Combined Analysis Of Variance
Combined analysis of variance (ANOVA) showed significant (P ≤ 0.001) variations for genotype, environment, and GEIs for all traits studied (Table 3). The combined mean grain yield, other phenotypic and disease traits of twenty-two sorghum genotypes were presented in Table 4. The grand mean values were 110 for DTF, 168 for DTM, 265.64 cm for PHT, 24.04 cm for PL, 6.13 cm for PW, 53.81 g for PWT, 23.14 g for TGW, and 2272.90 kg ha− 1 for GY. The combined mean grain yield ranged from 1,017.22 -3,306.59 kg ha− 1 (Table 4); indicating a wide range of genetic variability among tested sorghum genotypes. Statistically, there was no significant difference among genotypes for resistance to anthracnose, turcicum leaf blight, and grain mold diseases. However, some genotypes, including ETSL100808, ETSL100152, ETSL101524, ETSL100445, IS38326, ETSL101858, ETSL100474, and Merera showed resistant reactions to anthracnose, while others, such as ETSL100091 and IS38412, were susceptible to anthracnose. All genotypes exhibited resistance to turcicum leaf blight and grain mold (Table 4). Genotypes that exhibited resistant reactions to diseases; ETSL100808, Merera, and ETSL100474 also gave high grain yields.
Table 3
Combined analysis of variance for grain yield and related traits of sorghum genotypes across eight environments
Source | | DTF | | DTM | | PHT | | PL | |
| DF | SS | MS | SS | MS | SS | MS | SS | MS |
Env | 7 | 9885.33 | 1412.19*** | 43229.347 | 6175.62*** | 1559616.7 | 222802.39*** | 1449.56 | 207.08*** |
Gen | 21 | 53230.91 | 2534.81*** | 18641.148 | 887.67*** | 364017.8 | 17334.18*** | 6693.87 | 318.76*** |
Rep/Env | 16 | 2863.22 | 179.00*** | 962.68683 | 60.17ns | 86367.33 | 5397.96** | 361.55 | 22.60* |
Env x Gen | 147 | 29584.58 | 201.26*** | 14424.584 | 98.13*** | 252133.26 | 1715.19*** | 3202.83 | 21.79*** |
Residuals | 335 | 23395.95 | 69.84 | 20480.43 | 61.14 | 359301.21 | 1072.541 | 3812.59 | 11.38 |
Mean | | 110.02 | 168.17 | 265.64 | 24.04 |
| | PW | PWT | TGW | GY |
| DF | SS | MS | SS | MS | SS | MS | SS | MS |
Env | 7 | 206.51 | 29.50*** | 86718.73 | 12388.39*** | 1073.53 | 153.36*** | 120382615 | 17197516.00*** |
Gen | 21 | 205.6 | 9.79*** | 99803.56 | 4752.55*** | 3257.83 | 155.13*** | 2792470 | 7332972.00** |
Rep/Env | 16 | 94.71 | 5.92*** | 32437.41 | 2027.34*** | 690.88 | 43.18*** | 153992406 | 174529.00* |
Env x Gen | 147 | 257.69 | 1.75*** | 144144.87 | 980.58*** | 3066.61 | 20.86*** | 150623166 | 1024647.00*** |
Residuals | 335 | 329.78 | 0.98 | 169057.05 | 504.65 | 3719.79 | 11.1 | 30882447 | 91912 |
Mean | | 6.13 | 53.81 | 23.14 | 2272.90 |
***= Significant at 0.001 significance level, **= Significant at 0.01significance level, *=Significant at 0.05 significance level, ns = non-significant |
DF = degree of freedom, DTF = days to flowering, DTM = days to maturity, PHT = plant height (cm), PL = panicle length (cm), PW = panicle width (cm), PWT = panicle weight per plant (g), TGW = thousand grain weight (g), GY = grain yield (kg ha− 1), Gen = genotype, Env = environment, Rep = replication, SS = sum squares, MS = mean squares |
Table 4
Mean performance of twenty-two sorghum genotypes across eight environments for grain yield, agronomic, and disease traits
Gen | GY | DTF | DTM | PHT | PL | PW | PWT | TGW | ANTHa | TLBb | PGMRb |
ETSL101524 | 2466.94 | 124.67 | 176.63 | 292.31 | 26.50 | 6.87 | 64.39 | 22.85 | 2.3 | 1.9 | 1.3 |
