Ho = 66.4% in a newly derived broodstock population constructed from a number of different sources (GIT). The Expected Heterozygosity (He) was highest in GIM1 (He= 79.5%) and lowest in the GTR strain (He = 56.9%). The Effective number of alleles (Mae) showed that the GIM1, GIM2, GIM3, and GIT strain (5.4; 5.2; 4.8; 4.2) were significantly more variable than the other strains and that GTR strain had the lowest value (2.5). There appears to be substantial genetic variability within the commercial United Kingdom rainbow trout strains surveyed in this study. This appears to be the case despite very different management histories and levels of record keeping. The strains appear to be genetically distinct (based on population genetic analyses), though the reasons for this remain unclear (and possibly unanswerable given the poor records kept by the different companies).

The Glenwyllin farm strains (GIM) were chosen to form the base population for the project because of their high genetic variability, disease free status and because the farm produced around 20 million ova per year, so any genetic gains would have a widespread impact. The farm has an early (Strain A) and a late spawning (Strain B) and these were mated in a partial factorial design, 20 females and 20 neomales per strain (A & B) were chosen on the basis of maturity and gamete quality in November 2002 so that each male was crossed to 4 females (2 in the same strain and 2 in the other), a total of 160 families were created. All broodstock were biopsied to enable them to be genotyped. The families were reared separately up to the eyed stage at which point the eggs from each family were divided into three to generate three communal replicate populations. One of these was sent to a fingerling producer (Iwerne Spring) for ongrowing to fingerling size and formed the basis of a commercial production trial at Test Valley Trout farm (TVT) in Hampshire.

When the fish reached an average weight of 5 g they were transferred from Iwerne Spring to TVT and 1500 were randomly selected, PIT tagged and biopsied to enable them to be assigned to their family using 11 multiplexed microsatellite loci. Parental assignment was based on exclusion (FAP) but the results were compared with another parental assignment based on likelihood (PAPA). Of the 1500 offspring (OIM) PIT tagged 1242 82.8% could be assigned to a single family utilizing different combinations of more than 6 loci (6 to 11).

The growth of the 1500 OIM fish was tracked throughout the grow out period before they were finally harvested and fully processed. The results of OIM strain at the end of the trial period were mean weight of 415.5 g, and a mean length of 314.5 mm. The visual measurement of colour gave a mean flesh colour values of 26.01 on the 20-34 scale (SalmoFan™), and 11.0 with the colotimetry evaluation of colour (a*). The heritability results for the IOM strain were 43 ± 9% for weight, 42 ± 9% for gutted, and 28 ± 8% for length. The heritability estimates for the visual colour variables were 19 ± 7% and when using the colorimeter, the red chromaticity (a*) heritability was 14 ± 6%. Therefore, the heritability results of the IOM strain indicate that there are opportunities of substantial and rapid improvement of the growth rate and flesh colour traits. Also no line effects were observed or indications of non-additive genetic variation. In contrast to these last results, the overall survival of the GIM strain from the time of the physical tagging with PIT until harvest was 52.8%, and survival heritability was extremely low, 3 ± 2%, hardly significant.