Impaginato 337 Adv. Hort. Sci., 2020 34(3): 337­347 DOI: 10.13128/ahsc­7814 Identification of promising tomato breeding lines with determinate growth by selection index M. Viera Nascimento 1 (*), M.C. Ribeiro Ávila 1, M. Fiori de Abreu­Tarazi 1, A.P. Oliveia Nogueira 2, L.F. Cardoso Campos 1, A. Dos Reis Nascimiento 1 1 Department of Horticulture, School of Agronomy, Federal University of Goiás, Goiania, Goiás, Brazil. 2 Federal University of Uberlândia, Uberlândia, Minas Gerais, Brazil. Key words: genetic parameters, processed tomato, Solanum lycopersicum. Abstract: Source of important vitamins, fibers, and minerals, the tomato (Solanum lycopersicum L.) stands out in the world agricultural scenario for its economic and social relevance and versatility. The Brazilian market is dominat­ ed by multinationals companies, and this market segment obtains cultivars from other countries, with genetics accurate to climatic conditions and cultiva­ tion method very different from those used in Brazil. As a result, the local culti­ vation of tomatoes plants becomes dependent on market variations and has required a material that has limited production efficiency. This study aimed to estimate genetic parameters from agronomic traits and to select industrial tomato lines using the selection index. A randomized block experimental design with three replications was used. Eighty­five industrial tomato lines from the germplasm bank of the Vivati Plant Breeding Ltda were evaluated. Each plot had 12 plants. The two central plants of each plot were evaluated. The evaluations were carried out using adapted morphological descriptors described in the guidelines for carrying out the distinguishability, homogeneity, and stability (DHE) tests of the Ministry of Agriculture, Livestock, and Supply of Brazil (MAPA). The genotypic determination coefficient (H2) of the traits related to fruit pericarp thickness, fruit firmness, fruit yield, average cycle, average number of fruits per plant, and soluble solids was high. The base index and the classic index presented the largest gain from selection for the fruit yield trait. Rank summation index and genotype­ideotype distance index had the highest total selection gain values. The tomato lines PXT­601 and PXT­610 stood out as superior genotypes by the methods of direct selection and by selection indexes. 1. Introduction Tomato is grown in different regions of the world and stands out as the most produced vegetable in the world, second only to potatoes in the cultivated area (Geraldini et al., 2018). Part of the success of tomatoes comes from its diversity in food and nutritional aspects that help human health. The fruit is rich in vitamins A and C and lycopene, substances that help prevent cancer of the gastrointestinal tract (Simão and Rodriguez, (*) Corresponding author: nascimento_mariana1@hotmail.com Citation: VIERA NASCIMENTO M., RIBEIRO ÁVILA M.C., FIORI DE ABREU­TARAZI M., OLIVEIA NOGUEIRA A.P., CARDOSO CAMPOS L.F., DOS REIS NASCI­ MIENTO A., 2020 ­ Identification of promising tomato breeding lines with determinate growth by selection index. ­ Adv. Hort. Sci., 34(3): 337­ 347 Copyright: © 2020 Viera Nascimento M., Ribeiro Ávila M.C., Fiori de Abreu­Tarazi M., Oliveia Nogueira A.P., Cardoso Campos L.F., Dos Reis Nascimiento A. This is an open access, peer reviewed article published by Firenze University Press (http://www.fupress.net/index.php/ahs/) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Competing Interests: The authors declare no competing interests. Received for publication 11 January 2020 Accepted for publication 18 February 2020 AHS Advances in Horticultural Science http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ Adv. Hort. Sci., 2020 34(3): 337­347 338 2008). Brazil is in ninth place in the world tomato pro­ duction ranking. At the top is China, accounting for 31% of production, followed by India with 11%, and the United States with 8% of global production (Dossa and Fuchs, 2017). In 2017, 36.688 hectares of tomato were grown in Brazil, 47.40% of the produc­ tion was destined for fresh consumption, and 52.60% for processing industries (Marcomini and Molena, 2018). The Brazilian national market is dominated by multinationals and acquires imported cultivars, with genetic characteristics adapted to climatic conditions and cultivation systems very different from those found in Brazil. As a result, the Brazilian cultivation of tomato becomes dependent on market swings. It obtains cultivars with productive potential restricted if compared to the yields reached in the environment that they were developed. Also, the plants may suffer losses by climate intolerances and plant diseases, when facing the Brazilian growing conditions. Due to the economic importance of the crop, tomatoes produced for processing industries have been the focus of research, especially in genetic breeding with the aim of produce cultivars that pos­ sess genes able to assist in the adaptation and toler­ ance to biotic and abiotic stresses, which can result in important contributions to the sector (Parmar et al., 2017). In a breeding