A desirable tester that elicits greater genetic difference in Striga resistance among test crosses in a breeding program has not been reported. Therefore, this study was conducted to characterize 30 Striga resistant yellow endosperm maize inbred lines and three testers with varying resistance levels to Striga using DArTseq SNP markers and agronomic traits to identify a suitable tester for resistance hybrid breeding. Marker-based and agronomic trait-based genetic distances were estimated for yellow endosperm maize inbred lines and testers with varying resistance levels to Striga. The Marker-based cluster analysis separated the Striga resistant lines and testers into two distinct groups. Although the susceptible tester (T3) was the most distantly related to the 30 Striga resistant inbred lines, it exhibited a narrower range in genetic distance estimates and poor agronomic performance under Striga infestation in crosses with the resistant lines. In contrast, the resistant tester (T2) showed a broader range in genetic distance estimates in pairs with the 30 resistant lines. Also, it formed many high yielding hybrids with desirable traits under parasite pressure. Furthermore, the most significant positive association between agronomic trait-based and marker-based distance estimates (r= 0.389, P= 0.01) was observed when T2 has paired with the Striga resistant maize inbred lines. It thus appears that T2 may be used as a suitable tester to determine the breeding value of lines in hybrid maize resistance breeding programs. T2 was the most suitable tester, with a tolerant tester (T1) as an alternative tester to characterize the combining ability of Striga resistant maize inbred lines. This result can also encourage other breeders to investigate testers relative discriminating ability with varying levels of resistance in hybrid breeding for resistance to diseases, pests, and other parasitic plants.

Genetic Materials

The germplasm used in this study was developed at IITA, Ibadan, Nigeria. These consisted of thirty yellow endosperm Striga resistant maize inbred lines and three testers with varying levels of resistance to Striga. Pedigree of the inbred lines and testers were described by Zebire et al. [37].  Lines were designated as L1-L30, and testers presented as T1, T2 and T3 for the tolerant, resistant and susceptible tester, respectively. The 30 inbred lines each were crossed to the three testers to generate 90 test crosses, which were evaluated along with two checks having known tolerant and susceptible reactions to the parasite under Striga infested and non-infested conditions at Abuja and Mokwa in Nigeria for two years.  All procedures used to precondition Striga seed, Striga infestation and data collection for field trail were described by Zebire et al. [37].

DNA Extraction and Genotyping

Young growing leaves at the 3 to 4 leaf stage were collected from maize plants grown on the field for DNA extraction. Leaf samples were collected from 4 to 15 leaves of each line and tester. The samples were stored in a deep freezer at -80°c. Each sample was dried in a Labconco Freezone 2.5L system lyophilizer (Marshall Scientific, USA) before genomic DNA extraction. The extraction of the genomic DNA (gDNA) was carried out according to the DArT protocol (www.diversityarrays.com/files/ DArT_DNA_isolation.pdf). The quality and quantity of the DNA were checked by gel electrophoresis using 0.8% agarose gel and NANODROP® spectrophotometer (Thermo Fisher Scientific Inc., Denver, CO, USA). The samples were sent to the Diversity Array technology company [38]. All protocols, including Library construction, sequencing, and SNP calling, were performed at the Diversity Arrays facility, Canberra, Australia. ApeK1 restriction enzyme was used to digest the gDNA, and genotyping-by-sequencing (GBS) libraries were constructed in 96-plex for the samples and sequenced on Illumina HiSeq2000. Raw flow cell output was processed to genotype calls using the trait analysis by association, evolution and linkage (TASSEL)-GBS pipeline. The information of reads and tags found in each sequencing result was aligned to the Zea mays L. genome reference, version AGPV4 (B73 RefGen v4 assembly).

Data Analysis

Marker-based genetic diversity analysis

A total of 27,874 SNP markers generated from the present maize panel was received from the DArTseq platform. Quality control was performed to retain only bi-allelic sites, and the SNPs were further filtered using the TASSEL software [39] to maintain only polymorphic SNPs with a maximum of 10% missing values and a minimum and maximum allele frequency of 0.05 and 0.95, respectively. The final filtered data comprised 6081 SNP markers spanning the ten chromosomes of maize matched the quality criteria and were used for further analysis. Markers were used to calculate polymorphic information content (PIC), minor allele frequency, the number of alleles, gene diversity and heterozygosity using powerMarker software 3.25 version [40]. Genetic distance was estimated between a pair of inbred lines from 6081 markers using Roger's genetic distance (GD) in powerMaker version 3.25 [40]. A relative kinship matrix was calculated between pairs of inbred lines and testers from 6081 SNPs to understand the extent of relatedness using TASSELv.5.2.48 [39]. 

Cluster analysis was performed for the inbred lines and testers based on the genetic distance matrix with the unweighted pair group method of the arithmetic mean clustering algorithm (UPGMA dendrogram) in the PowerMarker version 3.25 [40] and viewed using MEGA, version 6.0 [41]. Population structure analysis was estimated from 6081 SNPs, based on a physical distance of 11 kb between adjacent markers. An admixture model-based clustering method was also used to infer the 33 inbred lines and testers population structure using STRUCTURE, version 2.3.4 [42]. Individuals with a probability of membership ≥ 60% were assigned to the same group, while those with < 60% probability memberships in any single group that did not show an ancestry proportion higher than this value was assigned to a "mixed" group [43 and 44]. The most probable value of K was estimated using the ad hoc statistic ΔK [45], depending on the rate of change in the log probability of data between successive K values. Also, principal coordinates analyses (PCoA) was performed based on the 6081 SNPs to distinguish among groups formed by the Striga resistant inbred lines and testers using GenAlEx 6.5 software [46].

Agronomic trait-based diversity assessment

Agronomic trait-based diversity analysis was carried out using eleven traits, namely grain yield, days to anthesis and silking, anthesis-silking interval, plant height, Striga damage ratings and Striga emergence counts at 8 and 10 weeks after planting (WAP), ear aspect and ears per plant from the 30 yellow Striga maize inbred lines and three testers. Means of the selected traits was first standardized in SAS version 9.4 [47]. Correlation among the different traits was analyzed using statistical analysis software (SAS). The principal component analysis was computed in SAS using the correlation matrix of trait means-centred averaged over envi­ronments. Measurements of genetic dissimilarity were then estimated from standardized data using the Euclidian distance matrix, after which has been subjected to cluster the lines and tester using Ward's clustering method [48]. The associations between the agronomic trait-based Euclidean distance matrix and marker-based genetic distance were calculated using GenAlEx 6.5 software [46], and the mantel test was used to determine the significant association between these data sets.