ETSL100445 | 3005.11 | 131.58 | 182.21 | 291.61 | 21.85 | 7.96 | 95.22 | 22.81 | 2.3 | 1.9 | 1.1 |
ETSL100091 | 1971.60 | 124.92 | 180.73 | 256.71 | 22.12 | 5.73 | 54.64 | 24.76 | 3.1 | 2.3 | 1.3 |
ETSL101312 | 2188.22 | 98.21 | 164.83 | 285.13 | 24.65 | 6.44 | 58.21 | 22.37 | 2.8 | 2.2 | 1.1 |
IS25531 | 1831.16 | 116.38 | 172.92 | 252.33 | 24.56 | 5.67 | 46.13 | 24.23 | 2.5 | 2.2 | 1.6 |
ETSL101732 | 2098.36 | 104.46 | 163.50 | 240.34 | 19.99 | 5.72 | 42.93 | 20.92 | 2.9 | 2.3 | 1.4 |
ETSL100152 | 2928.16 | 102.88 | 162.17 | 254.42 | 29.02 | 5.65 | 54.92 | 23.53 | 2.2 | 2.0 | 1.6 |
ETSL100561 | 2188.07 | 94.00 | 161.75 | 274.22 | 25.19 | 6.50 | 45.21 | 23.83 | 2.7 | 2.2 | 1.2 |
ETSL100474 | 2910.82 | 115.58 | 170.67 | 291.71 | 29.22 | 6.22 | 62.02 | 25.20 | 2.4 | 2.0 | 1.1 |
ETSL101128 | 1579.58 | 102.67 | 166.63 | 285.30 | 24.80 | 5.82 | 49.28 | 23.35 | 2.5 | 2.2 | 1.5 |
Bonsa | 2301.19 | 100.88 | 167.42 | 185.03 | 27.88 | 5.89 | 56.50 | 21.12 | 2.5 | 2.2 | 1.2 |
ETSL100808 | 3306.59 | 114.33 | 164.96 | 311.44 | 26.40 | 6.02 | 64.18 | 21.21 | 2.2 | 2.1 | 1.0 |
IS38326 | 2259.63 | 117.00 | 169.75 | 271.30 | 24.32 | 6.91 | 57.39 | 24.42 | 2.3 | 2.1 | 1.6 |
Merera | 3044.62 | 107.38 | 165.71 | 269.52 | 16.66 | 7.07 | 66.13 | 20.26 | 2.4 | 2.0 | 1.2 |
IS38331 | 2406.40 | 114.13 | 171.29 | 270.54 | 25.73 | 5.75 | 56.61 | 30.15 | 2.3 | 2.1 | 1.2 |
ETSL101858 | 1975.63 | 92.54 | 164.79 | 269.63 | 26.19 | 5.91 | 36.83 | 25.06 | 2.6 | 2.3 | 1.3 |
ETSL100072 | 2410.86 | 119.50 | 175.21 | 284.94 | 28.82 | 6.23 | 62.17 | 27.40 | 2.5 | 2.0 | 1.1 |
ETSL100771 | 1855.40 | 110.02 | 162.85 | 259.15 | 21.12 | 5.74 | 43.74 | 21.38 | 2.9 | 2.2 | 1.3 |
ETSL101173 | 1572.89 | 111.88 | 164.65 | 245.32 | 18.74 | 5.57 | 37.66 | 22.98 | 2.8 | 2.5 | 1.4 |
IS38412 | 1017.22 | 98.54 | 160.50 | 224.81 | 17.51 | 4.97 | 22.98 | 22.43 | 3.2 | 2.8 | 1.7 |
ETSL101712 | 2065.34 | 110.50 | 164.83 | 274.27 | 26.10 | 5.98 | 44.04 | 19.59 | 2.9 | 2.1 | 1.1 |
ETSL100587 | 2619.95 | 107.75 | 165.38 | 252.72 | 21.49 | 6.20 | 60.27 | 19.09 | 2.8 | 2.1 | 1.3 |
GM | 2272.90 | 109.99 | 168.15 | 265.80 | 24.04 | 6.13 | 53.70 | 23.13 | 2.6 | 2.2 | 1.3 |
LSD (5%) | | 4.911 | 4.446 | 19.746 | 1.957 | 0.597 | 13.187 | 1.937 | 0.30 | 0.28 | 0.231 |
CV (%) | | 7.60 | 4.7 | 12.3 | 14.0 | 16.19 | 21.8 | 14.4 | 19.3 | 21.5 | 29.9 |
***=significant at 0.001 significance level, **=significant at 0.01 significance level, *= significant at 0.01 significance level, ns = non-significant; a anthracnose disease severity score (1–5 scale), 1 = immune, 2 = resistant, 3–5 = susceptible, b disease severity score (1–5 scale); 1 = immune, 2 = resistant, 3 = moderately resistant, 4–5 susceptible, |
DTF = days to flowering, DTM = days to maturity, PHT = plant height (cm), PL = panicle length (cm), PW = panicle width (cm), PWT = panicle weigh per plant (g), TGW = thousand grain weight (g), ANTH = anthracnose, TLB = turcicum leaf blight, PGMR = panicle grain mold rating, Gen = genotype, GM = grand mean |
Additive Main Effects And Multiplicative Interaction (Ammi) Analysis
The result of AMMI ANOVA revealed that the genotype, environment, and GEI effects and the first two IPCAs were significant (P < 0.001) for all the traits except for TGW. The total variations explained due to treatment were over 60% for all traits studied (Table 5). Among these traits, GY explained the highest variation of about 92.7% for treatment and with the remaining 4.5% attributed to error.