program, the objective is to enhance the main phenotypic trait and conserve or improve the expression of secondary traits at the same time (Nogueira et al., 2012). However, the direct selection of quantitative traits is influenced by the environ­ ment, which may cause unfavorable changes in other traits (Vasconcelos et al., 2010). One way to improve this process is to use the simultaneous selection of a group of important agro­ nomic traits, that is, to use the selection indexes. These indexes relate information of different traits and make it possible to perform a selection effective­ ly, which increases the probability of success in a plant breeding program (Cruz et al., 2012; Vianna et al., 2013; Rezende et al., 2014). Considering the importance of the industrial pro­ cessing of tomatoes and the market demand for cul­ tivars that meet the requirements of this industrial chain, it is indispensable to know the relationship between agronomic traits and the study of the index­ es. This makes it possible to obtain the best predic­ tion of gains and yields and greater efficiency in the selection process. Given the above, this aimed to estimate genetic parameters for agronomic traits and to select industrial tomato lines using the selection index. 2. Materials and Methods The study was conducted in the experimental area of Vivati Plant Breeding Ltda, in Abadia de Goiás Unit, Goiás, Brazil, at 16°45’26” S, 49°26’15” W, and 898 m of altitude. The climate, according to Koppen, is classified as tropical humid, characterized by rainy summer with high temperatures and dry winter, with an average annual rainfall of 1.575 mm. The genotypes analyzed in this study are owned by Vivati Plant Breeding Ltda, which use their own selection and maintain methods. The seeds were sown in 450­cell polystyrene trays, filled with a sub­ strate composed of coconut fiber, rice husk, and peat and covered with vermiculite. The trays were kept in a greenhouse for 35 days when the seedlings had from two to three true leaves, and they were able to transplant to the field. The soil preparation was carried out with a tractor and rotary tiller. Seedbeds were prepared with 1.0 m wide, 0.20 m high, with 1.0 m spacing between beds. At the transplanting, 1.500 kg ha­1 of the NPK formu­ lation 04­30­10 was applied. As topdressing fertiliza­ tion, 20 kg ha­1 of MAP, 75 kg ha­1 of ammonium sul­ fate, 100 kg ha­1 of ammonium nitrate, and 200 kg ha­ 1 of potassium chloride were divided into four appli­ cations with a 20­day interval after transplanting. The seedlings were manually transplanted to the fi e l d 3 5 d a y s a ft e r s o w i n g ( D A S ) , w i t h 0 . 4 0 m between plants and 1.0 m between rows. Irrigation was performed by a drip system, supplying the water requirement based on the parameters for crop irriga­ tion management. Weed control was performed weekly to avoid competition. Insecticide baits were placed through­ out the field to identify the insect infestation rate and help the decision of pesticide application. Phytosanitary control was carried out whenever nec­ essary, to maximize fruit production (FAO, 2006). The tomato lines were characterized by morpho­ logical traits contemplated in the guidelines for per­ forming the distinguishability, homogeneity, and sta­ bility (DHE) assays by the MAPA, which were modi­ fied by the authors. A randomized block experimen­ tal design with three replications was used. Eighty­ Vieira Nascimento et al. ‐ Identification of tomato strains in Brazil 339 five industrial tomato lines were evaluated. Each plot had 12 plants. The two central plants of each plot were evaluated. The descriptors analyzed are shown in Table 1. It was estimated the genotypic determination coefficient (H2), according to the estimator below: Where: H2 = genotypic determination coefficient; ∅̂ = quadratic genetic component; QMT = mean square of genotypes; QMR = mean square of the residue; and ϒ = number of replications. Genotypes were grouped based on the Scott­ K n o tt t e s t a t t h e 1 % a n d 5 % p r o b a b i l i t y l e v e l . Subsequently, the selection gains estimates were reached by the aid of the selection index methodolo­ gies cited by Cruz (2006): direct and indirect selec­ tion; classic index proposed by Smith (1936) and Hazel (1943); rank summation index of Mulamba and Mock (1978); base index of Williams (1962); and genotype­ideotype distance index (GID). The selec­ tion criterion applied was to increase the traits: fruit pericarp thickness (FPT), fruit firmness (FF), yield (YLD), average number of fruits per plant (NFP), and soluble solids (SS). The index proposed by Smith (1936) and Hazel (1943) was established by the selection index (I) and the genotypic aggregate (H) described below: where: n = number of traits evaluated; b = vector of dimension 1 xn of the selection index weighting coefficients to be estimated; Table 1 ­ Descriptors for industrial tomatoes (adapted from MAPA, 2005) and details on their analysis Traits Trait description Description code Comments 01. Fruit pericarp thickness Slim S The analysis was performed using a digital