Identification of soybean genotypes resistant to stink bug complex and key agronomic traits for selection

Identificação de genótipos de soja resistentes ao complexo de percevejos e características agronômicas-chave para seleção

Abstract

The stink bug complex is characterized as an important soybean pest that demands the development of cultivars resistant to the attack of these insects to reduce losses in soybean quality and productivity and, therefore, develop a more sustainable agriculture. This study aimed at selecting soybean genotypes with good agronomic traits, including resistance to stink bug complex and identify the key traits that assist in this selection. 256 genotypes obtained by crossing IAC-100 (resistant) × CD-215 (susceptible) were evaluated under natural field infestation, in the 2014/15 harvest using the16x16alpha-lattice design. The multivariate statistical analysis of principal components and non-hierarchical K-means clustering method tested 11 agronomic traits of interest. The genotypes selected for resistance to stink bugs complex were IAC 100, 2, 18, 22, 25, 38, 48, 69, 71, 73, 110, 121, 132, 133, 149, 150, 171, 178, 181, 202, 214, 217, 219, 241 and 242. The studied traits as yield, good seed weight, leaf retention, and 100 seed weight can be used to select resistant genotypes.

Keywords:
Glycine max; Euschistus heros; Nezara viridula; Piezodorus guildinii

Resumo

O complexo de percevejos é caracterizado como uma importante praga da soja, que demanda o desenvolvimento de cultivares resistentes ao ataque desses insetos, a fim de reduzir perdas na qualidade e na produtividade da cultura, e, consequentemente, promover uma agricultura mais sustentável. Este estudo teve como objetivo selecionar genótipos de soja com boas características agronômicas, incluindo resistência ao complexo de percevejos, bem como identificar os principais aspectos que auxiliam nessa seleção. Foram avaliados 256 genótipos obtidos a partir do cruzamento entre IAC 100, resistente, e CD 215, suscetível, sob infestação natural em campo, na safra 2014/15, utilizando o delineamento alfa-látice 16 × 16. A análise estatística multivariada por componentes principais e o método de agrupamento não hierárquico K-means foram aplicados para avaliar 11 características agronômicas de interesse. Os genótipos selecionados quanto à resistência ao complexo de percevejos foram: IAC 100, 2, 18, 22, 25, 38, 48, 69, 71, 73, 110, 121, 132, 133, 149, 150, 171, 178, 181, 202, 214, 217, 219, 241 e 242. Os atributos/traços/aspectos estudados, como produtividade, bom peso de sementes, retenção foliar e peso de 100 sementes, podem ser utilizados para a seleção de genótipos resistentes.

Palavras-chave:
Glycine max; Euschistus heros; Nezara viridula; Piezodorus guildinii

1. Introduction

Soybean Glycine max (L.) Merr. is considered one of the most important legumes in the world agricultural system because it is an important source of protein and vegetable oil, which can be used as both human and animal food. Brazil is the world's largest producer of soybeans and produced 135.9 million tons in the 2020/2021 agricultural year (Brasil, 2022).

Phytophagous insects, especially those belonging to the family Pentatomidae, including Euschistus heros (F.), Nezara viridula (L.), and Piezodorus guildinii (West.), are considered key pests of soybean cultivation in Brazil, mainly due to the severe crop damage they cause, the high costs involved, and the difficulty of their control (Godoi and Pinheiro, 2009; Musser et al., 2011).

The main losses caused by the attack of these insects occur during the period of seed filling and maturation since the insects insert the stylet in the seeds releasing salivary secretions to facilitate their feeding at this growth stages (Depieri and Panizzi, 2011). The resulting damage to the seed can change their nutritional value and create a gateway for pathogen infection, reducing their quality and vigor (Panizzi and Slansky, 1985; Quirino, 2012), causing losses up to 18% (Bueno et al., 2015).

These bugs are usually controlled by chemical products applied to the plant, which are harmful to the environment and result in higher production costs (Maia et al., 2009). In addition, the reduced number of insecticides registered for their control hinders the chemical management and allows the emergence of resistant insects, making control increasingly more difficult and less efficient (Bueno et al., 2011, 2015), requiring an increasing number of applications and broad-spectrum products (Sosa-Góméz and Silva, 2010).

The use of resistant plants is an alternative to bypass the losses caused by the stink bug complex and also to promote the development of a more sustainable agriculture. In addition, resistant plants have more stable production and production costs, reduced chemical use, do not require transfer of new technologies, and allow integration with other management methods (Pinheiro et al., 2005).

The genotype IAC-100 has been identified in the literature as one of the most resistant to the stink bug complex and defoliant insects in Brazil (Lourenção et al., 2000; McPherson et al., 2007; Rossetto et al., 1995). Rossetto et al. (1995) reported that the resistance of the cultivar IAC-100 results from at least five different resistance mechanisms, such as shorter seed filling period, higher seed number per plant, damaged pod abortion and replacement with new pods, normal senescence with fall of leaves at maturation and resistance to Nematospora coryli transmitted by the stink bugs at feeding time.

Thus, this study aims at selecting soybean genotypes with good agronomic traits and resistant to the stink bug complex while identifying the main traits to aid in breeding selection.

