Acta Scientiarum http://periodicos.uem.br/ojs ISSN on-line: 1807-8664 Doi: 10.4025/actascitechnol.v45i1.61270 TECHNOLOGICAL INFORMATION Acta Scientiarum. Technology, v. 45, e61270, 2023 Near infrared spectroscopy as a tool for agricultural expertise: identification of tomato seedlings Amanda Rodrigues Vinhandelli 1 , Annelisa Arruda De Brito 1 , Raquel Cintra de Faria 1 , Luiz Fernandes Cardoso Campos 1 , Gilberto Alessandre Soares Goulart 1 , Gustavo Henrique de Almeida Teixeira 2 , Abadia dos Reis Nascimento 1 and Luís Carlos Cunha Junior 1* 1 Escola de Agronomia, Universidade Federal de Goiás, Avenida Esperança, s/n., 74690-900, Goiânia, Goiás, Brazil. 2 Faculdade de Ciências Agrárias e Veterinárias, Universidade Estadual Paulista, Jaboticabal, São Paulo, Brazil. *Author for correspondence. E-mail: cunhajunior.l.c@ufg.br ABSTRACT. Tomatoes are one of the most prominent vegetables globally, with significant cultural and economic relevance in various nations, including Brazil. The term ‘safe food’ is becoming more popular as consumer preferences and supply chain dynamics become evolved in these processes. In light of these issues, the use of safety and quality management methods for fruits and vegetables have increased dramatically, with traceability being one of these solutions worth highlighting. When it comes to traceability, evaluation of tomato seedlings, plants, and fruits to identify groups or hybrids becomes particularly crucial throughout the marketing process, since the consumer of seedlings or fruit has difficulties recognizing whether that product truly belongs to the group indicated by the merchant. Thus, the potential of near infrared spectroscopy (NIRS) combined with the PC-LDA and PLS-DA algorithms was tested for the discrimination of two significant commercial groups, Salada and Saladete, as well as eleven cultivars belonging to these groups, which were tested for this purpose. The results show that, by using the PLS-DA model, the portable NIR equipment is capable of differentiating tomato seedlings in nurseries of the Salada and Saladete groups, with an accuracy of 99.7% and sensitivity of 100%. The technique showed to be efficient for individual models of tomato seedlings in the Salada group, with accuracy over 90% and sensitivity above 93% for all models. For the Saladete group's individual models, the technique proved effectiveness for the hybrids Parma, BS-110012, Giácomo, Guara, and Tyna. Keywords: hybrids; PC-LDA; PLS-DA; traceability; Solanum lycopersycum L. Received on October 20, 2021. Accepted on May 17, 2022. Introduction Tomato cultivation for fresh consumption has a high production cost, of which the inputs, labor and seeds correspond, on average, for 60% of the total cost (Pagliuca, Deleo, Boteon, Mueller, & Valmorbida, 2017). Thus, profitability begins with the choice of the tomato and hybrid group to be implemented, as well as the production and delivery of healthy seedlings (Diniz, Guimarães, & Luz, 2006). However, it is not uncommon for loads or seedlings to be exchanged due to lack of inputs, such as lack of seeds of a certain hybrid, resulting in the delivery of tomato seedlings from the same group, but not of the same genetic material, causing economic damage that normally occurs at the time of production. These cases end up being prosecuted, and consequently subjected to expertise. Regarding the performance of judicial expertise, the elucidation is technically effective, but ineffective for reducing damages to the producer. This ambiguity occurs because such identification process is carried out in the fruit production stage, a period of easy distinction between genetic materials. However, at this point, the economic damage has already been done, as the implantation and input values were already spent at the time of the inspection. According to Pagliuca et al. (2017), these stages present values that can reach up to 50% of the total cost of production. Thus, it is necessary to study methods and techniques that allow an expert to identify plant material quickly and before planting. In this context, near infrared spectroscopy (NIRS) can be an alternative technology. In the field of non- destructive testing, the technique of spectroscopy in the near infrared range presents itself as a fast tool, which allows real-time analysis, thus demonstrating reliable results, which reduces the cost and time spent on routine analysis in laboratories (Brimmer & Hall, 2001; Muñiz, Magalhães, Carneiro, & Viana, 2012). Dale