caliper, measuring the diameter (mm) from the outer wall to the inner wall of the pericarp Average A Thick T 02. Fruit: firmness Soft S The analysis was performed by subjecting the fruits to pressure at one point in the middle region, measuring the resistance of the pulp to penetration, using Instrutherm model PTR­300 digital penetrometer, and obtaining the values expressed in Newton (N) Medium M Firm F 03. Maturation cycle Precocious P It was evaluated from the transplanting of seedlings Medium M Late L 04. Yield Low L It was determined by the weight and number of fruits per plant Average A High H 05. Number of fruits per Low L It was counted all fruits of each plant, including the green and damaged ones Average A High H 06. Soluble solids Low L The analysis was performed by transferring a drop of the fruit juice to the Hanna Instruments model HI 96801 digital refractometer prism and then reading it, expressed in °Brix Average A High H Adv. Hort. Sci., 2020 34(3): 337­347 340 y = nxp dimension matrix (plants) of phenotypic val­ ues of traits; a = is the 1 xn dimension vector of previously estab­ lished economic weights; g = nxp dimension matrix of unknown genetic values of the n traits considered. The vector b = P ­1 Ga, where P ­1 is the inverse of the matrix, of dimension nxn of phenotypic variance and covariance between traits. G is the nxn dimen­ sion matrix of genetic variance and covariance between traits. The expected gain for trait j was expressed by: Where: Ag j(i) = gj (i): expected gain for trait j, with selection based on index I; DS j(i) = selection differential of trait j, with selection based on index I; h2j = heritability of trait j. In the rank summation index of Mulamba and Mock (1978), the orders of each genotype were s u m m e d , r e s u l ti n g i n t h e s e l e c ti o n i n d e x , a s described below: I= r1+r2+....+rn Where: I = index value for a given individual or family; rn = an individual’s rank (or rank) from the j th trait; n = number of traits considered in the index. The weights were given by: L=p1r1+p2r2+....+pnrn Where: pj = economic weight attributed to the jth trait. For the base index of Williams (1962), the follow­ ing index was used as selection criteria: Where: y = are the means; a = are the economic weights of the traits studied. For the index of genotype­ideotype distance (Cruz, 2006), the mean and maximum and minimum values for each variable were calculated. Xij was con­ sidered as the mean phenotypic value of the ith genotype concerning the ith trait. As well, we consid­ ered the value Yij representing the transformed mean phenotypic value and Cj as a constant relative to the average genotype depreciation. Thus, we had: LIj as the lower limit to be presented by the geno­ type, relative to the characteristic j, LSj as the upper limit to be presented by the genotype and VOj as the optimal value to be presented by the genotype, under selection. If LIj LSj, Yij = Xij + VOj ­ LSj + Cj. In the methodology, it was considered Cj = LSj ­ LIj. The Cj value ensured that any value of Xij within the range of variation around the optimum resulted in a value of Yij of magnitude close to the optimal value (VOj), as opposed to the values of Xij outside this range. Thus, the Xij transformation was per­ formed to ensure the depreciation of phenotypic val­ ues out of range. The Yij values obtained by transfor­ mation were later standardized and weighted by the weights assigned to each characteristic, obtaining the Yij values, as described below: Where: S (Yj) = standard deviation of the mean phenotypic values obtained by the transformation; aj = weight or economic value of the characteristic. T h e n , w e c a l c u l a t e d t h e G I D i n d e x v a l u e s expressed by the distances between the genotypes and the ideotype, as illustrated: From these indexes, the best genotypes were identified, and the selection gains were calculated. All genetic and statistical analyzes were processed through the Computational Program in Genetics and Statistics ­ GENES Program (Cruz, 2016). 3. Results and Discussion Genetic variability was found for all traits by the F­ test at 1% or 5% probability level, which evidenced the ability to perform the selection of superior toma­ to lines. It was verified by values of coefficient of vari­ ation (CV) ranging from 1.36% to 29.03% for MC and N F P , r e s p e c ti v e l y . T h e h i g h e s t C V v a l u e s w e r e observed in trait NFP (29.03%), SS (18.28%), and YLD (18.05%) (Table 2). The genotypic coefficient of determination (H2) allows us to define the estimate of genetic gain to be achieved and to establish the most appropriate strat­ egy to be used in the breeding program (Baldissera et al., 2014). H2 values change according to each charac­ Vieira Nascimento et al. ‐ Identification of tomato strains in Brazil 341 teristic and are classified as high when they are high­ er than 0.7 (Alvares et al., 2016). The highest H2 values were found for the matura­ tion cycle (71.93%), fruit firmness (58.30%), and yield (51.48%). These values allow us to reach success by the phenotypic selection, which