Desenvolvimento e validação de um painel molecular para identificação da cultivar BRS Integra (Urochloa ruziziensis)

Abstract

The BRS Integra cultivar of Urochloa ruziziensis, commercially released in March 2022 by Embrapa Gado de Leite, represents the first cultivar of the species developed through a breeding program for Brazilian conditions. Its morphological similarity to cv. Comum makes visual distinction in the field difficult, generating demand for molecular tools to certify the identity and genetic purity of commercialized seeds. Thus, the present study aimed to develop and validate a molecular panel of microsatellite (SSR) markers for identifying and differentiating the BRS Integra cultivar from cv. Comum of U. ruziziensis. For this purpose, 500 individuals from four commercial seed producers were genotyped using eight SSR loci in a multiplex PCR system with capillary electrophoresis. Marker characterization revealed a mean PIC of 0.6499 and a combined probability of identity of 1:20,000,000 for the eight loci. Genetic structure analysis identified significant differentiation between the cultivars (FST=0.018; p<0.001) and detected one commercial lot labeled as BRS Integra with a genetic profile compatible with cv. Comum, evidenced by the convergence of four independent analyses (FST, Nei's distance, UPGMA dendrogram, and STRUCTURE), with PCoA providing additional qualitative evidence, not conclusive on its own. Application of the chi-square test with Bonferroni correction and a minimum contrast criterion of eight percentage points resulted in the selection of nine diagnostic alleles across six SSR loci, with Cramér's V ranging from 0.1087 to 0.2706. The panel showed a combined probability of discrimination greater than 99.99% for the five independent loci. The overall accuracy of 70.3% obtained by LOO-CV is biologically consistent with the actual genetic overlap between the cultivars, particularly in Population 3 (Q_Integra=0.51), and does not reflect the panel's performance in its practical application, which relies on comparing allele frequencies at the lot level. Qualitative validation in the breeding program's reference populations confirmed the directional pattern of all nine alleles. Bootstrap sample rarefaction analysis indicated a minimum sample size of 25 to 30 individuals per lot for reliable estimation of diagnostic allele frequencies with a maximum deviation of 5%. The molecular panel developed constitutes the first tool specifically designed for molecular differentiation between BRS Integra and cv. Comum, effective at the population level and viable for implementation in varietal certification and seed quality control processes for forage crops in Brazil.