Non-additive genetic effects in Nile tilapia

Abstract

Nile tilapia (Oreochromis niloticus) is the world’s fastest growing aquaculture species, in terms of annual increase in production (ca 10%) and is today produced world-wide. Most of the commercial and farmed Nile tilapia strains are derived from the genetically improved farmed tilapia (GIFT) strain established in the early 1990s. The systematic mixing of the eight different strains during the first 3 generations of GIFT, would prompt a hypothesis that there may have been substantial non-additive genetic effects in the base. Despite having large full-sib families, which enables the estimation of non-additive genetic effects, it is a paradox that these effects have been ignored in the commercial evaluations and the design of the breeding schemes. Thus, this thesis sets out to explore the possibility of utilizing non-additive genetic effects in Nile tilapia breeding programs using a purpose-bred population. This was achieved firstly by using classical methods utilizing pedigree to partition the variances into additive, dominance, maternal and environmental components (paper I), then by developing the necessary genomic resources (SNP-array and linkage maps) to better allow tilapia breeding to exploit new opportunities (paper II) and finally applying these resources to partition the variance components further into additive, dominance, epistasis and maternal environmental components based on marker information (paper III).

The Onil50-array for Nile tilapia containing more than 58K SNPs was developed from the whole genome sequence of 32 Nile tilapia sampled from the GenoMar breeding nucleus (paper II). The SNPs on the array were selected based on even physical distribution and the polymorphic information content. SNP performance of the array was evaluated on nearly 5000 samples, revealing high-quality genotype data for 43,588 SNPs. Then the integrated physical and genetic linkage map, containing 40,186 SNPs, was constructed. Most of the Linkage groups (LGs) were found to have sigmoid recombination pattern, with the recombination rate between males and females being 1:1.2.

A diallel design with reciprocal cross was applied to partition the variance components. The analyses by pedigree showed the presence of non-additive variation (paper I), identified to a large extent as a full-sib family component that was not associated with additive effects or maternal effects. This source is commonly assumed to arise from dominance using pedigree. Further analysis using data from the Onil50 SNP-array (paper III) showed that this variation, when present, was associated with additive-by-additive epistasis, and not dominance. These estimates were corrected for departures from HWE. Detrimental effects of inbreeding using genomics was reported for the commercial traits of Nile tilapia.

Substantial contributions of the non-additive genetic effects were observed (P<0.05) for two traits: body weight at harvest (BWH) and body depth (BD). Further, substantial contribution of maternal variance (P<0.05) was observed for BWH, BD, fillet weight (FW) and body length (BL); estimates based on both genomics and pedigree approaches being comparable. Unlike non-additive genetic effects, including maternal component in the models was found to cause substantial consequences on the rankings for both genomic and pedigree BLUP models. Thus, ignoring maternal effect was found to inflate the heritability and introduce bias in the genetic evaluations, over-predicting the potential gains.

Rather than depending on the interactions of unknown genes contributing to the non-additive genetic variance, targeted exploitation of this effect in the future in Nile tilapia breeding program might depend on finding out the genes or genomic regions associated with the heterosis for the traits. However, the creation of maternal lines in Tilapia breeding schemes may be a possibility if this variation is found to be heritable. Though the marker information has been used in tilapia breeding for parental assignment for almost 20 years already, the genomic resources developed here have opened a new door of genomic era in Nile Tilapia breeding and has also the potential to improve the genetic gain through genomic selection.

Nil-tilapia (Oreochromis niloticus) er den akvakulturarten som har raskest produksjonsvekst i verden, med ca 10 % årlig økning, og produseres i dag stort sett over hele verden. De fleste av de kommersielle Niltilapiastammene er kommer fra den såkalte GIFT-stammen (Genetically improved farmed tilapia) som ble etablert tidlig på 1990-tallet. Her ble åtte ulike stammer systematisk blandet i løpet av de første 3 generasjonene, noe som gir grunn til å tro at det kan ha blitt generert betydelige ikkeadditive genetiske effekter i denne populasjonen. Til tross for at tilapia har store fullsøsken￢familier, som gjør det mulig å estimere ikke-additive genetiske effekter, har disse effektene hittil blitt ignorert i planlegging av avlsdesign og når en gjør avlsverdivurderinger. Derfor har denne studien valgt å undersøke muligheten for å utnytte disse ikke-additive, genetiske effektene i en Niltilapia-avlspopulasjon. Ved å bruke klassiske analysemetoder, der vi kun tar hensyn til slektskap mellom individer, beregnet vi additive-, dominans-, maternale- og miljøeffekter (artikkel 1), deretter utviklet vi en såkalt SNP-chip (SNP er en type markører) og et genetisk koblingskart, som er nyttige verktøy når en vil gjøre seleksjon i en ny art som tilapia (artikkel 2) og til slutt brukte vi disse verktøyene for å enda bedre kunne skille mellom additive-, dominans-, maternale- og miljøeffekter, som i artikkel 1, men i tillegg også additive samspillseffekter, dvs epistatiske geneffekter (artikkel 3).