Cornell University

This is a past event. Its details are archived for historical purposes.

The contact information may no longer be valid.

Please visit our current events listings to look for similar events by title, location, or venue.

Genomic predictions with inclusion of environmental covariates to improve cassava for disease resistance and yield

Wednesday, August 8, 2018 at 12:20pm to 1:10pm

Emerson Hall, 135

Cassava is the fourth largest source of calories in the developing countries after sugarcane, maize and rice. However, yield losses due to viral diseases; cassava mosaic disease (CMD) and cassava brown streak disease (CBSD) impact the production of cassava in Asia and Africa. In Sub-Saharan Africa, CBSD is considered more destructive particularly in East, Central and Southern parts of Africa. One major obstacle in breeding cassava for traits of preference is the long breeding cycle (8 to 10 years). Genomic selection, which uses genome-wide DNA markers and phenotypic records from the training population (TP), could help shorten the cycle by enabling estimation of the breeding values (GEBVs) and total genetic merit for selection candidates without phenotyping.

The National Crops Resources Research Institute (NaCRRI), in Uganda is among the first cassava breeding programs to implement genomic selection. The present study covers three main areas; First, we assessed genetic variation, level of inbreeding and trait correlations in genomic selection breeding cycles to understand the impact of accelerated breeding on genetic diversity, inbreeding and trait correlations in cassava. Secondly, we tested genomic prediction accuracies for agronomic and disease traits in light of genotype-by-environment (GxE) interactions, a common problem in breeding for a wide adaptation. In the third objective, we tested genomic prediction accuracies for CBSD-related traits across breeding program (predictions of CBSD resistance in W. African clones) as pre-emptive breeding strategy.

The key results and conclusions of the study are: (i) There was genetic progress made for most traits from genomic selection cycle zero (C0) to cycle one (C1). Selection based on GEBVs did not erode the original genetic diversity of lines bred under GS enabled breeding system. Based on these results, we do not expect genomic selection to cause rapid inbreeding, as breeding populations are moved from one cycle of genomic selection to the next. (ii) Inclusion of GxE information in genomic prediction showed moderate to high prediction accuracies for CBSD-related traits, dry matter content (DMC) and harvest index (HI) under different cross-validation prediction schemes. However, the predictive ability of GxE models were generally low for root and shoot weight per plot, except for a scenario of predicting unobserved environments. This result implies that selection can be made accurately for CBSD, DMC, HI based on GxE models, while additional phenotypic information is required on the genotypes to select for fresh root yield. (iii)Moderate predictive accuracies were recorded for CBSD in West African clones for foliar disease symptom expression, but low prediction accuracies were observed for root necrosis. Based these results, building a training set comprising West African clones is recommended to predict CBSD resistance in West Africa germplasm where the disease is yet non-existent.

The output of this study provides vital information to breeders to aid inter-regional prediction and reduce multi-locational evaluation costs, without comprising the genetic diversity across generations. The implementation of genomic assisted breeding will greatly contribute to improved cassava production to meet calories of people in the developing world.

Google Calendar iCal Outlook
Event Type

Special Event


Plant Breeding and Genetics


pbg, sips, sipsseminar


Alfred Ozimati

Speaker Affiliation

Plant Breeding and Genetics, Cornell University

Recent Activity