Table 5
AMMI ANOVA for grain yield and related traits of 22 sorghum genotypes across eight environments.
| DTF | DTM | PHT | PL |
Source | DF | SS | MS | % TV | SS | MS | % TV | SS | MS | % TV | SS | MS | % TV |
Total | 527 | 119249 | 226.3*** | | 97804 | 185.6*** | | 2622389 | 4976*** | | 15529 | 29.47*** | |
Trt | 21 | 92841 | 530.5*** | 77.9 | 76304 | 436*** | 78.0 | 2174506 | 12426*** | 82.9 | 11349 | 64.85*** | 73.1 |
Block | 16 | 2955 | 184.7*** | | 967 | 60.4*** | | 88302 | 5519*** | | 362 | 22.61*** | |
Gen | 175 | 53176 | 2532.2*** | 44.6 | 18632 | 887.2*** | 19.1 | 368353 | 17541*** | 14.1 | 6721 | 320.05*** | 43.3 |
Env | 7 | 10044 | 1434.8*** | 8.4 | 43295 | 6185ns | 44.3 | 1544126 | 220589*** | 58.9 | 1429 | 204.12*** | 9.2 |
GxE | 147 | 29621 | 201.5*** | 24.8 | 14377 | 97.8*** | 14.7 | 262027 | 1782*** | 10.0 | 3199 | 21.76*** | 20.6 |
IPCA 1 | 27 | 17763 | 657.9*** | 60 | 6476 | 239.8** | 45 | 128782 | 4770* | 49.1 | 1561 | 57.81** | 49.1 |
IPCA 2 | 25 | 5253 | 210.1*** | 77.7 | 2837 | 113.5ns | 64.8 | 48197 | 1928* | 67.5 | 529 | 21.14* | 67.5 |
Error | 336 | 23453 | 69.8 | | 20533 | 61.1 | | 359581 | 1070 | | 3819 | 11.37 | |
| PW | PWT | TGW | | GY |
Source | DF | SS | MS | % TV | SS | MS | % TV | SS | MS | % TV | SS | MS | % TV |
Total | 527 | 1095.1 | 2.078*** | | 536186 | 1017*** | | 11818 | 22.43*** | | 458673103 | 870347 | |
Trt | 21 | 670.8 | 3.833*** | 61.3 | 333563 | 1906*** | 62.2 | 7404 | 42.31*** | 62.7 | 424998186 | 7332972*** | 92.7 |
Block | 16 | 94.4 | 5.903*** | | 32640 | 2040*** | | 691 | 43.21*** | | 2792470 | 174529** | |
Gen | 175 | 206.4 | 9.83*** | 18.9 | 102751 | 4893*** | 19.2 | 3258 | 155.16*** | 27.6 | 143992406 | 2428561*** | 31.4 |
Env | 7 | 207.2 | 29.6*** | 18.9 | 85804 | 12258*** | 16.0 | 1078 | 153.97*** | 9.1 | 120382615 | 17197516*** | 26.3 |
GxE | 147 | 257.1 | 1.749*** | 23.5 | 145008 | 986*** | 27.0 | 3068 | 20.87*** | 26.0 | 150623166 | 1024647*** | 32.8 |
IPCA 1 | 27 | 115.4 | 4.275*** | 48.8 | 74217 | 2749*** | 51.2 | 1322 | 48.97*** | 43.1 | 70046208 | 2594304*** | 46.5 |
IPCA 2 | 25 | 57.2 | 2.29*** | 65.3 | 25484 | 1019** | 68.8 | 666 | 26.62*** | 64.8 | 24113663 | 964547*** | 62.5 |
Error | 336 | 329.9 | 0.98 | | 169982 | 506 | | 3723 | 11.08 | | 20671354 | 91912 | 4.5 |
***=significant at 0.001 significance level, **=significant at 0.01 significance level, *= significant at 0.01 significance level, ns = non-significant |
DF = degree of freedom; SS = sum squares; MS = mean squares; Gen = genotype; Env = environment; Rep = replication; TV = percentage of total variance explained; G×E = genotype by environment interaction; IPCA = interaction principal component axis; Trt = treatment; DTF = days to flowering; DTM = days to maturity; PHT = plant height (cm); PL = panicle length (cm); PW = panicle width (cm); PWT = panicle weight per plant (g); TGW = thousand grain weight (g); GY = grain yield (kg ha− 1); IPCA1 = first interaction principal component axis; IPCA2 = second interaction principal component axis |
The proportion of the total variance explained by genotype, GEI, and environment for DTF was 44.6%, 24.8%, and 8.4%, respectively. For DTM, genotype variance accounted for 19.1% of the total variance whereas environment and GEI contributed 44.3% and 14.7% to the total variance, respectively. In the case of PHT, genotype accounted for 14.1% of the total variance whilst environment and GEI contributed 58.9% and 10%, respectively. Similarly, of the total variation the variances due to genotype for other agronomic traits such as PL, PW, PWT, and TGW accounted for 43.3%, 18.9%, 19.2%, and 27.6% whereas variation due to environment was 9.2%, 18.9%, 16%, and 9.1% and variances explained by GEI was 20.6%, 23.5%, 27.0%, and 26.0%, respectively. For GY, the genotype, environment, and GEI explained 31.4%, 26.3%, and 32.8% total variance, in respective order. The highest variance of 32.8% was explained due to GEI for GY indicating the interaction effect of genotype and environment as a major factor influencing the yield performance of genotypes. This result revealed differential yield performance among genotypes across testing environments and the presence of strong genotypes and GEI.
