12.07.2015 Views

marker-assisted selection in wheat - ictsd

marker-assisted selection in wheat - ictsd

marker-assisted selection in wheat - ictsd

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

320Marker-<strong>assisted</strong> <strong>selection</strong> – Current status and future perspectives <strong>in</strong> crops, livestock, forestry and fishrate of 6.7 x 10 -4 when 13 464 gametes and<strong>marker</strong> <strong>in</strong>formation from 12 loci werescreened <strong>in</strong> a turbot population. Mutationsand scor<strong>in</strong>g error have the sameeffect on exclud<strong>in</strong>g potential parents, giv<strong>in</strong>grise to <strong>in</strong>correct assignments.• Unl<strong>in</strong>ked loci and l<strong>in</strong>kage equilibrium.L<strong>in</strong>kage and l<strong>in</strong>kage disequilibriumbetween the loci will reduce the effectivenumber of loci used for the parentalassignment. Note that the power to assignparental pairs correctly can thereby differbetween different sets of microsatellites.Estoup et al. (1998) quantified the difference<strong>in</strong> the power of microsatellite <strong>marker</strong>sets used to assign parents correctly <strong>in</strong>turbot (eight loci, eight alleles per loci)and ra<strong>in</strong>bow trout (eight loci, four allelesper loci) populations by calculat<strong>in</strong>g thefrequency of good and unique parentassignments (f gu ). The more variable setof microsatellites resulted <strong>in</strong> higher f gu forlarger maximal mat<strong>in</strong>g schemes for turbotthan the less variable set of microsatellitesfor ra<strong>in</strong>bow trout. In general, a set of lociwith an equal number of alleles has thehighest exclusion probability (Weir, 1996;Jamieson and Taylor, 1997).Walk-back <strong>selection</strong> schemesPractical breed<strong>in</strong>g schemes us<strong>in</strong>g molecular<strong>marker</strong>-based parental assignment havebeen reported. Doyle and Herb<strong>in</strong>ger (1994)proposed carry<strong>in</strong>g out parentage assignmentfor <strong>in</strong>dividuals us<strong>in</strong>g genetic <strong>marker</strong>s,such that full-sib families would not have tobe kept separately until tagg<strong>in</strong>g but, rather,would be held <strong>in</strong> one large tank. Note that<strong>in</strong>dividuals that are genotyped also need tobe physically tagged, so that genotyp<strong>in</strong>gresults can be traced back to particular<strong>in</strong>dividuals. At the time for <strong>selection</strong>, fish<strong>in</strong> the tank were first ranked on their phenotypicvalue, assum<strong>in</strong>g that <strong>selection</strong> wasfor only one trait that could be measuredon the <strong>selection</strong> candidates (e.g. weight).Then the <strong>in</strong>dividual with the highest phenotypicvalue was selected and genotypedfor family identification. Thereafter, the<strong>in</strong>dividual with the second highest phenotypicvalue was genotyped and selected if itwas not a full- or half-sib of other, alreadyselected<strong>in</strong>dividuals, such that with<strong>in</strong>-family<strong>selection</strong> was performed. This procedurecont<strong>in</strong>ued until the desired number ofbrood stock was selected. This approachof progress<strong>in</strong>g through the performancerank<strong>in</strong>g, genotyp<strong>in</strong>g and select<strong>in</strong>g the bestperform<strong>in</strong>g<strong>in</strong>dividuals with<strong>in</strong> families wastermed “walk-back” <strong>selection</strong>. Mat<strong>in</strong>gssubsequently would be made betweenfamilies, a strategy preferred because itwould keep the rate of <strong>in</strong>breed<strong>in</strong>g low(Falconer and Mackay, 1996). Herb<strong>in</strong>geret al. (1995) reported sett<strong>in</strong>g up a ra<strong>in</strong>bowtrout breed<strong>in</strong>g programme us<strong>in</strong>g the walkback<strong>selection</strong> programme of Doyle andHerb<strong>in</strong>ger (1994) and genetic <strong>marker</strong>s toestimate full/half relationships among thecandidates us<strong>in</strong>g a likelihood ratio methodand thereafter perform<strong>in</strong>g with<strong>in</strong>-family<strong>selection</strong>.Us<strong>in</strong>g stochastic simulations, Sonesson(2005) studied a comb<strong>in</strong>ation of optimumcontribution <strong>selection</strong> and walk-back <strong>selection</strong>.Optimum contribution is a <strong>selection</strong>method that maximizes genetic ga<strong>in</strong> witha restriction on the rate of <strong>in</strong>breed<strong>in</strong>g(see earlier <strong>in</strong> this chapter). Hence, thecomb<strong>in</strong>ation of optimum contribution andwalk-back <strong>selection</strong> ensures that the rateof <strong>in</strong>breed<strong>in</strong>g is under control, while thegenetic ga<strong>in</strong> is higher than <strong>in</strong> the with<strong>in</strong>family<strong>selection</strong> schemes of Doyle andHerb<strong>in</strong>ger (1994) because <strong>selection</strong> is bothwith<strong>in</strong> and between families. In the studyby Sonesson (2005), batches of candidateswere pre-selected from a s<strong>in</strong>gle tank on

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!