02739nam a2200205 a 450000100080000000500110000800800410001910000160006024501920007626001080026850000290037652019910040565300220239670000150241870000160243370000240244970000180247370000210249170000210251220384852023-03-23       2015    bl uuuu     u00u1 u    #d1 aLIMA, B. M.  aGenome-wide prediction of individual tree ranking for growth, chemical and physical wood properties in Eucalyptus based on High-Density SNP Data from the EUChip60K.h[electronic resource]  aIn: PLANT & ANIMAL GENOME CONFERENCE, 23., 2015, San Diego, CA. [Abstracts...]. San Diego: [s.n.]c2015  aPAG 2015. Pôster P1230.  aDespite their fast growth, eucalypt breeding cycles still take &gt12 years and wood properties assessment is limited to a sample of trees in late stages of selection, preventing the full exploitation of the existing genetic variation. We examined fifteen growth and wood chemical and physical traits in 1,000 progeny individuals derived from an elite Eucalyptus breeding population of International Paper do Brasil. Near-infrared spectroscopy models were developed and used for high-throughput wood phenotyping. A total of 29,090 SNPs were retained for genomic prediction analyses, providing 1 SNP/22 kbp in the Eucalyptus genome, a higher marker density than any previous study in trees. RR-BLUP and BLASSO were used for genomic predictions, and compared to the performance of BLUP phenotypic selection. Predictive abilities reached similar estimates with the two approaches, varying from a low of 0.10, for microfibril angle, to 0.42 for volume growth, and up to 0.83, for syringyl:guaiacyl lignin ratio. Correlations between genomic and phenotypic predictions ranged between 0.771 and 0.929, and were best for wood chemical traits, density and growth. Both genomic prediction models yielded a coincidence &gt70% for the top 30 trees ranked by phenotypic selection for volume growth, wood density and S:G ratio, and &gt60% in the top 10 trees. When tandem multi-trait selection was applied to these traits simultaneously, 15 out of the top 25 trees selected based on phenotypes were also selected by genomic selection (GS). These results corroborate earlier results by which GS could significantly reduce the length of a breeding cycle in Eucalyptus by applying ultra-early selection of genomically multi-trait ranked seedlings precluding the progeny trial stage. Top ranked seedlings for GEBV would be subject to early flower induction and inter-mated to create the next generation of breeding and/or deployed into field validation clonal trials depending on the breeding objectives.  aEucalypt breeding1 aGARCIA, C.1 aALMEIDA, A.1 aSILVA JUNIOR, O. B.1 aVENCOVSKY, R.1 aMANSFIELD, S. D.1 aGRATTAPAGLIA, D.