can be proven by the results found in the CVg/CVe ratio, which were close to 1.0 for these traits. The lowest H2 values were observed for the number of fruits per plant (29.38%) and soluble solids (33.29%). The medium and high results of the heritability coef­ ficient and coefficient of genetic variation are related to higher selective accuracy, higher genetic variability, and the probability of successfully choosing genotypes with optimal agronomic traits (Storck and Ribeiro, 2011). The CVg/CVe ratio was close to 1.0 only for the Table 2 ­ Mean square, coefficient of variation, and genetic parameters of agronomic traits and yield of 85 industrial tomato lines FPT= fruit pericarp thickness, FF= fruit firmness, YLD= yield, MC= maturation cycle, NFP= number of fruits per plant, SS= soluble solids, H2= genotypic coefficient determination, CV= coefficient of variation, CVg= coefficient of genetic variation, CVe= coefficient of experi­ mental variation. ** and * significant by F­test at 1% and 5% probability, respectively. Source of Variation DF Mean square FPT FF YLD CM NFP SS Blocks 2 0.65 0.59 488.22 2.89 1663.12 37.75 Lines 84 2.08 ** 0.56 ** 698.53 ** 8.16 ** 1664.94 * 0.82 * Residue 168 1.35 0.23 338.89 2.29 1175.83 0.55 CV (%) ­ 15.96 16.69 18.05 1.36 29.03 18.28 CVg/CVe ­ 0.42 0.68 0.59 0.92 0.37 0.41 H2 ­ 34.84 58.30 51.48 71.93 29.38 33.29 medium cycle. The CVg/CVe ratio can be accepted as an indicator of the obtaining of more relevant genet­ ic gains in the selection of superior genotypes (Cruz et al., 2012). The constitution of tomato fruits for the industry has been remodeled through genetic improvement, to select cultivars with desirable characteristics for processing. As a general rule, the desired tomato lines are those that combine higher yield with quali­ ty, and that meet the needs of the industry, which currently are firm fruits, with a high content of solu­ ble solids, a shorter cycle, a higher number of fruits per plant and higher fruit pericarp thickness (Iglesias et al., 2015; Peixoto et al., 2017). Fruit pericarp thickness ranged from 5.36 to 9.04 mm (Table 3). Only 3.7% of the tomato lines had a Table 3 ­ Fruit pericarp thickness (FPT), fruit firmness (FF), yield (YLD), maturation cycle (MC), number of fruits per plant (NFP), and solu­ ble solids (SS) of 85 industrial tomato lines Lines Traits FPT mm FF N YLD t ha­1 MC days NFP n° plant­1 SS °Brix PXT­102 5.5 b 1.79 b 103.79 a 109 b 123.67 a 4.17 a PXT­104 6.55 b 1.98 b 93.96 b 107 b 111 b 4.57 a PXT­106 7.59 a 2.39 b 108.19 a 111 a 130.83 a 3.07 b PXT­107 6.11 b 2.31 b 111.50 a 109 b 109 b 4.07 a PXT­108 7.20 a 2.67 b 123.94 a 110 b 96.33 b 4.13 a PXT­109 7.91 a 2.30 b 85.28 b 110 b 92.33 b 3.77 b PXT­111 7.75 a 2.58 b 68.98 b 107 b 74.83 b 4.43 a PXT­113 7.15 a 3.15 a 85.76 b 112 a 99.17 b 4.77 a PXT­114 7.40 a 2.38 b 72.87 b 110 b 112.83 b 3.03 b PXT­115 8.08 a 2.55 b 86.27 b 107 b 106.00 b 4.20 a PXT­116 7.36 a 1.94 b 117.45 a 113 a 107.17 b 3.67 b PXT­117 8.08 a 2.65 b 119.54 a 112 a 130.83 a 4.40 a PXT­118 6.90 b 2.70 b 110.33 a 110 b 141.17 a 3.77 b PXT­120 7.17 a 2.13 b 114.81 a 113 a 128.17 a 4.73 a PXT­121 6.00 b 2.63 b 115.12 a 112 a 135.17 a 3.17 b PXT­122 7.39 a 2.40 b 106.26 a 113 a 129.33 a 4.10 a PXT­123 8.58 a 2.68 b 103.88 a 112 a 143.33 a 4.30 a PXT­124 6.15 b 2.41 b 113.87 a 113 a 77.67 b 3.87 b PXT­125 5.69 b 3.14 a 93.54 b 113 a 159.00 a 4.80 a PXT­126 6.15 b 3.05 a 115.37 a 109 b 155.83 a 4.83 a To be continued... Means followed by the same letters belong to the same group by the Scott­Knott test at 5% probability level. 342 Adv. Hort. Sci., 2020 34(3): 337­347 Table 3 ­ Fruit pericarp thickness (FPT), fruit firmness (FF), yield (YLD), maturation cycle (MC), number of fruits per plant (NFP), and solu­ ble solids (SS) of 85 industrial tomato lines Lines Traits FPT mm FF N YLD t ha­1 MC days NFP n° plant­1 SS °Brix PXT­401 5.37 b 3.08 a 95.68 b 113 a 173.83 a 3.60 b PXT­402 5.90 b 2.74 b 107.19 a 112 a 139.00 a 4.57 a PXT­403 6.37 b 3.33 a 122.89 a 112 a 146.00 a 4.00 b PXT­404 6.74 b 2.11 b 118.68 a 113 a 93.67 b 3.37 b PXT­405 5.50 b 2.88 b 117.25 a 112 a 115.50 b 4.30 a PXT­406 6.37 b 3.30 a 100.62 a 110 b 156.17 a 4.40 a PXT­407 6.43 b 2.85 b 103.61 a 111 a 99.33 b 3.80 b PXT­408 6.01 b 2.62 b 90.30 b 112 a 126.00 a 4.87 a PXT­409 6.03 b 3.22 a 99.00 b 111 a 100.33 b 4.87 a PXT­410 7.42 a 2.80 b 106.20 a 112 a 112.00 b 4.63 a PXT­411 7.81 a 2.46 b 101.26 a 113 a 96.83 b 4.00 b PXT­412 7.58 a 2.65 b 110.89 a 113 a 80.92 b 4.63 a PXT­413 7.95 a 3.12 a 111.25 a 113 a 116.83 b 3.10 b PXT­501 9.04 a 2.85 b 90.56 b 110 b 120.00 a 4.63 a PXT­502 8.44 a 3.35 a 116.72 a 110 b 101.67 b 3.57 b PXT­503 7.72 a 3.35 a 100.34 a 113 a 124.67 a 4.17 a PXT­504 7.10 a 3.48 a 95.05 b 112 a 122.83 a 4.17 a PXT­505 7.44 a 2.84 b 113.92 a 108 b 127.00 a 4.37 a PXT­506 6.79 b 3.06 a 114.68 a 111 a 110.83 b 3.97 b PXT­551 6.66 b 2.97 a 95.56 b 112 a 128.33 a 4.07 a PXT­552 7.16 a 2.68 b 112.40 a 111 a 109.00 b 3.47 b PXT­553 6.92 b 2.99 a 112.46 a 111 a 146.83 a 3.63 b PXT­554 8.69 a 2.67 b 83.15 b 112 a 69.50 b 4.47 a PXT­555 6.96 b 2.65 b 104.99 a 108 b 154.17 a 4.10 a