Selection of Parent Materials for Alfalfa Recurrent Selection Using a Logistic Model

In alfalfa breeding, traditional recurrent selection methods often rely on extensive field trials and empirical judg-ment, which are inefficient and lack accuracy. This experiment attempts to introduce a logistic regression model combined with the analysis of alfalfa agronomic traits to select hybrid parents for alfalfa materials, thereby improving the efficiency and accuracy of recurrent selection. Using 20 alfalfa materials as subjects, the experiment involved agronomic trait analysis, variation analysis, cluster analysis, and the construction of a logistic model to evaluate and screen the alfalfa materials. The results showed that the 20 alfalfa materials were clustered into four clusters with similar performances. Based on the growth performance at the initial flowering stage, the best-performing alfalfa in autumn and spring was in cluster II. Around the 3.5th week of spring, cluster III > cluster II, showing the fastest growth. According to the predictions from the logistic fitting curve, the growth performance of cluster IV alfalfa surpassed that of cluster II around the 7th week, which was inconsistent with the growth performance before the initial flowering stage, revealing the genetic potential of cluster IV alfalfa in plant height traits. The results indicate that the Logistic model can improve the selection accuracy in alfalfa breeding, avoid the waste of genetic resources, and provide important reference value for the selection of parents in recurrent selection of alfalfa.