Multi‐trait genomic selection (MT‐GS) has the potential to improve predictive ability by maximizing the use of information across related genotypes and genetically correlated traits. In this study, we extended the use of sparse phenotyping method into the MT‐GS framework by split testing of entries to maximize borrowing of information across genotypes and predict missing phenotypes for targeted traits without additional phenotyping expenditure. Using 300 advanced breeding lines from North Dakota State University (NDSU) pulse breeding program and ∼200 USDA accessions that were evaluated for 10 nutritional traits, our results show that the proposed sparse phenotyping aided MT‐GS can further improve predictive ability by >12% across traits compared with univariate (UNI) genomic selection. The proposed strategy departed from the previous reports that weak genetic correlation is a limitation to the advantage of MT‐GS over UNI genomic selection, which was evident in the partially balanced phenotyping‐enabled MT‐GS. Our results point to heritability and genetic correlation between traits as possible metrics to optimize and further improve the estimation of model parameters, and ultimately, prediction performance. Overall, our study offers a new approach to optimize the prediction performance using the MT‐GS and further highlight strategy to maximize the efficiency of GS in a plant breeding program. The sparse‐testing‐aided MT‐GS proposed in this study can be further extended to multi‐environment, multi‐trait GS to improve prediction performance and further reduce the cost of phenotyping and time‐consuming data collection process.

Maintenance of genome integrity is critical for proper cell growth. This occurs through accurate DNA replication and repair of DNA lesions. A key factor involved in both DNA replication and the DNA damage response is the heterotrimeric single-stranded DNA (ssDNA) binding complex Replication Protein A (RPA). Although the RPA complex appears to be structurally conserved throughout eukaryotes, the primary amino acid sequence of each subunit can vary considerably. Examination of sequence differences along with the functional interchangeability of orthologous RPA subunits or regions could provide insight into important regions and their functions. This might also allow for study in simpler systems. We determined that substitution of yeast Replication Factor A (RFA) with human RPA does not support yeast cell viability. Exchange of a single yeast RFA subunit with the corresponding human RPA subunit does not function due to lack of inter-species subunit interactions. Substitution of yeast Rfa2 with domains/regions of human Rpa2 important for Rpa2 function (i.e., the N-terminus and the loop 3–4 region) supports viability in yeast cells, and hybrid proteins containing human Rpa2 N-terminal phospho-mutations result in similar DNA damage phenotypes to analogous yeast Rfa2 N-terminal phospho-mutants. Finally, the human Rpa2 N-terminus (NT) fused to yeast Rfa2 is phosphorylated in a manner similar to human Rpa2 in human cells, indicating that conserved kinases recognize the human domain in yeast. The implication is that budding yeast represents a potential model system for studying not only human Rpa2 N-terminal phosphorylation, but also phosphorylation of Rpa2 N-termini from other eukaryotic organisms.

The superiority of multi-trait genomic selection (MT-GS) over univariate genomic selection (UNI-GS) can be improved by redesigning the phenotyping strategy. In this study, we used about 300 advanced breeding lines from North Dakota State University (NDSU) pulse breeding program and about 200 USDA accessions evaluated for ten nutritional traits to assess the efficiency of sparse testing in MT-GS. Our results showed that sparse phenotyping using MT-GS consistently outperformed UNI-GS when compared to partially balanced phenotyping using MT-GS. This strategy can be further extended to multi-environment multi-trait GS to improve prediction performance and reduce the cost of phenotyping and time-consuming data collection process. Given that MT-GS relies on borrowing information from genetically correlated traits and relatives, consideration should be given to trait combinations in the training and prediction sets to improve model parameters estimate and ultimately prediction performance. Our results point to heritability and genetic correlation between traits as possible parameters to achieve this objective.