Phenotypic stability in sorghum lines evaluated by Multiple Factor Analysis

ROSAS, BELÉN; ORTIZ, DIEGO; GUILLERMO RAUL PRATTA; GUIAMET, J.J.

Córdoba . Formato virtual

Simposio; Primer Simposio de Mejoramiento Vegetal; 2021

INTA - UNC

Multiple factor analysis (MFA), developed by Escofier and Page҅s (1), is a 3-way factorial method adapted to the analysis of data tables in which the same set of individuals is described through several groups of variables. These groups can arise from the joint use of variables of different nature - quantitative or qualitative - or when evaluated under different environmental conditions. The objective was to study the relationships between phenotypic variables in a representative panel of sorghum lines and evaluate their stability over three years using a 3-way multivariate analysis. The AFM method was applied to a group of 137 sorghum genotypes belonging to the Sorghum Association Panel. The trials were carried out at EEA INTA Manfredi (39° 49 ́ S, 63° 46 ́O) in α-lattice designs, for three years. Five quantitative variables were evaluated: plant height (ALT), leaf chlorophyll content (SPAD), specific leaf area (AFE), stomatal density (DE) and captured energy efficiency (Fv ́/Fm ́). For the statistical analysis in each year, a table of two modes (genotypes by variables) was obtained and then all the information was considered in a set of three tables of genotypes by variables. The information was thus summarized in multiway matrices, in which each data was originated by three independent factors: genotypes, variables and years. All results were obtained through the R (FactoMiner Package) software. In this work the factorial plane generated by the first two components was considered for the analysis, since it shows the main characteristics, similarities and differences between the sorghum lines. In the analysis of vector-variables most of the vectors representing the same variable in the three years are in close proximity, which indicates that they are stable, except for DE whose vectors fall into different quadrants. However, the latter vectors are very short so they do not make a significant contribution to the overall variability. When comparing years, we can see that Year 1 and Year 2 are similar on Dim 1, highly associated in opposite directions (one positive and one negative) to ALT and SPAD in all three years, while Year 1 and Year 3 are similar on Dim 2, highly associated with the same variables, but in the same sense. Of all the variables analyzed, ALT and SPAD show the greatest variability in this group of genotypes and remain stable over the years. Although the total variability explained by the first two dimensions is intermediate to low (37.6%), these variables are the ones that best allow to differentiate the genotypes in two groups on the first two dimensions. The multiple factor analysis allowed to verify the phenotypic stability during three years of evaluation as well as to distinguish the main sources of variability for quantitative characters of agronomic interest in a representative panel of sorghum genotypes. References:(1) Escofier, B. and Page҅s, J.,1990. Single and Multiple Factor Analysis: Objectives, Methods and Interpretation. University of the Basque Country, Bilbao, 285 p.