Genomic Prediction of Morphometric and Colorimetric Traits in Solanaceous Fruits: Optimizing Plant Breeding Through Machine Learning and Genomic Selection

Introduction

In the realm of horticulture, plant breeding holds immense significance, particularly for the Solanaceae family, which includes widely cultivated crops such as tomato, pepper, and eggplant. Plant breeders strive to select high-performance lines exhibiting desirable traits, a process traditionally carried out through conventional breeding methods. While effective, these methods face limitations in terms of precision and labor intensity, often hindering breeding efficiency.

Fortunately, advancements in machine learning and genomic prediction, particularly Genomic Selection (GS), have emerged as game-changers in plant breeding. GS harnesses the power of genetic markers, such as Single Nucleotide Polymorphisms (SNPs), to estimate the breeding values of unseen lines, offering unprecedented opportunities to revolutionize breeding efficiency.

Current Research Focus

To fully exploit the potential of GS, researchers are actively pursuing two primary objectives:

1. Overcoming the Limitations of Marker-Assisted Selection (MAS): MAS, a traditional marker-based selection method, has certain limitations. Researchers are integrating various GS models, encompassing regression-based, classification-based, and deep learning models, to enhance the predictability of diverse traits. This integration aims to surpass the capabilities of MAS, enabling more accurate and efficient selection.

2. Refining GS Models for Minor Gene Effects: Minor gene effects, often overlooked in traditional breeding, can significantly contribute to complex traits. Researchers are refining GS models to capture these subtle effects, thereby streamlining breeding processes and unlocking the full potential of genetic variation.

Study: “Genomic Prediction of Morphometric and Colorimetric Traits in Solanaceous Fruits”

To further explore the potential of GS in Solanaceae fruits, a groundbreaking study titled “Genomic Prediction of Morphometric and Colorimetric Traits in Solanaceous Fruits” was published in March 2022 in Horticulture Research. This study delved into the performance of GS models for morphometric (shape-related) and colorimetric (color-related) traits in tomatoes and peppers.

Key Findings

The study yielded several notable findings, shedding light on the intricacies of GS in Solanaceae fruits:

1. The Power of TA Traits and Marker SNPs: The combination of Tomato Analyser (TA) traits and marker SNPs proved instrumental in predicting morphological and color-related traits in Solanaceae fruits. TA, a specialized tool employing longitudinal and transverse cut fruit images, provided valuable phenotypic data, enhancing the accuracy of GS models.

2. Pepper’s Superior Predictability: Across all TA traits, pepper exhibited higher predictability than tomato. This observation suggests that pepper’s genetic architecture might be more conducive to GS, potentially due to its higher genetic diversity or simpler genetic control of certain traits.

3. Predictability Varies Across Traits: The study revealed that basic traits, such as fruit size, were more easily predicted from genomic information compared to complex traits. This finding highlights the influence of trait complexity on GS accuracy.

4. Multi-Trait GS Models: For specific chromaticity traits in peppers, multi-trait GS models demonstrated a slight edge over single-trait models. This finding suggests that considering multiple traits simultaneously can improve prediction accuracy for certain characteristics.

5. Impact of Genetic Diversity: Testing on an independent population of wild tomatoes revealed lower predictability for all TA traits. This disparity is likely attributed to the genetic distance between the training (cultivated) and testing (wild) populations, emphasizing the importance of genetic diversity in GS.

6. Enhancing Predictability with Diverse Germplasm: Remarkably, adding just a few wild accessions dramatically improved the predictability, underscoring the significance of incorporating diverse germplasm into GS models. This finding reinforces the value of genetic diversity in breeding programs.

7. Comparison of GS Models with Traditional Phenotyping: For traits scored by traditional descriptors (CDs), a classification-based GS model was compared with TA phenotyping. The predictability of CD traits was generally lower than related TA traits, indicating the superiority of TA phenotyping in predicting fruit size and yield-related traits.

Conclusion

The study titled “Genomic Prediction of Morphometric and Colorimetric Traits in Solanaceous Fruits” provides valuable insights into the optimization of GS models for improved plant breeding outcomes. The varied predictabilities observed across different GS models and traits underscore the complexity of GS in Solanaceae fruits.

Moreover, the impact of genetic diversity on model performance and the superiority of TA over conventional phenotyping methods highlight the need for continuous refinement of GS approaches. These findings pave the way for further research and development, ultimately leading to more efficient and precise plant breeding strategies.

References

1. Hao Tong1,2,3, Amol N. Nankar1, Jintao Liu2, Velichka Todorova4, Daniela Ganeva4, Stanislava Grozeva4, Ivanka Tringovska4, Gancho Pasev4, Vesela Radeva-Ivanova4, Tsanko Gechev1, Dimitrina Kostova1,4 and Zoran Nikoloski1,2,3,*

Affiliations

1. Center of Plant Systems Biology and Biotechnology, Plovdiv, 4000, Bulgaria
2. Systems Biology and Mathematical Modeling, Max Planck Institute of Molecular Plant Physiology, Potsdam, 14476, Germany
3. Bioinformatics, Institute of Biochemistry and Biology, University of Potsdam, Potsdam, 14476, Germany
4. Maritsa Vegetable Crops Research Institute, Plovdiv, 4003, Bulgaria.

About Zoran Nikoloski

1. Zoran Nikoloski, a renowned Professor at the University of Potsdam and group leader at the Max Planck Institute of Molecular Plant Physiology, Potsdam-Golm, has made significant contributions to the field of computational systems biology.

2. Nikoloski’s early work focused on developing computational approaches for integrating heterogeneous data from high-throughput molecular profiling technologies.

3. Nikoloski possesses extensive expertise in computational systems biology, aiming to reconstruct networks from molecular profiles and understand how network structure determines the abundance of network components.

4. Nikoloski’s primary research areas include:

– Data-driven qualitative and quantitative modeling of genome-scale metabolic and gene-regulatory networks.
– Analysis of evolutionary and optimization processes in biological networks.
– Characterization of system’s functions emerging from molecular interactions.