Genotype By Environment Interaction
The mean values of each trait in each environment based on AMMI analysis (IPCA1 and IPCA2 scores) and the four top-ranking genotypes for each trait at each environment are presented in Table 6. In AMMI analysis, low IPCA1 scores imply a high contribution to genotype stability and a low contribution to GEI. The IPCA1 scores showed that Chawaka and Gute were the main contributors to the stability of genotypes for GY. On the contrary, the most contributor environments for genotype stability in terms of PWT were Gute, Bilo, and Jimma, respectively. Haro Sabu, Bilo, Bako, and Gute environments contributed most to the genotype's stability in TGW. In this study, among testing environments, Gute contributed the most to genotype stability in GY, TGW, and PWT.
Table 6
AMMI analysis based on mean for phenotypic and IPCAe values, and four top ranking genotypes for each trait in each environment
Traits | Environment | Mean | IPCAe1 | IPCAe2 | 1 | 2 | 3 | 4 |
GY | Asosa | 1442 | -38.44 | -21.35 | G7 | G2 | G12 | G9 |
Bako | 2905 | 27.57 | 20.11 | G14 | G12 | G22 | G9 |
Bilo | 2329 | 20.82 | 9.33 | G14 | G12 | G9 | G13 |
Chawaka | 2194 | 0.74 | 13.11 | G14 | G2 | G1 | G9 |
Gute | 1657 | 10.92 | -4.52 | G14 | G12 | G2 | G22 |
Haro Sabu | 2811 | -29.48 | 21.96 | G22 | G7 | G2 | G17 |
Jimma | 2351 | 27.83 | -34.60 | G12 | G13 | G9 | G1 |
Uke | 2494 | -19.95 | -4.04 | G7 | G12 | G2 | G15 |
PWT | Asosa | 45.6 | 3.07 | 3.73 | G2 | G9 | G4 | G14 |
Bako | 65.92 | 4.38 | 1.46 | G2 | G14 | G9 | G4 |
Bilo | 45.05 | 0.79 | 2.19 | G14 | G9 | G4 | G2 |
Chawaka | 66.45 | 2.10 | -5.79 | G2 | G1 | G14 | G11 |
Gute | 30.86 | 0.78 | 2.69 | G9 | G14 | G17 | G11 |
Haro Sabu | 70.67 | -10.68 | -0.35 | G7 | G10 | G15 | G2 |
Jimma | 48.32 | 1.89 | -5.33 | G2 | G13 | G11 | G12 |
Uke | 56.75 | -2.33 | 1.40 | G2 | G12 | G17 | G7 |
TGW | Asosa | 23.15 | -3.55 | 1.18 | G14 | G6 | G13 | G9 |
Bako | 24.41 | 0.50 | 0.37 | G15 | G17 | G20 | G13 |
Bilo | 21.94 | 0.27 | 0.56 | G15 | G17 | G9 | G2 |
Chawaka | 26.16 | 2.32 | -0.32 | G15 | G16 | G17 | G9 |
Gute | 21.62 | 0.73 | 1.20 | G15 | G17 | G16 | G10 |
Haro Sabu | 22.34 | -0.24 | 0.15 | G15 | G17 | G9 | G5 |
Jimma | 23.38 | -1.03 | -3.38 | G3 | G15 | G1 | G5 |
Uke | 22.06 | 1.00 | 0.26 | G15 | G2 | G3 | G17 |
PWT = panicle weight per plant (g), TGW = thousand grain weight (g), GY = grain yield (kg ha− 1), IPCAe1 and IPCAe2 = the first and the second interaction principal component axis score of environments, G1 = ETSL101524, G2 = ETSL100445, G3 = ETSL100091, G4 = ETSL101312, G5 = IS25531, G6 = ETSL101732, G7 = ETSL100152, G8 = ETSL100561, G9 = ETSL100474, G10 = ETSL101128, G11 = Bonsa, G12 = ETSL100808, G13 = IS38326, G14 = Merera, G15 = IS38331, G16 = ETSL101858, G17 = ETSL100072, G18 = ETSL100771, G19 = ETSL101173, G20 = IS38412, G21 = ETSL101712, G22 = ETSL100587 |
