Association between Grain Size, Shape and Thousand Kernel Weight in Pakistani Wheat Landraces

Abstract

Wheat (Triticum aestivum. L) grain size is considered to be one of the main criteria of yield constituents by wheat breeders. In order to detect phenotypic diversity and relationship between thousand kernel weight, a collection of 204 landraces from different parts of Pakistan was planted across two years (2012-2014). High throughput method based on seed imaging was used to measure the kernel size and shape. The correlation analysis revealed significant positive correlation between thousand kernel weight (TKW) with perimeter of vertical (PV), thickness (T), area of vertical (AV), area of horizontal (AH), perimeter of horizontal (PH), width (W) and Length (L). Bioplot showed that accessions with high seed shape parameters have higher TKW. By taking TKW as a dependent variable multiple regression analysis was performed. Regression summary indicated that 31% of the variations in TKW are explained by the independent variables. It was shown that grain thickness, length and width are most important for predicting TKW. Based on image analysis this study provides useful information about the relationship between TKW, kernel size and shape in Pakistani wheat landraces that may help to improve grain weight in a breeding program.