Los Baños, Philippines
11 September 2024
In the fast-paced world of crop breeding, timing is everything. Breeders face immense time pressure, particularly during the critical crossing period, where multiple tasks must be handled simultaneously from controlled pollination, environmental management, seed development, harvesting, to analysis. In many cases, with only three to four days to make crucial decisions before moving on to the next cross, breeders often find themselves bogged down by manual calculations. This leaves them with little time to thoroughly evaluate all the factors that could impact their success.
The CGIAR Genetic Innovation teams have developed an advanced breeding analytics pipeline to support crop breeders and their own teams. The tool automates complex calculations, freeing up valuable time and allowing breeders to focus on making informed decisions that drive better outcomes.
Three ways data analytics simplifies the work of breeders
- Understanding the impact of environment on genotypes
In nature, the genotype of plants and animals is influenced by various environmental interactions and conditions, which could significantly affect breeding efforts by decreasing the genetic signal. Through agronomic management, breeders can manage these factors by adjusting shade levels, improving drainage, increasing water supply, and so on. In addition, experimental design is carried out to dissect the genetic signal through post-harvest statistical analysis when the genetic effects are not easily discernible by sight alone.
One effective way to reduce environmental noise is through analytics, which assesses the location, year, block, incomplete block effects among others and eliminates them. As agriculture becomes increasingly sophisticated, the volume of data collected has grown substantially. With breeding analytics and the power of big data, we can better explain the sources of variation that cause plants to behave in certain ways.
For instance, modern breeding involves collecting multiple data points over time, together with environmental information such as different temperatures and precipitation levels. But how do these variables impact plant behavior?
Leveraging analytics, can promote a deeper understanding of plant phenotypes and the potential of parent plants, unveiling the true value of a product. Removing noise enables better decision making in a complex world. Ultimately, this leads to the selection of better parents and superior products.
- Identifying genes behind traits
Another key application of analytics in breeding is identifying the genes responsible for specific traits. For plants, this could mean better yield, reduced fertilizer needs, etc. Analytics pinpoint the regions of the genome responsible for particular traits. The variation of these genes, or alleles, is also crucial and can be identified through breeding analytics such as genome-wide association models. This process begins with gene discovery, followed by trait introgression, where these genes are integrated into breeding programs. Traditionally, introgressing genes through multiple backcrossing is time-consuming. Analytics can cut this process by half.
- Predictive analytics in breeding
The third critical use of analytics is in making predictions. Predictive analytics helps us understand what the phenotype of an individual plant is likely to be in a given environmental condition, aiding breeders in deciding whether or not to plant it in the first place. These predictions, based on probabilities, provide direction and are only possible with the help of analytics, which are based on quantitative genetic models developed in the last 100 years. Marker effects and other models can be coded into computer programs, allowing scientists to use these tools. Furthermore, applications can be developed without coding skills, enabling scientists without programming expertise to design and run these models.
This is where CGIAR’s breeding analytics pipeline plays a crucial role. It democratizes access to advanced tools, empowering breeders and scientists alike to make better, data-driven decisions that are transforming the future of crop breeding.
CGIAR breeding data analytics pipeline is already available
The CGIAR Breeding Analytics Pipeline is an intuitive interface designed for breeders and partners, particularly those who lack coding skills but need to calculate breeding values and develop predictive models. This platform provides access to algorithms that harness evolutionary forces in genetics — selection, mutation, gene flow and drift. These forces drive the evolution of species worldwide since the dawn of time and are fundamental to any breeding process.
In breeding, breeders utilize these forces to drive desired changes. So, the idea is to virtually implement selection to achieve specific effects on the phenotype, predicting outcomes and making fieldwork more efficient. The CGIAR Breeding Analytics Pipeline enables breeders to do just that—eliminating environmental noise from phenotypes to rescue genetic effects, facilitating gene mapping, and enhancing the understanding of genetic variation.
The tool automates many routine workflows, making breeding more efficient, saving time, and supporting better decision-making. In many cases, tasks that once took hours or days to complete manually can now be done in just five to 10 minutes. While there is a learning curve, most users can run analyses after just a few hours of training, saving weeks of work in the long run.
To ensure it continues to meet breeders' needs, the tool's development is overseen by the Breeding Resources Initiative, with feedback from the Accelerated Breeding team. And soon, the use of the Breeding Analytics Pipeline will become a requirement for CGIAR breeding programs to reach standardization of methods.
Members of the CGIAR-NARES network can access the Breeding Analytics Pipeline through the Breeding Resources Service Request Portal, with free access to cloud computing included for these users. Those outside the network can still access the demo server, which offers the same capabilities except for cloud storage.
Training support for CGIAR and NARES partners is available through the Service Request Portal or our biometrics unit, where users can request a demo and learn how to use the tool. Training is sponsored by Breeding Resources and will soon be funded by CGIAR. For further guidance, a video series is also available.
Resources:
The development of the BA Pipeline is funded by Crops to End Hunger, a broad multi-funder initiative substantially supported by the Federal Ministry for Economic Cooperation and Development (BMZ), and channelled through the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ). We extend our heartfelt appreciation to all the generous supporters of this initiative, with special recognition to GIZ. We also express our gratitude to the CGIAR research funders for their invaluable contributions.
Topics
Species