The AMMI-2 biplot shows environment scores with the first two interaction principal component axis ( IPCA1 and IPCA2) for GY, TGW, PWT, PHT, PL, and PW (Fig. 1 and Table 5). According to AMMI-2 biplot analysis, environments with low IPCA1 and IPCA2 scores placed close to the origin have a high contribution to the stability of genotypes and a low contribution of genotypes by environment interaction. Accordingly, Gute, Chawaka, and Bilo environments showed low IPCA1 and IPCA2 scores positioned close to the origin in terms of GY indicating the contribution of the environments to genotype stability in GY. Similarly, Uke, Bilo, and Gute environments had a contribution to genotype stability for PWT. Whereas, Haro Sabu, Bako, and Bilo environments contributed more to genotype stability in terms of TGW.
Genotype Performance And Ammi Stability Analysis
Genotype means performance and AMMI-2 stability analysis were presented in Table 4 and Table 7, respectively. The mean performance indicated that genotypes ETSL100808, Merera, ETSL100445, ETSL100152, ETSL100474, and ETSL100587 had high mean GY (kg ha− 1) of 3,306.59, 3,044.62, 3,005.11, 2,928.16, 2,910.82, and 2,619.95, respectively. Two of the genotypes (IS38412 and ETSL101173) had the least GY and PWT (Table 4). The higher PWT was recorded for five genotypes (ETSL100445, Merera, ETSL101524, ETSL100808 and ETSL100072), and high TGW for IS38331, ETSL100072, ETSL100474, ETSL101858, and ETSL100091. The three high-yielding genotypes (ETSL100808, Merera, and ETSL100445) had also higher PWT (g) of, 64.18, 66.13, and 95.22, respectively. The longest panicle was recorded in ETSL100474 (29.22 cm), ETSL100152 (29.02 cm), and ETSL100072 (28.82cm) whereas Merera (16.66 cm) had the shortest panicle. For PW, ETSL100445 (7.96 cm), Merera (7.07 cm), and IS38326 (6.91 cm) had the widest panicle, whilst IS38412 and ETSL101173 genotypes were recorded as the narrowest panicles with 4.97 cm and 5.57 cm, respectively.
Table 7
The ASV, GSI, and combined mean performance of twenty sorghum genotypes for grain yield and panicle weight.
Gen | Mean | GY | PWT | |
IPCAg1 | ASV | RYI | GSI | Mean | IPCAg1 | RASV | RYI | GSI |
ETSL101524 | 2467 | -1.8 | 5.3 | 7 | 12.3 | 64.4 | 1.5 | 5.9 | 3 | 8.9 |
ETSL101128 | 1580 | -20.8 | 61.6 | 20 | 81.6 | 49.3 | -7.0 | 20.3 | 14 | 34.3 |
Bonsa | 2301 | 15.3 | 44.4 | 10 | 54.4 | 56.5 | 0.6 | 3.9 | 11 | 14.9 |
ETSL100808 | 3307 | 4.0 | 14.6 | 1 | 15.6 | 64.2 | -0.5 | 1.9 | 4 | 5.9 |
IS38326 | 2260 | 16.7 | 52.2 | 11 | 63.2 | 57.4 | -2.2 | 6.8 | 9 | 15.8 |
Merera | 3045 | 19.1 | 57.3 | 2 | 59.3 | 66.1 | 3.3 | 9.5 | 2 | 11.5 |
IS38331 | 2406 | 1.3 | 20.7 | 9 | 29.7 | 56.6 | -2.7 | 8.0 | 10 | 18.0 |
ETSL101858 | 1976 | 8.0 | 24.7 | 16 | 40.7 | 36.8 | 1.9 | 5.5 | 21 | 26.5 |
ETSL100072 | 2411 | -10.0 | 29.9 | 8 | 37.9 | 62.2 | -0.6 | 2.9 | 5 | 7.9 |
ETSL100771 | 1855 | 1.5 | 7.6 | 18 | 25.6 | 43.7 | 1.3 | 4.7 | 18 | 22.7 |
ETSL101173 | 1573 | 10.9 | 31.8 | 21 | 52.8 | 37.7 | 2.1 | 6.3 | 20 | 26.3 |
ETSL100445 | 3005 | -21.9 | 63.7 | 3 | 66.7 | 95.2 | 1.3 | 5.4 | 1 | 6.4 |
IS38412 | 1017 | 5.9 | 20.1 | 22 | 42.1 | 23.0 | 2.2 | 6.4 | 22 | 28.4 |
ETSL101712 | 2065 | -5.1 | 14.8 | 15 | 29.8 | 44.1 | 1.3 | 3.8 | 17 | 20.8 |
ETSL100587 | 2620 | -4.4 | 33.2 | 6 | 39.2 | 60.3 | -1.7 | 4.9 | 7 | 11.9 |
ETSL100091 | 1972 | -4.6 | 22.0 | 17 | 39.0 | 54.6 | 0.5 | 1.7 | 13 | 14.7 |
