**2.2 GGE analysis**

In GGE biplot analysis, the first two principal components (PC1 and PC2), derived by subjecting the environment-centered yield bi-plot (**Figure 1**), to singular value decomposition (SVD). Principal component (PC1) was significant and location accounted for 17% of the total sums of squares. Genotype by environment variation was greater in the two seasons confirming that food productivity is threatened by environmental variables. The small yield variation due to location is relevant to cultivar evaluation. The yield obtained from across environments selected the following genotypes as they combine yield and stability and this should be considered during genotype selection. Genotypes with the above mean performance were; IESV 92207DL, IESV92024SH × ICSB 497-1, on the basis of the

ATC (Average Tester coordinate X-axis) or (Average Tester coordinate Y –axis, the stability axis). The research identified two mega environments for sorghum under rain-fed areas. This has several implications for future breeding and genotype evaluations of sorghum that is, warm-dry ecology (Kumi) and sub-humid environments; (Namutumba, Pallisa and Iganga)**.** The closer an environment is to this virtual environment (ATC axis); the better it is as a test environment [11]. Thus Pallisa and Namutumba are relatively favorable test environments and most representative and as well discriminative, suitable for multiple stress evaluation with a yield above 2000 kg/ha. Kumi was a most discriminative environment, probably due to the high level of striga from low fertile sandy soils which enhance water stress, hence strong genotype by environment interaction. The large part of the genotype × environment interaction was also indicated by a positive correlation between different yield components (**Figure 2**).

#### **2.3 Cluster dendrograms based on grain yield**

A dendogram with clusters was created from 12 elite lines based on their level of similarity in grain yield/ha (**Figure 3**). This is important for planning a crossing program; for example, using IESV24029SH × ICSB 497 and KAK-7780 could have a good combination of favorable alleles. They were also among the best four highly ranked genotypes from Additive Main effects and Multiplicative Interaction (AMMI) analysis (**Table 5**). Genotype KAK-7780 was selected from a landrace population that could possess both major and minor gene systems for stress protection which contributes to yield.

*DOI: http://dx.doi.org/10.5772/intechopen.112322 Accelerating Breeding for Drought Tolerance in Sorghum (*Sorghum bicolor*): An Integrated…*

#### **Figure 2.**

*Dendogram showing how sorghum founder lines can be clustered together in groups on the basis of important biochemical composition.*

#### **Figure 3.**

*Dendrogram with clusters based on level of similarity (%) in grain yield (Kgm/ha).*


*NPT 1 = NAROSORGH3, NPT 5 = IESV24029SH, NPT9 = IESV24029SH × ICSB479-1, NPT22 = SESO1. NPT14 = IESV92207DL. NPT15 = IESV92034DLSEL2, NPT21 = NAROSORGH 2, NPT23 = SESO3.*

#### **Table 5.**

*AMMI ranking of the best four selections.*

#### **2.4 Cluster dendogram based on plant height**

Using IESV92034DLSEL2, KAK-7780, NAROSORGH 2 and IESV92207DL will be very useful for improved yield components. These differences could have resulted from differences in biomass production and seed weights and hence they are the best ranking candidates for rain-fed areas (**Figure 4**). Since the genotype cannot restrict their heights significantly across environments, it implies resistance operates against major sorghum production challenges (drought, disease and pests) contributing to stability. More research needs to be conducted to unravel the underlying principles for plant-stress interaction with respect to plant height so that they can be incorporated into breeding.

#### **2.5 Correlation between traits**

Genetic variability for yield and seedling vigor components exits and they include; seed size, days to flowering (maturity), plant height, panicle length, panicle width, pest and disease resistance. They are a result of active physiological processes driven by active translocation and they influence dry matter accumulation. Plant height is strongly positively correlated to panicle length (PL) and moderately to days to flowering (DF) and hundred seed weight (HSWT). On the other hand, stem borer (SB) resistance is moderately positively correlated to stay green (SC), plant height (PHT) and days to flowering (**Figure 2**), where the correlation is low (R = 0.2) or negative, or both, little progress can be made. For example, the relationship between hundred seed weight and stay green could be due to competition for sink source relationship. So grain yield may not be true determinant for drought tolerance [12].

This could be because hundred seed weight is calculated after seed cleaning. For breeding purposes, it is therefore important to compare hundred seed weight (seed size) among genotypes with respect to the standard check/commercial varieties. The data should be interpreted based on physiological time of maturity, growth patterns, dry matter accumulation, partitioning of sink and genetic differences. Although many traits have been studied for their use in breeding for drought resistance, there is a general consensus among breeders that only a few of them can be recommended for use in practical breeding programs at this time. The study identified maturity (50% flowering), plant height, stay green [13].


#### **Figure 4.**

*Level of similarity (%) based on plant height.*

*DOI: http://dx.doi.org/10.5772/intechopen.112322 Accelerating Breeding for Drought Tolerance in Sorghum (*Sorghum bicolor*): An Integrated…*

#### **2.6 Sectional conclusion**

It is possible to select stable genotypes combining the number of yield components since genetic merit for these traits exist. For example, tall plants of a height of above 170 cm (centimeters) can be used to select improved grain yield components. Considering the strong weak negative correlation between turscicum leaf blight (TLB) infection with plant height and days to flowering, implies delayed flowering is associated with decreased diseased levels and is common among tall plants. The key traits are plant height, stay grain and days to 50% flowering which are most likely to improve the rate of genetic gains for drought tolerance. There is a need to exploit information gained from correlated traits and pedigree for the selection of parents with multiple traits [14, 15].
