Truncation selection for BLUP-EBV and phenotypic values in fish breeding schemes

We have compared rate of inbreeding (ΔF) and genetic gain (ΔG) in truncation selection schemes for fish on best linear unbiased prediction (BLUP) estimated breeding values (EBVs; BTS) or phenotypic values (PTS) with one constraint on the number of tanks (100 or 200) to rear the families, e.g., until tagging size, and one constraint on the total number of selection candidates (3200 or 6400). Single trait selection was practiced for a trait with heritability of 0.1, 0.4 or 0.7 and 5, 10, 20, 50, 100 or 200 selected sires and dams were selected. When going from PTS to BTS, we found an overall increase of ΔG and ΔF of a factor of 0.07 and 1.45, respectively; that is, the increase of ΔG was moderate and accompanied by a dramatic increase of ΔF. In general, ΔF was reduced by increasing the number of tanks at a given number of selection candidates and by decreasing the number of selection candidates at a given number of tanks. ΔF was also reduced for schemes with high heritability because BLUP breeding value estimation then increases the weight on the individual's phenotype, which leads to a reduced correlation between BLUP estimated breeding values within a family, and fish from more families become selected. In practical fish breeding schemes, PTS seems preferable over BTS because BTS generally give unacceptable high ΔF (?0.01) for these schemes with few but large families. PTS can, however, only be used for traits measured on the selection candidate, which makes PTS less valuable for schemes with comprehensible breeding goals, including, for example, growth, disease resistance, maturity and fillet quality traits. Several traits are measured today on sibs only.