ETSL101312 | 2188 | 4.8 | 14.5 | 12 | 26.5 | 58.2 | 2.7 | 8.0 | 8 | 16.0 |
IS25531 | 1831 | 3.7 | 10.8 | 19 | 29.8 | 46.1 | 0.0 | 0.5 | 15 | 15.5 |
ETSL101732 | 2098 | 10.7 | 31.6 | 14 | 45.6 | 42.9 | 1.1 | 4.4 | 19 | 23.4 |
ETSL100152 | 2928 | -48.4 | 140.5 | 4 | 144.5 | 54.9 | -7.1 | 20.8 | 12 | 32.8 |
ETSL100561 | 2188 | 3.8 | 11.1 | 13 | 24.1 | 45.2 | 0.5 | 1.4 | 16 | 17.4 |
ETSL100474 | 2911 | 11.3 | 33.7 | 5 | 38.7 | 62.0 | 1.5 | 5.1 | 6 | 11.1 |
Gen = genotype, ASV: AMMI stability value, RYI = rank of mean yield, GSI: genotype selection index, IPCAg1: first interaction principal component axis scores for genotype, PWT: panicle weight per plant (g), GY: grain yield (kg ha− 1), |
Table 8
Mean yield, regression coefficients, and deviation from regression using Eberhart and Russell model
Gen | Mean | Rank | bi | Rank | s2di | Rank | YSi | |
ETSL101524 | 2466.94 | 7 | 1.37 | 16 | 3.29 | 9 | 7 | |
ETSL100445 | 3005.11 | 3 | 0.68 | 13 | 5.62 | 19 | 3 | |
ETSL100091 | 1971.6 | 17 | 0.64 | 15 | 1.62 | 7 | 17 | |
ETSL101312 | 2188.22 | 12 | 0.94 | 5 | 4.93 | 2 | 8.5 | |
IS25531 | 1831.16 | 19 | 0.99 | 2 | 11.28 | 10 | 19 | |
ETSL101732 | 2098.36 | 14 | 1.26 | 11 | 5.27 | 5 | 15 | |
ETSL100152 | 2928.16 | 4 | 0.29 | 19 | 4.15 | 21 | 4 | |
ETSL100561 | 2188.07 | 13 | 1.0 | 1 | 2.12 | 3 | 10.5 | |
ETSL100474 | 2910.82 | 5 | 1.09 | 6 | 2.68 | 8 | 5 | |
ETSL101128 | 1579.58 | 20 | 0.12 | 22 | 2.26 | 17 | 20 | |
Bonsa | 2301.19 | 10 | 1.05 | 4 | 3.69 | 13 | 12 | |
ETSL100808 | 3306.59 | 1 | 1.21 | 9 | 0.76 | 15 | 1 | |
IS38326 | 2259.63 | 11 | 1.13 | 8 | 3.75 | 21 | 13 | |
Merera | 3044.62 | 2 | 1.29 | 12 | 0.32 | 21 | 2 | |
IS38331 | 2406.4 | 9 | 1.76* | 21 | 7.87 | 11 | 10.5 | |
ETSL101858 | 1975.63 | 16 | 0.48 | 18 | 2.25 | 6 | 16 | |
ETSL100072 | 2410.86 | 8 | 1.03 | 3 | 0.54 | 12 | 8.5 | |
ETSL100771 | 1855.4 | 18 | 1.38 | 17 | 2.46 | 4 | 18 | |
ETSL101173 | 1572.89 | 21 | 1.13 | 7 | 1.45 | 14 | 21 | |
IS38412 | 1017.22 | 22 | 0.78 | 10 | 17.75 | 16 | 22 | |
ETSL101712 | 2065.34 | 15 | 0.64 | 14 | -0.02 | 1 | 14 | |
ETSL100587 | 2619.95 | 6 | 1.74* | 20 | 1.94 | 18 | 6 | |
GM | 2272.9 |
Yi: Mean response, bi: regression coefficient, sdi2 = deviations from regression component of interaction, YSi: yield and stability index, GM = grand mean, |
A low AMMI stability value (ASV) in AMMI analysis denotes good genotype stability and low GEI (Purchase et al., 2000). The genotype with the lowest ASV score is the one that adapts across environments, and the higher the ASV value, whether positive or negative, the more specifically a genotype is adapted to a given environment. Genotypes ETSL100771, IS25531, ETSL100561, and ETSL101312 with mean GY of 1,855, 1,831, 2,188, and 2,188 kg ha− 1, respectively showed high stability having low ASV (Table 7). However, these genotypes gave GY below the average mean. On the other hand, ETSL100808, ETSL100474, Merera, and ETSL100445 were identified as high-yielding and stable based on the genotype selection index (GSI). For PWT, ETSL100808, ETSL100474, and ETSL100445had high stability. Therefore, ETSL100808, ETSL100474, and ETSL100445 had high GY and PWT and were stable across environments.
