Evaluation the Mean Performance and Stability of Lentil Genotypes by Combining Features of AMMI and BLUP Techniques

In this research, 12 selective advanced genotypes of lentil with Kimia and Gachsaran checks were grown for three growing years (2010-2013) in four locations including Gachsaran, Gonbad, Khoramabad and Moghan using randomized complete block design with three replicates in each location. The heatmap plot indicated the variation of seed yield of genotypes in different environments. Mosaic plot showed that the portion of sum squares of genotype (G) and sum squares of genotype by environment interaction (GEI) in total sum of squares (TSS) were 17.53% and 82.47%, respectively. The likelihood ratio test (LRT) indicated that the effect of GEI was significant on seed yield and therefore for evaluation of stability of genotypes, singular value decomposition (SVD) was performed on the matrix of best linear unbiased prediction (BLUPs) of GEI. The screet test showed that the first five principal components had a significant contribution in the GEI matrix derived from BLUP, as the first and second principal components explained only 32.28% and 26.95% of the GEI variation, respectively. The bilot of first principle component (PC1) of the environment versus nominal yield showed that genotypes 8, 4, 3, 14 and 7, due to the lowest scores of the PC1, had a small share in the GEI and were more stable. Biplot of seed yield versus WAASB (weighted average of absolute scores) placed genotypes in four regions, so that the genotypes 4, 6, 8, 9, 10 and 12 in the fourth region were very productive and stabile due to the large value of response variable (high seed yield) and high stability (low values of WAASB). Identification of genotypes with WAASBY (weighted average of the stability (WAASB) and mean performance (Y)) showed genotypes 4, 6, 8 and 9 as high yielding and stable, and therefore can be candidate for cultivar introduction.