PXT­556 8.49 b 3.06 a 101.46 a 109 b 91.83 b 4.63 a PXT­557 6.92 b 3.08 a 104.24 a 112 a 96.67 b 3.60 b PXT­558 7.53 a 3.69 a 102.49 a 109 b 164.00 a 4.23 a PXT­559 7.31 a 3.77 a 103.30 a 112 a 98.76 b 3.26 b PXT­560 6.88 b 2.50 b 109.27 a 109 b 93.00 b 4.50 a PXT­561 8.22 a 2.70 b 85.59 b 109 b 102.50 b 3.63 b PXT­562 7.73 a 2.88 b 102.51 a 110 b 94.00 b 5.03 a PXT­563 7.22 a 3.05 a 111.74 a 113 a 90.50 b 4.07 a PXT­564 6.95 b 2.50 b 131.20 a 113 a 98.33 b 3.26 b PXT­565 7.12 a 3.07 a 104.29 a 111 a 114.83 b 4.10 a PXT­566 8.44 a 3.38 a 114.83 a 111 a 128.17 a 4.43 a PXT­567 8.35 a 2.84 b 91.36 b 112 a 148.83 a 3.83 b PXT­568 7.23 a 3.42 a 105.30 a 111 a 122.50 a 3.80 b PXT­569 8.06 a 2.87 b 76.93 b 112 a 121.83 a 4.13 a PXT­570 8.02 a 3.34 a 88.84 b 113 a 139.33 a 3.40 b PXT­571 7.23 a 2.80 b 90.29 b 112 a 94.17 b 3.96 b PXT­572 8.14 a 2.76 b 83.28 b 110 b 104.33 b 4.76 a PXT­601 6.4a b 3.31 a 137.92 a 112 a 154.66 a 4.26 a PXT­602 7.59 a 2.76 b 122.97 a 110 b 114.17 b 5.26 a PXT­603 6.19 b 3.08 a 98.68 b 112 a 136.67 a 3.66 b PXT­604 7.60 a 2.98 a 117.62 a 109 b 105.33 b 3.37 b PXT­605 7.82 a 2.94 a 77.11 b 111 a 116.17 b 3.60 b PXT­606 7.67 a 2.82 b 99.15 b 112 a 144.50 a 4.03 a PXT­608 8.82 a 2.79 b 93.17 b 110 b 84.33 b 3.03 b PXT­609 7.29 a 2.78 b 97.48 b 109 b 104.17 b 3.77 b PXT­610 7.69 a 3.08 a 145.98 a 113 a 132.67 a 4.27 a Means followed by the same letters belong to the same group by the Scott­Knott test at 5% probability level. To be continued... Vieira Nascimento et al. ‐ Identification of tomato strains in Brazil 343 high thickness of the pericarp. According to Vieira et al. (2019), the thickness of the pericarp, together with the resistance of the epidermis and the texture of the placenta tissue, influences the firmness of the fruit (the relationship between the volume of the pericarp and volume of the locular material). Only 3.7% of the tomato lines had high values of fruit firmness. Firmer fruits present less degradation of the cell wall and increase the resistance of the fruits during the transport process. The fruit firmness ensures resistance to mechanical damage during mechanized harvesting and bulk transport. Fruits that are not firm are more susceptible to the transforma­ tion and breakage of the skin, releasing cellular juice and causing fermentation and deterioration of the fruits before the arrival in the industry (Vieira et al., 2019). The fruit firmness is extremely important for the industry, because, between the harvest and the unloading process in the industry, there are many losses, due to a large number of disintegrated fruits, r e l a t e d t o e x c e s s i v e c o m p r e s s i o n ( M o u r a a n d Golynski, 2018). One of the main characteristics to be used in the selection of the ideal genotype for the tomato pro­ cessing industry and mainly for the producers is fruit yield. Among the tomato lines evaluated, again, 3.7% of them obtained high values, above 131 t ha­1. The average yield of the state of Goiás, where the tomato lines were evaluated, were 85 and 94 t ha­1 in the 2017 and 2018 harvests, respectively (Globo Rural, 2018). The average cycle ranged from 106 to 113 days. Only 5.88% of tomato lines evaluated had a short cycle. Most cultivars marketed by seed companies have a cycle between 95 and 125 days (Kelley et al., 2010), which demonstrates that all tomato lines eval­ uated are classified between the short and middle cycles. The use of short­cycle genotypes is desirable in breeding programs, as it allows for a shorter stay in the field, where they will be subject for a shorter time to effects of biotic and abiotic factors such as disease and drought stress (Gatut­Wahyu et al., 2014). The number of fruits per plant ranged from 69.50 for PXT­554 to 173.83 for PXT­401. Cultivars with a low number of fruits per plant are not recommended because they have lower yield during the harvesting process (Santos, 2015). High soluble solids content is one of the main characteristics that an industrial tomato material must­have. According to Figueiredo et al. (2015), the higher the soluble solids content, the higher the effi­ ciency of industrial production, and the lower the energy expenditure during the pulp concentration procedure. In practice, for each addition of a °Brix in the pulp, there is a 20% increase in industrial produc­ tion. Values above 4.5°Brix are higher than the Brazilian average. Among 85 tomato lines evaluated, 23.17% is above this value, reaching the maximum value of 5.23 °Brix. Table 3 ­ Fruit pericarp thickness (FPT), fruit firmness (FF), yield (YLD), maturation cycle (MC), number of fruits per plant (NFP), and solu­ ble solids (SS) of 85 industrial tomato lines Lines Traits FPT mm FF N YLD t ha­1 MC days NFP n° plant­1 SS °Brix PXT­611 7.94 a 2.84 b 104.42 a 113 a 110.67 b 4.00 b PXT­613 7.75 a 3.03 a 92.45 b 109 b 106.50 b 4.40 a PXT­614 7.46 a 3.31 a 99.55 b 112 a 107.00 b 3.50 b PXT­615 7.37 a 2.67 b 83.69 b 110 b 79.33 b 3.73 b PXT­616 7.69 a 3.91 a 98.75 b 111 a 122.50 a 4.87 a PXT­617 8.27 a 3.04 a 85.38 b 113 a 92.83 b 3.97 b PXT­618 8.68 a 3.04 a 114.23 a 110 b 96.33 b 4.57 a