Among evaluated genotypes, ETSL100587 and ETSL100152 had high mean GY but were placed far from the biplot origin suggesting that they were not stable. These genotypes had their specific adaptation environment; ETSL100587 at Haro Sabu and ETSL100152 at Asosa. The positive interaction of ETSL100091and IS38326 with environment Jimma, and ETSL100587 and IS38331 with environment Bako for GY indicated the specific adaption for respective environments (Table 4 and Fig. 1b).
Gge Biplot Analysis
Which-won-where polygon view of GGE biplot
Stability analysis of the genotypes based on their IPCA scores using the GGE biplot analysis and the polygon view of the GGE biplot showed the best-performing genotypes and interaction patterns among genotypes and environments (Fig. 2). In the current study, the GGE biplot analysis of the twenty-two sorghum genotypes tested in eight environments showed that the sum of the first two principal components explained 74.9, 77.20, 76.90, 81.56, 85.36, and 72.17% for GY, PWT, TGW, PHT, PL, and PW, respectively (Fig. 1 and Fig S1). In this GGE biplot, the polygon of lines is made by connecting vertex genotypes with straight lines and the rest of the genotypes falling inside the polygon. The vertex genotypes for GY were ETSL100808, Merera, ETSL100152, ETSL101128, IS38412, ETSL101173, and IS38326 whereas Merera, ETSL100445, ETSL100152, ETSL101128, and IS38412 were taken as the vertices for PWT. These two sets of genotypes were the most responsive to environmental interactions for GY and PWT. Therefore, these genotypes are either the best or poorest genotypes in some or all environments because they are farthest from the origin. Similarly, genotypes (IS38331, ETSL100445, ETSL100587, Merera, IS38412, and IS25531) forming vertices of the polygon for TGW, ETSL100474, ETSL100808, ETSL101524, Bonsa, ETSL101173, ETSL100771, and IS38326for PHT; ETSL100474, ETSL100152, ETSL100561, ETSL100587, Merera, IS38412, Bonsa and ETSL100072 for PL, and Merera, ETSL101524, ETSL100091, IS38412 and ETSL100587for PW were the most responsive (Fig S1).
In the present study, three mega environments (MGEs) were formed for GY, PL, and PWT whilst two MGEs were formed for PHT, PWT, and TGW. Genotypes that were positioned inside the sector of the environment showed top performance for the studied trait. For GY, the first MGE contained Jimma, Bako, and Bilo with winning genotype Merera; the second MGE contained Gute and Chawaka with winning genotypes ETSL100808 and ETSL100474; the third MGE contained Uke, Haro Sabu and Asosa with winning genotypes ETSL100445and ETSL100152 (Fig. 2). On other hand, for PWT and PHT, environments Bako, Bilo, Chawaka, Uke, Jimma, Gute, and Asosa formed with ETSL100808 as a better-performing genotype whereas environment Haro Sabu was separated from MGE. For PL and PW, Asosa was separated solely as MGE where IS25531 performed well. Environment Uke, Gute, and Jimma jointly formed while Haro Sabu, Chawaka, Bako, and Bilo were separate MGE for these traits with winning genotypes ETSL100152 and ETSL100474, respectively (Fig S2). In addition to Asosa, a single MGE for PW, environments Haro Sabu, Uke, Gute, Bilo, and Jimma jointly formed as the second MGE whereas Bako and Chawaka were another third MGE for this trait. For the second and third MGEs, ETSL100445 and Merera are best in terms of their PW. On the contrary, environments Bako, Bilo, Gute, Chawaka, Uke, Haro Sabu, and Jimma were grouped into the same MGE whereas Asosa was separated by another MGE for TGW.
