Selection of Major Yield Contributing Traits by Multiple Linear Regression Model in Finger Millet (Eleusine coracanaL.)

Y. A. Nanja Reddy *

All India Coordinated Small Millets Improvement Project, Bangalore-560065, Karnataka, India and Department of Crop Physiology University of Agricultural Sciences, GKVK, Bangalore-560065, Karnataka, India.

M. V. Channabyre Gowda

All India Coordinated Small Millets Improvement Project, Bangalore-560065, Karnataka, India.

*Author to whom correspondence should be addressed.


Abstract

Yield improvement of finger millet through yield-based selection has plateaued in the last two decades. Hence, breeding efforts are required to prioritize the trait-based improvement in addition to yield per se. Using the backward multiple linear regression, and path analysis, the independent parameters, mean ear-head weight (MEW), ear-head number/ plant (ENo.) and the threshing percentage, were found significant contributors to grain yield. Genetic resources for the higher mean ear-head weight of >10.22g (GE-4683, GE-4596) and ear-head number >104.7m-2 (RAU-8, PR-202) were selected over the popular variety, GPU-28 (7.59g and 72.7m-2, respectively). Theoretically, incorporating the higher MEW and ear-head number/plant from the identified lines into popular variety, GPU-28 (shy tillering) in multiple regression model has predicted the possibility of increasing yield of GPU-28 by 17.8% and 29.5%, respectively. Therefore, a shift towards trait-based improvement could be appropriate to break the yield plateau of finger millet.

Keywords: Genetic resources, path analysis, multiple linear regression, yield prediction, finger millet


How to Cite

Reddy , Y. A. Nanja, and M. V. Channabyre Gowda. 2024. “Selection of Major Yield Contributing Traits by Multiple Linear Regression Model in Finger Millet (Eleusine coracanaL.)”. International Journal of Plant & Soil Science 36 (5):533-43. https://doi.org/10.9734/ijpss/2024/v36i54550.