Genetic improvement is considered the major contributor to crop productivity. Advances in biotechnology, such as the availability of whole genome sequences, high throughput genotyping and phenotyping tools, as well as data management, and analytical services, enable breeders to better understand the genetic basis of agriculturally important traits in crops and predict the breeding values of individual plants or lines in a plant breeding program. Additionally, the decreasing cost of using molecular techniques enables breeders to screen large populations, thus increasing the efficiency of their application.
The various approaches for accelerated breeding include marker-assisted backcrossing (MABC), a quick and effective way of transferring a gene from a donor line to another line that is deficient in the trait of interest; markerassisted recurrent selection (MARS), which allows the accumulation of a relatively large number of favorable alleles, represented by quantitative trait loci, using selected markers that are significantly associated with target traits, and genomic selection (GS) which helps to predict the genetic values of breeding progenies using a statistical model based on markers distributed across the genome. IITA researchers are developing and deploying these cutting-edge genomic tools and techniques for innovative and accelerated breeding of crops such as maize, cassava, cowpea, and yam.
Fast-track breeding of stress-tolerant maize
DNA markers linked to key traits, such as tolerance to drought and Striga, have been identified and applied in our maize breeding to save time and reduce the costs associated with extensive field evaluation. Two cycles of genotypic selection have been completed in four IITA MARS populations drawn from different maturity groups, adapted to the low to medium altitude, and with various agronomically superior attributes. Lines derived from the various marker-based cycles of selection have been evaluated in multilocational field trials to estimate genetic gains. The primary focus was on the performance of lines under drought and no drought conditions as well as Striga infestation. In the two populations analyzed so far, MARS increased the frequencies of favorable alleles, suggesting the efficiency of genotypic selection.
Catalyzing genetic gain in cassava through genomic selection
Cassava breeding through phenotypic recurrent selection has achieved remarkable success, with a large number of disease resistant and high yielding improved varieties being currently deployed throughout sub-Saharan Africa. However, it takes between 4 and 6 years of field phenotyping to identify good parents for generating the next cycle of selections. The breeding cycle has now been reduced to 1-2 years through genomics-assisted breeding. IITA has teamed with Cornell University and the national programs of Nigeria and Uganda to embark on genomic selection-based breeding within the framework of the Next Generation Cassava Breeding Project (www.nextgencassava.org).
Since the onset of the project, three cycles of genomic selection and recombination have been undertaken. In each of these annual cycles, about 100 clones with good breeding values for key traits (fresh root yield, drymatter content, and resistance to cassava mosaic disease), undergo controlled crosses producing 5,000-10,000 seeds. To accelerate the breeding cycle, the seeds are germinated during the off-season and about 2,500 of these are selected based on parental breeding values for genotyping-by-sequencing (GBS). Each seedling is genotyped at more than 100,000 genomic positions to generate the breeding values for traits of interest. Superior progenies are then selected for the next cycle of controlled crosses in the main season. This shortened breeding cycle allows the breeding program to respond to changes in breeding targets and meet the demands of smallholder farmers. As a result, new cohorts of improved varieties have been channeled towards the product development pipeline.
Additionally, several other aspects of this pipeline have been strengthened. For example, trait measurements on individual plants and plots are captured using android apps running on tablets and smartphones, making the data instantly available for uploading to the breeding database. To store, analyze, and ensure open access, the Cassava Breeding Program is currently depositing all field-trial data on http://cassavabase.org. This database not only provides access to data but also hosts tools for breeders and other researchers that include genomic selection algorithms and analysis capacity, a cassava genome browser, cassava ontology tools, phenotyping tools, and social networking.
The promise of molecular techniques in yam
Among the major staple food crops, yam (Dioscorea spp.) is a challenge to breeders. The biology of the crop makes it less amenable to genetic improvement as it is a polyploid dioecious species with a significant period of tuber dormancy that prolongs the growth cycle. The yam collections maintained at IITA have been characterized using GBS. This has allowed varietal identification and description of genetic diversity, linkage mapping, and QTL analysis of target traits (anthracnose disease, sex-determination, and other agronomic traits) for accelerated breeding through marker-assisted selection. A simple and efficient Agrobacterium-mediated transformation system for D. rotundata has been established, opening up an avenue for further genetic studies. Through the AfricaYam project (www.africayam.org), genomics and marker-assisted breeding platforms are being established to fast-track the development of new varieties and training of NARS partners.
Novel approaches for advancing cowpea improvement
In cowpea, genomic tools are being developed to enhance progress in breeding improved varieties with attributes preferred by farmers and consumers. To this end, the genetic diversity of 365 lines, a subset representing the entire 15,000 accessions maintained at IITA, has been characterized by GBS. This has classified the accessions into five distinct groups. In addition, a set of about 215 recombinant inbred lines (RILs) has been genotyped and phenotyped for resistance to aphids and other desirable traits. A wild cowpea relative, which is resistant to aphids, is one of the two parents for the RILs. Trait-linked candidate markers are also being tested for their efficacy in facilitating genetic improved breeding line to two released varieties which lack resistance to the parasitic weed.
Data management and decision support tools
As depicted in the above examples, the low cost of sequencing allows molecular breeding approaches with large populations producing vast amounts of raw data. These data need to be processed promptly to extract the information needed by the breeders for selecting progenies with good breeding values. The data analysis pipeline, including sequence cleaning, sequence polymorphism search, comparative genomics, and decision-making, needs specific hardware and software to be able to cope with the immense data load and complex analysis. By establishing a bioinformatics platform with the correspondingly adapted infrastructure, IITA is now able to efficiently analyze the big data produced by diverse breeding programs. The capacity of the computing infrastructure allows the breeders to analyze thousands of genotypes simultaneously producing up to 100,000 data points. Genotypic data, in combination with phenotypic and other metadata, are organized and stored in effective open access data management systems such as Breeding Management System (BMS, www.integratedbreeding.net) and customized crop databases (CassavaBase, YamBase, MusaBase). These systems provide the downstream decision-making tools with a vast quantity of high quality data for a precise selection of improved offspring for subsequent breeding steps.