Yield Estimation of Dolichos Bean (Lablab purpureus L.) Using Different Clustering Techniques

Abstract

Dolichos bean (Lablab purpureus L.) is an important leguminous vegetable crop. Diversity of genes to a high extent in the gene pool provides better opportunities to develop appropriate plant varieties. Hence, finding advantageous traits or their combinations is the key step. Genetic diversity is important for parents to be chosen for potential breeding programmes. In general, genetically divergent parents are used in segregating generations to produce the optimal recombinants. Green seed yield of the crop is influenced by number of yield contributing characters such as, green pod yield, pods per plant, days to 50 percent flowering, number of primary branches, raceme per plant, nodes per raceme, raceme length and 100 fresh seed weight. The association between the characters and yield provides basis for further breeding programme. Cluster analysis was used to classify the germplasm accessions into different clusters based on the morphological traits of the accessions. The data on germplasm accessions (64) was collected from the Department of Genetics and Plant Breeding, UAS, GKVK, Bengaluru, Karnataka. Euclidean distance metric was used to measure the difference between germplasms. Elbow method indicated three optimum genotype clusters. Three clustering methods were used viz., single, complete and average linkage. Dunn and Silhouette indices were used to validate the clusters. It was found that Average linkage method was best among the three. Further, in the third cluster by average linkage method with 12 genotypes was found to have highest green seed yield and green pod yield. Hence these 12 genotypes can be used for further breeding programmes.