Genotype Ranking Based On Their Mean Performance And Stability
The higher mean performance of a genotype is indicated by an average environment axis (AEA) in the ranking biplot, which is depicted by a single arrowhead line passing through the origin. Genotypes with above-average means for GY were ETSL100808, Merera, ETSL100445, ETSL100152, ETSL100474, ETSL100587, IS38331, ETSL101524, ETSL100072, and Bonsa while IS38326, ETSL100561, ETSL101312, ETSL101712, ETSL101732, ETSL101858, ETSL100091, ETSL100771, IS25531, ETSL101128, ETSL101173, and IS38412 were genotypes which had below-average means of GY. For PWT, genotypes ETSL100445, Merera, ETSL101524, ETSL100808, ETSL100072, ETSL100474, ETSL100587, ETSL101312, IS38326, IS38331, Bonsa, ETSL100152, and ETSL100091 had above average PWT whereas genotypes ETSL101128, IS25531, ETSL100561, ETSL101712, ETSL100771, ETSL101732, ETSL101173, ETSL101858, and IS38412 had below average means of PWT. On the other hand, genotypes ETSL100152, ETSL100474, ETSL100072, Bonsa, and ETSL100808 showed high mean PL, and genotypes ETSL100445, Merera, IS38326, ETSL101524, and ETSL100561 had high mean PW. In contrast, genotypes Merera, IS38412, ETSL101732, ETSL101173, and ETSL100771, and IS38412, ETSL101173, IS25531, ETSL100091, and ETSL100771 showed low for PL and PW, respectively. For TGW, genotypes IS38331, ETSL100072, ETSL100474, ETSL101858, and ETSL100091had high mean values while genotypes Bonsa, ETSL101732, Merera, ETSL101712, and ETSL100587 showed lower mean values.
The stability of genotypes was evaluated based on the length of the vector (dotted line in the graph) between the genotype positions and the AEA in the ranking biplot (Fig. 2 and Fig S2). Genotypes such as ETSL100808, ETSL100474, ETSL100445, and Merera were the most stable with high grain yield that had shorter vectors from AEA whereas ETSL100152 and ETSL101128 were the least stable having the longest vector from AEA (Fig. 3a). For PWT, ETSL100445 and ETSL101524 were the most stable while ETSL100152 and ETSL101128 were the least stable (Fig. 3b). Contrastingly, the most stable genotypes for PL were ETSL100474, ETSL100152, and ETSL100808 whereas ETSL100587, ETSL100561 and Bonsa were the least stable. The most stable genotypes for PW were ETSL101524 and Merera while ETSL100561 was the least stable. Genotypes IS38331 and ETSL100072 were the most stable in terms of their TGW whereas ETSL100445 is the least stable (Fig S2).
Evaluation Of Environments In Comparison To Biplots
In the biplots analysis, environment-focused scaling of comparison of the GGE biplot shows AEA, AEC, and concentric circles which helps to evaluate the tested environments. A test environment that has a small angle with AEC i.e close to the center of concentric circles is the more representative/ideal of other test environments.
In the present study, the concentric circles on the comparison GGE biplot graph showing the distance of the environments to AEA, AEC, and the biplot origin were displayed in Figs. 4 and Fig S3. Accordingly, Chawaka, Gute, and Uke were the ideal environments for grain yield as they are the closest to the center of concentric circles and have the smallest angle with AEA. Conversely, Asosa was the most discriminative environment for GY as it was placed far from the biplot origin. On the other hand, Jimma was a representative environment for PWT whereas Haro Sabu was the ideal environment for PL and PW. For TGW, Bako was the ideal environment. Asosa was the discriminative environment for TGW, PL, and PW whilst Haro Sabu was the discriminative environment for PHT and PW.
In this study, the biplot showed that the ideal genotype for GY and PWT was ETSL100808 and ETSL100445, respectively. Similarly, the biplot showed that four genotypes (ETSL100808, ETSL100474, ETSL100445, and IS38331) were ideal for PHT, PL, PW, and TGW, respectively (Fig S4). Based on these criteria, ETSL100808, ETSL100474, ETSL100445, and Merera had regression close to unity and deviation from regression close to zero with high mean grain yield and hence could be considered as stable genotypes and variety, respectively. Hence, the Merera sorghum variety can be demonstrated across the testing environments as a variety with broad adaptation. In advance, the ETSL100808 genotype determined for its stability and grain yield greater than standard checks (Bonsa and Merera) can be proposed for variety verification and would be released as a potential new cultivar across testing environments.
In addition, ETSL100072 and ETSL100587 also showed regression close to unity and deviation from regression close to zero; but their grain yield yet not exceeds that of standard checks to be proposed as variety. Genotypes such as ETSL100561 and ETSL101712 were stable across environments, however, showed mean grain yield less than average and hence were not selected for possible release.