Skip to main content
Premium Trial:

Request an Annual Quote

Cornell Researcher Spells Out Strategy for Next-Gen Cassava Breeding Project at PAG

SAN DIEGO (GenomeWeb News) – Members of the Next Generation Cassava Breeding project are gearing up to implement genomic selection in cassava breeding in sub-Saharan Africa — a strategy that's expected to help breeders of the staple crop develop new cassava varieties more quickly.

Late last year, a Cornell University-led team secured $25.2 million in funding from the Bill and Melinda Gates Foundation and the UK's Department of International Development for the five-year Next Generation Cassava Breeding effort, which also involves researchers from other centers in the US, Uganda, and Nigeria.

During a session on cassava genomics at the International Plant and Animal Genomes meeting, Cornell University's Martha Hamblin presented results from a pilot study using genomic selection in cassava. She also described plans for the main phase of the project.

Cassava is able to grow in marginal soils and low water conditions. But there is still a ways to go in maximizing the production potential of cassava in Africa and other parts of the world. And traditional breeding methods are time-consuming, Hamblin explained, since it takes several years to breed plants, propagate them over multiple generations, and select clones with the most favorable features.

To speed up that process, researchers are searching for genetic markers that can be used to predict cassava traits at the earlier seeding stage, significantly cutting the time needed for a breeding cycle.

Genomic selection "evaluates seedlings based on genotype alone, bypassing the need for expensive and time-consuming yield trials," Hamblin and her colleagues wrote in the abstract accompanying the PAG presentation, "with the goal of dramatically increasing the rate of genetic improvement of cassava."

For the pilot phase of the Next Generation Cassava Breeding project, the team took advantage of a cassava clone collection at the International Institute of Tropical Agriculture, or IITA, one of the centers collaborating on the project.

Using 600 phenotypically characterized cassava lines from this resource, researchers split the set into 480 training lines and 120 validation lines.

After doing genotyping-by-sequencing at more than 2,000 SNPs in each of the training set lines, the investigators trained their genomic selection model on this variant data, Hamblin explained.

From there, they tried to predict traits in the validation group based on genotype information, alone, giving them a chance to look at how well their marker-based predictions jibed with plants' actual traits.

Additional cross validation studies supported the notion that such markers often corresponded quite well with cassava traits, Hamblin said. Generally speaking, the predictive accuracy of these markers was higher for more heritable traits, she added, though that wasn't always the case.

Going forward, Hamblin noted that there are several viable avenues for increasing the accuracy of genomic selection in cassava — from doing more detailed and quality phenotyping of plants to optimizing models and considering more markers in the genome.

She also highlighted some of the work that's gone into designing a feasible genomic selection workflow that takes into account the capacity and seasonal constraints at participating breeding centers as well as the time needed for steps such as DNA extraction and sequencing. For instance, the IITA will fold the genomic selection research into an existing annual breeding program already in place there, while Uganda's National Crops Resources Research Initiative and the National Root Crops Research Institute in Nigeria will use two-year breeding cycles.

On the analysis side, study collaborators are also hammering out the most appropriate methods for calling SNPs, filtering data, and tackling other informatics stages of the project.

During another presentation at the same session, researcher Lukas Mueller of the Boyce Thompson Institute for Plant Research touched on the development of CassavaBase, a clearinghouse for phenotypic information, genotyping data, geolocation and other data on cassava lines, and other information collection during the Next Generation Cassava Breeding project.