PXT­619 6.61 b 3.57 a 97.11 b 111 a 152.00 a 4.17 a PXT­651 7.45 a 3.37 a 56.63 b 109 b 125.83 b 3.60 b PXT­652 8.16 a 4.10 a 74.64 b 109 b 157.67 a 3.57 b PXT­653 8.39 a 3.31 a 82.29 b 112 a 103.83 b 2.93 b PXT­654 6.84 b 3.43 a 84.09 b 111 a 166.17 a 3.87 b PXT­655 7.55 a 2.35 b 108.04 a 109 b 124.17 a 4.27 a PXT­656 6.90 b 3.43 a 114.19 a 110 b 149.50 a 3.90 b PXT­687 6.70 b 2.95 a 98.99 b 109 b 113.83 b 4.77 a Means followed by the same letters belong to the same group by the Scott­Knott test at 5% probability level. Adv. Hort. Sci., 2020 34(3): 337­347 344 Direct selection resulted in higher individual gains (Table 4). This selection is directed only for one trait of interest and comprises the obtention of maximum gains of a single trait for which selection is practiced. According to how this trait is associated with others, favorable or unfavorable results may occur in traits of secondary importance (Cruz, 2016). Direct selection for FPT, NFP, and SS resulted in direct gains for fruit firmness, with values of 1.97%, 8.69%, and 2.93%, respectively. Noteworthy was the direct selection for the number of fruits per plant, which resulted in the largest indirect gain for fruit firmness. The indexes of selection consist of an alternative that allows the simultaneous selection to perform effectively by combining different traits (Rosado et al., 2012). In general, the index of the rank summa­ tion index of Mulamba and Mock (1978) showed the largest gain of yield (7.89%) and soluble solids (4.02%), followed by the Smith (1936) and Hazel ( 1 9 4 3 ) i n d e x , w i t h 7 . 2 0 % o f t h e g a i n o f y i e l d . However, these two indexes had low selection gain values for the other traits (Table 5). The rank summation index of Mulamba and Mock (1978) had the highest gain for all the traits and the highest total gain, with values of 22.92%. The geno­ type­ideotype distance index obtained the second­ highest total gain value, with 22.54%. These indices presented a balanced distribution of selection gains. In the research carried out by Rosado et al. (2012), the authors reported that the rank summation index of Mulamba and Mock (1978) was the most appropri­ ate, allowing for a balanced distribution of selection gains for a larger number of yellow passion fruit progenies. The top ten genotypes, selected by all selection methods used in this study and their values of fruit pericarp thickness (Table 6), fruit firmness (Table 7), yield (Table 8), number of fruits per plant (Table 9), and soluble solids (Table 10) are shown in the tables below. The lines PXT­601 and PXT­610 were selected in all selection methods applied, verifying the superi­ ority of these genotypes. 4. Conclusions The rank summation index of Mulamba and Mock (1978) and the classical index proposed by Smith (1936) and Hazel (1943) applied to agronomic traits of eighty­five industrial tomato lines turned out to the largest selection gain for the yield trait. Rank summation index of Mulamba and Mock (1978) has the highest total genetic gain values. The lines of tomato PXT­601 and PXT­610 stand out as superior genotypes by the direct selection method and selection indexes. Acknowledgements To Vivati Plant Breeding Ltda, for support in con­ ducting the project and providing access to the germplasm bank, and to CNPq, for the master’s scholarship granted to the first author. Table 4 ­ Genetic gain estimates obtained for five traits evaluat­ ed by direct and indirect selection for 85 industrial tomato lines Traits Genetic gain (%) FPT FF YLD NFP SS FPT 6.21 0.43 ­0.43 ­3.25 ­1.94 FF 1.97 14.41 ­3.34 8.69 2.93 YLD ­1.41 ­1.51 12.05 ­0.67 ­1.47 NFP ­2.32 5.05 ­0.05 10.26 0.2 SS ­0.12 ­0.23 ­0.44 1.06 6.81 Total 4.33 18.15 7.79 16.09 6.53 FPT= fruit pericarp thickness, FF= fruit firmness, YLD= yield, NFP= number of fruits per plant, SS= soluble solids. Table 5 ­ Genetic gain estimates obtained for five traits by selection by the classical index proposed by Smith (1936) and Hazel (1943), rank summation index of Mulamba and Mock (1978), base index of Williams (1962), and genotype­ideotype distance index for 85 industrial tomato lines FPT= fruit pericarp thickness, FF= fruit firmness, YLD= yield, NFP= number of fruits per plant, SS= soluble solids. Traits Genetic gains (%) Smith (1936) and Hazel (1943) Mulamba and Mock (1978) Williams (1962) Genotype­ ideotype distance FPT ­4.55 1.88 ­2.98 2.31 FF 5.34 5.12 5.83 7.07 YLD 7.20 7.89 6.71 6.69 NFP 7.22 4.01 8.77 4.12 SS 0.29 4.02 0.57 2.35 Total 15.50 22.92 18.90 22.54 Vieira Nascimento et al. ‐ Identification of tomato strains in Brazil 345 Table 6 ­ Fruit pericarp thickness (FPT) in mm from ten superior genotypes selected by direct selection for fruit pericarp thickness, and classic index proposed by Smith (1936) and Hazel (1943), rank summation index of Mulamba and Mock (1978), base index of Williams (1962), and genotype­ideotype distance index (GID) Selection indexes Williams (1962) and direct selection of fruit pericarp thickness Smith (1936) and Hazel (1943) Mulamba and Mock (1978) Genotype­ideotype distance Lines FPT Lines FPT Lines FPT Lines FPT