Scalable Sparse Testing Genomic Selection Strategy for Early Yield Testing Stage.

To enable a scalable sparse testing genomic selection (GS) strategy at preliminary yield trials in the CIMMYT maize breeding program, optimal approaches to incorporate genotype by environment interaction (GEI) in genomic prediction models are explored. Two cross-validation schemes were evaluated: CV1, predicting the genetic merit of new bi-parental populations that have been evaluated in some environments and not others, and CV2, predicting the genetic merit of half of a bi-parental population that has been phenotyped in some environments and not others using the coefficient of determination (CDmean) to determine optimized subsets of a full-sib family to be evaluated in each environment. We report similar prediction accuracies in CV1 and CV2, however, CV2 has an intuitive appeal in that all bi-parental populations have representation across environments, allowing efficient use of information across environments. It is also ideal for building robust historical data because all individuals of a full-sib family have phenotypic data, albeit in different environments. Results show that grouping of environments according to similar growing/management conditions improved prediction accuracy and reduced computational requirements, providing a scalable, parsimonious approach to multi-environmental trials and GS in early testing stages. We further demonstrate that complementing the full-sib calibration set with optimized historical data results in improved prediction accuracy for the cross-validation schemes.

Maximizing efficiency of genomic selection in CIMMYT's tropical maize breeding program.

KEY MESSAGE: Historical data from breeding programs can be efficiently used to improve genomic selection accuracy, especially when the training set is optimized to subset individuals most informative of the target testing set. The current strategy for large-scale implementation of genomic selection (GS) at the International Maize and Wheat Improvement Center (CIMMYT) global maize breeding program has been to train models using information from full-sibs in a "test-half-predict-half approach." Although effective, this approach has limitations, as it requires large full-sib populations and limits the ability to shorten variety testing and breeding cycle times. The primary objective of this study was to identify optimal experimental and training set designs to maximize prediction accuracy of GS in CIMMYT's maize breeding programs. Training set (TS) design strategies were evaluated to determine the most efficient use of phenotypic data collected on relatives for genomic prediction (GP) using datasets containing 849 (DS1) and 1389 (DS2) DH-lines evaluated as testcrosses in 2017 and 2018, respectively. Our results show there is merit in the use of multiple bi-parental populations as TS when selected using algorithms to maximize relatedness between the training and prediction sets. In a breeding program where relevant past breeding information is not readily available, the phenotyping expenditure can be spread across connected bi-parental populations by phenotyping only a small number of lines from each population. This significantly improves prediction accuracy compared to within-population prediction, especially when the TS for within full-sib prediction is small. Finally, we demonstrate that prediction accuracy in either sparse testing or "test-half-predict-half" can further be improved by optimizing which lines are planted for phenotyping and which lines are to be only genotyped for advancement based on GP.

Stacking Tolerance to Drought and Resistance to a Parasitic Weed in Tropical Hybrid Maize for Enhancing Resilience to Stress Combinations.

Maize is a food security crop cultivated in the African savannas that are vulnerable to the occurrence of drought stress and Striga hermonthica infestation. The co-occurrence of these stresses can severely damage crop growth and productivity of maize. Until recently, maize breeding in International Institute of Tropical Agriculture (IITA) has focused on the development of either drought tolerant or S. hermonthica resistant germplasm using independent screening protocols. The present study was therefore conducted to examine the extent to which maize hybrids simultaneously expressing resistance to S. hermonthica and tolerance to drought (DTSTR) could be developed through sequential selection of parental lines using the two screening protocols. Regional trials involving 77 DTSTR and 22 commercial benchmark hybrids (STR and non-DTSTR) were then conducted under Striga-infested and non-infested conditions, managed drought stress and fully irrigated conditions as well as in multiple rainfed environments for 5 years. The observed yield reductions of 61% under managed drought stress and 23% under Striga-infestation created desirable stress levels leading to the detection of significant differences in grain yield among hybrids at individual stress and non-stress conditions. On average, the DTSTR hybrids out-yielded the STR and non-DTSTR commercial hybrids by 13-19% under managed drought stress and fully irrigated conditions and by -4 to 70% under Striga-infested and non-infested conditions. Among the DTSTR hybrids included in the regional trials, 33 were high yielders with better adaptability across environments under all stressful and non-stressful testing conditions. Twenty-four of the 33 DTSTR hybrids also yielded well across diverse rainfed environments. The genetic correlations of grain yield under managed drought stress with yield under Striga-infestation and multiple rainfed environments were 0.51 and 0.57, respectively. Also, a genetic correlation between yields under Striga-infestation with that recorded in multiple rainfed environments was 0.58. These results suggest that the sequential selection scheme offers an opportunity to accumulate desirable stress-related traits in parents contributing to superior agronomic performance in hybrids across stressful and diverse rainfed field environments that are commonly encountered in the tropical savannas of Africa.

Genetic Gains in Yield and Yield Related Traits under Drought Stress and Favorable Environments in a Maize Population Improved Using Marker Assisted Recurrent Selection.

The objective of marker assisted recurrent selection (MARS) is to increase the frequency of favorable marker alleles in a population before inbred line extraction. This approach was used to improve drought tolerance and grain yield (GY) in a biparental cross of two elite drought tolerant lines. The testcrosses of randomly selected 50 S1 lines from each of the three selection cycles (C0, C1, C2) of the MARS population, parental testcrosses and the cross between the two parents (F1) were evaluated under drought stress (DS) and well watered (WW) well as under rainfed conditions to determine genetic gains in GY and other agronomic traits. Also, the S1 lines derived from each selection types were genotyped with single nucleotide polymorphism (SNP) markers. Testcrosses derived from C2 produced significantly higher grain field under DS than those derived from C0 with a relative genetic gain of 7% per cycle. Also, the testcrosses of S1 lines from C2 showed an average genetic gain of 1% per cycle under WW condition and 3% per cycle under rainfed condition. Molecular analysis revealed that the frequency of favorable marker alleles increased from 0.510 at C0 to 0.515 at C2, while the effective number of alleles (Ne) per locus decreased from C0 (1.93) to C2 (1.87). Our results underscore the effectiveness of MARS for improvement of GY under DS condition.