Multi-omics assists genomic prediction of maize yield with machine learning approaches.

With the improvement of high-throughput technologies in recent years, large multi-dimensional plant omics data have been produced, and big-data-driven yield prediction research has received increasing attention. Machine learning offers promising computational and analytical solutions to interpret the biological meaning of large amounts of data in crops. In this study, we utilized multi-omics datasets from 156 maize recombinant inbred lines, containing 2496 single nucleotide polymorphisms (SNPs), 46 image traits (i-traits) from 16 developmental stages obtained through an automatic phenotyping platform, and 133 primary metabolites. Based on benchmark tests with different types of prediction models, some machine learning methods, such as Partial Least Squares (PLS), Random Forest (RF), and Gaussian process with Radial basis function kernel (GaussprRadial), achieved better prediction for maize yield, albeit slight difference for method preferences among i-traits, genomic, and metabolic data. We found that better yield prediction may be caused by various capabilities in ranking and filtering data features, which is found to be linked with biological meaning such as photosynthesis-related or kernel development-related regulations. Finally, by integrating multiple omics data with the RF machine learning approach, we can further improve the prediction accuracy of grain yield from 0.32 to 0.43. Our research provides new ideas for the application of plant omics data and artificial intelligence approaches to facilitate crop genetic improvements. Supplementary Information: The online version contains supplementary material available at 10.1007/s11032-024-01454-z.

Combining association with linkage mapping to dissect the phenolamides metabolism of the maize kernel.

Phenolamides are important secondary metabolites in plant species. They play important roles in plant defense responses against pathogens and insect herbivores, protection against UV irradiation and floral induction and development. However, the accumulation and variation in phenolamides content in diverse maize lines and the genes responsible for their biosynthesis remain largely unknown. Here, we combined genetic mapping, protein regulatory network and bioinformatics analysis to further enhance the understanding of maize phenolamides biosynthesis. Sixteen phenolamides were identified in multiple populations, and they were all significantly correlated with one or several of 19 phenotypic traits. By linkage mapping, 58, 58, 39 and 67 QTLs, with an average of 3.9, 3.6, 3.6 and 4.2 QTLs for each trait were mapped in BBE1, BBE2, ZYE1 and ZYE2, explaining 9.47%, 10.78%, 9.51% and 11.40% phenotypic variation for each QTL on average, respectively. By GWAS, 39 and 36 significant loci were detected in two different environments, 3.3 and 2.8 loci for each trait, explaining 10.00% and 9.97% phenotypic variation for each locus on average, respectively. Totally, 58 unique candidate genes were identified, 31% of them encoding enzymes involved in amine and derivative metabolic processes. Gene Ontology term analysis of the 358 protein-protein interrelated genes revealed significant enrichment in terms relating to cellular nitrogen metabolism, amine metabolism. GRMZM2G066142, GRMZM2G066049, GRMZM2G165390 and GRMZM2G159587 were further validated involvement in phenolamides biosynthesis. Our results provide insights into the genetic basis of phenolamides biosynthesis in maize kernels, understanding phenolamides biosynthesis and its nutritional content and ability to withstand biotic and abiotic stress.