CIMMYT

Displaying 1 - 2 of 2

Bekele Abeyo
CIMMYT-Ethiopia
Primary Author Email: 
b.abeyo@cgiar.org
Poster or Plenary?: 
Poster
BGRI Year: 
2015
Abstract Tags: 
geographic_area: 

CIMMYT wheat germplasm flow to Ethiopia started in the late 1960s. Over 90 bread wheat varieties were released over the decades. Of these, about 77% had CIMMYT origins or were derived from CIMMYT materials. Wheat is a traditional rainfed crop grown by 5 million small-scale farmers on 1.6 ha more or less. Yields have increased from 1.0 t/ha in the 1960s to 2.54 t/ha in 2014 mainly due to high yielding semi-dwarf bread wheat varieties and modern agronomic practices. Using such technologies, better farmers often get 5-6 t/ha. The rusts are the most important production constraints. For example, the 2010 yellow rust epidemic debilitated the mega varieties Kubsa and Galama in the highlands. In 2013/14, stem rust caused up to 100% yield losses in the widely adopted bread wheat variety Digalu in Arsi and Bale. This epidemic was caused by Pgt race TKTTF, which is virulent to the gene SrTmp that is present in Digalu, but is avirulent to Sr31, which is overcome by race Ug99 (TTKSK) and derivatives. To avert the increasing threat of rusts, CIMMYT developed a shuttle breeding program where germplasm moves back and forth between Mexico and Kenya and has increased nursery testing sites (Holetta, Kulumsa, Debre Zeit, Sinana, Adet, and Melkassa) in Ethiopia from two to six. The germplasm passes through rigorous tests against major diseases during both the main- and off-seasons. To obtain high yielding rust resistant germplasm, many hundreds of genotypes were introduced and tested over the last two years. In 2014/15, 266 (25%) lines with multiple disease resistances and high yield were promoted to national trials. CIMMYT continues to be an important source of germplasm. Fast tracked variety testing and release, accelerated seed multiplication, demonstration and popularization of new varieties with high yield, multiple disease resistance, and acceptable quality will continue.

Juliana
Plant Breeding and Genetics Section, School of Integrative Plant Science, Cornell University, USA
Primary Author Email: 
fp228@cornell.edu

The quest for durable rust resistance in wheat is burgeoning with the emergence of new virulent races. Breeders challenged with this unceasing plant-pathogen arms race have to devise strategies for effective evaluation and exploitation of the rust resistance genes. Considering the likely presence of useful variation for rust resistance in CIMMYT’s international bread wheat screening nurseries (IBWSN), we implemented genomic prediction in the 45th and 46th IBWSN entries to determine their genomic estimated breeding values (GEBV’s) for leaf, stem and stripe rust resistance. The 350 lines (45th IBWSN) and 329 lines (46th IBWSN) were phenotyped in replicated trials over two to three years in El Batan, Mexico (leaf rust); Njoro, Kenya (stem rust) and Toluca, Mexico (stripe rust). The filtered genotyping data for these two nurseries comprised of 6,786 and 11,218 genotyping by sequencing (GBS) markers. Our objective was to compare the GEBV’s estimated by four different models: multiple linear regression (MLR) with QTL-linked markers as fixed effects; Genomic-best linear unbiased prediction (G-BLUP); G-BLUP mixed model which includes QTL linked markers as fixed effects and Bayesian least absolute shrinkage and selection operator (LASSO). We observed that the prediction accuracies (calculated using 10-fold cross validation) were the highest for stripe rust (0.52 to 0.61), followed by stem rust (0.42 to 0.65) and leaf rust (0.15 to 0.45). Among the models, the MLR gave the lowest prediction accuracies (0.15,0.42 and 0.52), while G-BLUP (0.45,0.59 and 0.59), mixed G-BLUP (0.38,0.65 and 0.62) and the Bayesian LASSO (0.45,0.58 and 0.61) yielded relatively higher and almost similar accuracies. Overall, our results are promising and indicate that using genome-wide markers is advantageous than including only significant QTL-linked markers. We hope that implementing genomic prediction in breeding programs, would help to achieve rapid gains from selection and revolutionize our efforts in combating the rust pathogen.

Subscribe to CIMMYT