PXT­601 6.44 PXT­601 6.44 PXT­566 8.44 PXT­566 8.44 PXT­610 7.69 PXT­403 6.37 PXT­610 7.69 PXT­558 7.53 PXT­126 6.15 PXT­610 7.69 PXT­558 7.53 PXT­616 7.69 PXT­403 6.37 PXT­126 6.15 PXT­616 7.69 PXT­601 6.40 PXT­558 7.53 PXT­401 5.37 PXT­601 6.44 PXT­117 8.08 PXT­401 5.37 PXT­656 6.90 PXT­117 8.08 PXT­656 6.90 PXT­656 6.90 PXT­405 5.50 PXT­126 6.15 PXT­610 7.69 PXT­555 6.96 PXT­121 6.00 PXT­602 7.59 PXT­123 8.58 PXT­553 6.92 PXT­406 6.37 PXT­618 8.68 PXT­503 7.72 PXT­406 6.37 PXT­619 6.61 PXT­123 8.58 PXT­618 8.68 Table 7 ­ Fruit firmness (FF) in Newton from ten superior genotypes selected by the direct selection for fruit firmness, and classic index proposed by Smith (1936) and Hazel (1943), rank index of Mulamba and Mock (1978), base index of Williams (1962), and genotype­ideotype distance index (GID) Selection Indexes Williams (1962) and direct selection of fruit firmness Smith (1936) and Hazel (1943) Mulamba and Mock (1978) Genotype­ideotype distance Lines FF Lines FF Lines FF Lines FF PXT­601 3.31 PXT­601 3.31 PXT­566 3.38 PXT­566 3.38 PXT­610 3.08 PXT­403 3.33 PXT­610 3.08 PXT­558 3.69 PXT­126 3.05 PXT­610 3.08 PXT­558 3.69 PXT­616 3.91 PXT­403 3.33 PXT­126 3.05 PXT­616 3.91 PXT­601 3.31 PXT­558 3.69 PXT­401 3.08 PXT­601 3.31 PXT­117 2.65 PXT­401 3.08 PXT­656 3.43 PXT­117 2.65 PXT­656 3.43 PXT­656 3.43 PXT­405 2.88 PXT­126 3.05 PXT­610 3.08 PXT­555 2.65 PXT­121 2.63 PXT­602 2.76 PXT­123 2.68 PXT­553 2.99 PXT­406 3.30 PXT­618 3.04 PXT­503 3.35 PXT­406 3.30 PXT­619 3.57 PXT­123 2.68 PXT­618 3.04 Table 8 ­ Yield (YLD), in Mg ha­1, of ten superior genotypes selected by direct selection for yield, and classic index proposed by Smith (1936) and Hazel (1943), rank summation index of Mulamba and Mock (1978), base index of Williams (1962), and genotype­ ideotype distance index (GID) Selection Indexes Williams (1962) and direct selection of yield Smith (1936) and Hazel (1943) Mulamba and Mock (1978) Genotype­ideotype distance Lines YLD Lines YLD Lines YLD Lines YLD PXT­601 137.92 PXT­601 137.92 PXT­566 114.83 PXT­566 114.83 PXT­610 145.98 PXT­403 122.89 PXT­610 145.98 PXT­558 102.49 PXT­126 115.37 PXT­610 145.98 PXT­558 102.49 PXT­616 98.75 PXT­403 122.89 PXT­126 115.37 PXT­616 98.75 PXT­601 137.92 PXT­558 102.49 PXT­401 95.68 PXT­601 137.92 PXT­117 119.54 PXT­401 95.68 PXT­656 114.19 PXT­117 119.54 PXT­656 114.19 PXT­656 114.19 PXT­405 117.25 PXT­126 115.37 PXT­610 145.98 PXT­555 104.99 PXT­121 115.12 PXT­602 122.97 PXT­123 103.88 PXT­553 112.46 PXT­406 100.62 PXT­618 114.23 PXT­503 100.34 PXT­406 100.62 PXT­619 97.11 PXT­123 103.88 PXT­618 114.23 Adv. Hort. Sci., 2020 34(3): 337­347 346 References ALVARES R.C., SILVA F.C., MELO L.C., MELO P.G.S., PEREIRA H.S., 2016 ­ Estimation of genetic parameters and selection of high‐yielding, upright common bean lines with slow seed‐coat darkening. ­ Genet. Mol. Res., 15(4), 1­10. B A L D I S S E R A J . N . C . , V A L E N T I N I G . , C O A N M . M . D . , GUIDOLIN A.F., COIMBRA J.L.M. 2014 ­ Genetics factors related with the inheritance in autogamous plant popu‐ lations ­ Journal of Agroveterinary Sciences, 13: 181­ 189. CRUZ C.D., 2006 ­ Programa GENES: biometria. ­ UFV, Viçosa, Vol. I, pp. 382. CRUZ C.D., 2016 ­ Genes: a software package for analysis in experimental statistics and quantitative genetics. ­ Acta Sci. Agron., 35(3): 271­276. CRUZ C.D., REGAZZI A.J., CARNEIRO P.C.S., 2012 ­ Biometric models applied to genetic improvement. ­ UFV, Viçosa, Vol. IV, pp. 514. DOSSA D., FUCHS F., 2017 ­ Tomate: análise técnico‐econô‐ mica e os principais indicadores da produção nos mer‐ cados mundial, brasileiro e paranaense. ­ Ceasa Paraná, Paraná, Vol. III, pp. 7. FAO, 2006 ­ International standards for phytosanitary measures. ­ FAO, Roma, Vol. I, pp. 251. FIGUEIREDO A.S.T., RESENDE J.T.V., FARIA M.V., PAULA J.T., SCHWARZ K., ZANIN D.S., 2015 ­ Combining ability and heterosis of relevant fruit traits of tomato geno‐ types for industrial processing. ­ Crop Breed. Appl. Table 9 ­ Number of fruits per plant (NFP) of ten superior genotypes selected by direct selection for number of fruits per plant, and clas­ sic index proposed by Smith (1936) and Hazel (1943), rank summation index of Mulamba and Mock (1978), base index of Williams (1962), and genotype­ideotype distance index (GID) Selection Indexes Williams (1962) and direct selection of number of fruits per plant Smith (1936) and Hazel (1943) Mulamba and Mock (1978) Genotype­ideotype distance Lines NFP Lines NFP Lines NFP Lines NFP PXT­601 154.67 PXT­601 154.67 PXT­566 128.17 PXT­566 128.17 PXT­610 132.67 PXT­403 146.00 PXT­610 132.67 PXT­558 164.00 PXT­126 155.83 PXT­610 132.67 PXT­558 164.00 PXT­616 122.50 PXT­403 146.00 PXT­126 155.83 PXT­616 122.50 PXT­601 154.66 PXT­558 164.00 PXT­401 173.83 PXT­601 154.67 PXT­117 130.83 PXT­401 173.83 PXT­656 149.50 PXT­117 130.83 PXT­656 149.50 PXT­656 149.50 PXT­405 115.50 PXT­126 155.83 PXT­610 132.67 PXT­555 154.17 PXT­121 135.17 PXT­602 114.17 PXT­123 143.33 PXT­553 146.83 PXT­406 156.17 PXT­618 96.33 PXT­503 124.67 PXT­406 156.17 PXT­619 152.00 PXT­123 143.33 PXT­618 96.33 Table 10 ­ Soluble solids (SS), in °Brix, from ten superior genotypes selected by direct selection for soluble solids, and classic index pro­ posed by Smith (1936) and Hazel (1943), rank summation index of Mulamba and Mock (1978), base index of Williams (1962), and genotype­ideotype distance index (GID) Selection Indexes Williams (1962) and direct selection of soluble solids Smith (1936) and Hazel (1943) Mulamba and Mock (1978) Genotype­ideotype distance Lines SS Lines SS Lines SS Lines SS PXT­601 4.27 PXT­601 4.27 PXT­566 4.43 PXT­566 4.43 PXT­610 4.27 PXT­403 4.00 PXT­610 4.27 PXT­558 4.23 PXT­126 4.83 PXT­610 4.47 PXT­558 4.23 PXT­616 4.87 PXT­403 4.00 PXT­126 4.83 PXT­616 4.87 PXT­601 4.26 PXT­558 4.23 PXT­401 3.60 PXT­601 4.27 PXT­117 4.40 PXT­401 3.60 PXT­656 3.90 PXT­117 4.40 PXT­656 3.90 PXT­656 3.90 PXT­405 4.30 PXT­126 4.83 PXT­610 4.27 PXT­555 4.10 PXT­121 3.17 PXT­602 5.27 PXT­123 4.30 PXT­553 3.63 PXT­406 4.40 PXT­618 4.57 PXT­503 4.17 PXT­406 4.40 PXT­619 4.17 PXT­123 4.30 PXT­618 4.57 Identification of tomato strains in Brazil 347 Biotechnol., 15(3): 154­161. GATUT­WAHYU A.S., MANGOENDIDJOJO W., YUDONO P., KASNO A., 2014 ­ Mode of inheritance of genes control maturity in soybean. ­ ARPN J. Agric. Biol. Sci., 9(5): 178­182. GERALDINI F., DELEO J.P., JULIÃO L., MARAGON M., BOTEON M. 2018 ­ Anuário 2017‐2018. ­ Hort. Brasil., 16(174): 14­17. GLOBO RURAL., 2018 ­ Safra de tomate industrial bate recorde de toneladas por hectare. ­ Rev. Globo Rural, https://g1.globo.com/economia/agronegocios/globo­ rural/noticia/2018/11/04/safra­de­tomate­industrial­ bate­recorde­de­toneladas­por­hectare.ghtml HAZEL L.N., 1943 ­ The genetic basis for constructing selec‐ tion indexes. ‐ Genetics, 28(6): 476­490. IGLESIAS M.J., GARCÍA­LÓPEZ J., COLLADOS­LUJÁN J.F., LÓPEZ­ORTIZ F., DÍAZ M., TORESANO F., CAMACHO F., 2015 ­ Differential response to environmental and nutritional factors of high‐quality tomato varieties. ­ Food Chem., 176(1): 278­287. KELLEY W.T., BOYHAN G.E., HARRISON K.A., SUMNER P.E., LANGSTON D.B., SPARKS A., HURST W., FONSAH E.G., 2010 ­ Commercial tomato: production handbook. ­ Vol. I, UGA extension, Georgia, pp. 48. MAPA, 2005 ­ Instruções para execução dos ensaios de dis‐ tinguibilidade, homogeneidade e estabilidade de culti‐ vares de tomate (Lycopersicon esculentum Mill.) ­ Ministério da Agricultura Pecuária e Abastecimento, Diário Oficial da União, pp. 1­8. MARCOMINI L., MOLENA L.A., 2018 ­ Tomate: baixa renta‐ bilidade em 2017 limita investimentos no verão de 2018. ­ Hort. Brasil., 16(174): 14­17. MOURA L.E., GOLYNSKI, A., 2018 ­ Critical points of indus‐ trial tomato from field to processing. ­ Hort. Brasil., 36(4): 521­525. MULAMBA N.N., MOCK J.J., 1978 ­ Improvement of yield potential of the Eto Blanco maize (Zea mays L.) popula‐ tion by breeding for plant traits. ­ Egypt. J. Genet. Citol., 7(1): 40­51. NOGUEIRA A.P.O., SEDIYAMA T., SOUSA L.B., HAMAWAKI O.T., CRUZ C.D., PEREIRA D.G., MATSUO E., 2012 ­ Path analysis and correlations among traits in soybean grown in two dates sowing. ­ Biosci. J. 28(6): 877­888. PARMAR N., SINGH K.H., SHARMA D., SINGH L., KUMAR P., NANJUNDAN J., KHAN Y.J., CHAUHAN D.K., THAKUR A.K., 2017 ­ Genetic engineering strategies for biotic and abiotic stress tolerance and quality enhancement in horticultural crops: a comprehensive review. ­ Biotech., 7(4): 239­342. PEIXOTO J.V.M., NETO C.M.S., CAMPOS L.F.C., DOURADO W.S., NOGUEIRA A.P.O, NASCIMENTO A.R., 2017 ­ Indutrial tomato liines: morphological properties and productivity. ­ Genet. Mol. Res., 16(2): 1­15. REZENDE J.C., BOTELHO C.E., OLIVEIRA A.C.B., SILVA F.L., CARVALHO G.R., PEREIRA A.A., 2014 ­ Genetic progress in coffee progenies by different selection criteria. ­ Coffee Sci., 9(3): 347­353. ROSADO L.S.D., SANTOS C.E.M.D, BRUCKNER C.H., NUNES E.S., CRUZ C.D., 2012 ­ Simultaneous selection in proge‐ nies of yellow passion fruit using selection indices. ­ Rev. Ceres, 59(1): 95­101. SANTOS F.F.B., 2015 ­ Selection of tomato breeding lines with resistance to Tomato yellow vein streak virus. ­ Hortic. Bras., 33(3): 345­351. SIMÃO R., RODRIGUEZ T.A., 2008 ­ Evolução da produção do tomate de mesa no estado de Rondônia. ­ Sober., 46(8): 1­8. SMITH H.F., 1936 ­ A discriminant function for plant selec‐ tion. ‐ Annual Eugenics, 7(3): 240­250. STORCK L., RIBEIRO N.D., 2011 ­ Soybean pure lines genet‐ ics values predicted by using the Papadakis method. ­ Bragantia, 70(4): 753­758. VASCONCELOS E.S., FERREIRA R.P., CRUZ C.D., MOREIRA A., RASSINI J.B., FREITAS A.R., 2010 ­ Estimates of genetic gain by different selection criteria in alfalfa genotypes. ­ Rev. Ceres, 57(2): 205­210. VIANNA V.F., DESIDERIOSUP J.A., SANTIAGOSUP S., JUNIORSUP J.A.F., FERRAUDOSUP A.S., 2013 ­ The mul‐ tivariate approach and influence of characters in select‐ ing superior soybean genotypes. ­ Afr. J. Agric. Res., 8(30): 4162­4169. VIEIRA D.A.P., CALIARI M., SOUZA E.R.B., SOARES JÚNIOR M.S., 2019 ­ Mechanical resistance, biometric and physicochemical characteristics of tomato cultivars for industrial processing. ­ Food Sci. Technol., 39(2): 363­ 370. WILLIAMS J.S., 1962 ­ The evaluation of a selection index. ­ Biometrics, 18(3): 375­393.