2. Material and Methods

This research was carried out at Jaboticabal county wich is in the North-east Region of the State of São Paulo, at 21°15′22” south latitude and 48°18′58” west longitude, and at an altitude of 595 m. The regional climate is classified as Aw transitioning to Cwa (Koppen, 1948) with 22.2°C annual average temperature and 1451 mm precipitation, and Eutrophic Red Latosol soil with gently undulating relief (Vianna et al., 2013).

Sowing was carried out in conventional tillage using a plotter, at a seeding density of 16 seeds linear meter⁻¹. One day before sowing, the seeds were inoculated with the Gelfix 5 inoculant. The planting fertilization consisted of applying 350 kg ha⁻¹ of NPK (00-20-20) fertilizer according to crop requirements, following soil analysis.

Crop treatments followed the technical recommendations for the culture (EMBRAPA, 2013), except for insect control, which was not performed to allow the natural infestation of the area to evaluate the incidence of the stink bug complex. The infestation evolution was monitored weekly, starting when the first genotypes reached the R4 development stage and ending when they reached maturity (R8). The rag cloth method used to follow the stink bug population as described by Stürmer et al. (2012).

The experimental design was the 16 × 16 α-lattice, with three replicates. The experimental units consisted of four rows 5 m long and 0.5 m apart, with the two central rows leaving out 0.5 m at the borders, totaling 4 m⁻²considered as the useful area.

The genetic material used was a RILs (Recombinant Inbred Lines) population with 251 lines, obtained in the genetics department of the “Luiz de Queiroz” School of Agriculture, from the bi-parental cross between the cultivars IAC-100 (resistant to the stink bug complex) × CD-215 (high agronomic potential), homozygous by the SPD (Single Pod Descendent) method. In addition to the RILs population, the two parents and three controls, BMX-Potencia RR (T1), V-Max RR (T2) and FPS-Urano RR (T3) were part of the experiment, totaling 256 treatments.

The agronomic traits (variables) evaluated were plant height at maturity (APM, cm), measured from the soil surface until the insertion of the last reproductive node; insertion height of the first pod (AIV, cm), measured from the soil surface until the insertion of the first reproductive node, and number of days to maturity (NDM, days), period between the VE and R8 development stages. The agronomic value (VA) was evaluated at maturity using a visual scale to assign scores ranging from 1 to 5(1 plants with no agronomic value to 5 plants with excellent agronomic traits, such as high number of pods, height between 70 and 110 cm, vigorous plants, no lodging, lack of foliar retention and natural threshing, and reduced symptoms of diseases). The foliar retention (RF) was evaluated at maturity using a visual score scale varying from 1 to 5 (1 plants with normal senescence to 5, plants with stems and green leaves making mechanized harvest impracticable). Also, the number of branches per plant (NR) was determined by counting the branches produced per plant; number of pods per plant (NV), pod count per plant; number of reproductive nodes (NNR), count of the plant reproductive nodes; 100-seed weight (PCS, grams) determined by counting and weighing a100-seedsample after correcting to 13% moisture. Finally, good seed weight (PSB, grams) that consists of weighing the seeds that remained after discarding green and poorly shaped seeds, using a spiral that separates the seeds by gravity and centrifugation, standardized to 13% moisture, and grain yield (PG), obtained by weighing the material harvested in the useful plot, converted to Kg ha⁻¹ and corrected to 13% moisture.

After obtaining the data and all the assumptions were met, we estimated the variance components and the genotypic values by the REML/BLUP method. Next, we performed the exploratory multivariate analyzes, principal components, and non-hierarchical cluster analysis. The dissimilarity between the genotypes was measured by the Euclidean distance and the linkage between the groups was determined by the K-means method.

First, the data were standardized so that all variables presented zero mean and unit variance, using the following Equation 1:

$Z_{ij}=\frac{X_{ij}-{\overline{x}}_j}{S_j}$ (1)

where: i = 1, 2, ... n, objects; j = 1, 2, ... p, variables; ${\overline{x}}_j$ and S_j= mean and standard deviation from the j column.

Subsequently, to calculate the principal components, the original variables were decomposed into principal components (eigenvectors) from the eigenvalues of the covariance matrix. The percentage of total variance present in each of the principal components was obtained by the Equation 2:

$C{P}_h=\frac{{\lambda }_h}{Trace\left(C\right)}.100$ (2)

where: C = covariance matrix from the standardized original data; λ_h= h-th characteristic root (eigen value) of matrix C; Trace (C) = λ₁ + λ₂ .....+ λ_h.

Kaiser (1958) proposed that only eigenvalues above one generate components with quality information from the relevant original variables. Thus, from this assumption, only the three principal components with eigenvalues above one are considered.

The correlation between the variables with the principal components was given by the Equation 3:

$r_{xj}\left(C{P}_h\right)=\frac{a_{jh\sqrt{{\lambda }_h}}}{S_j}$ (3)

where: S_j = variable standard deviation j; a_jh = coefficient of variable j in the h-th principal component; λ_h = h-th characteristic root (eigenvalue) of the covariance matrix.

Loads higher than 0.55 were considered since Hair et al. (2009) stated that load values of the components above 0.35 could be accepted in the exploratory analyses.

For non-hierarchical K-means grouping, we first calculated the Euclidean distance from the geometric distance between the objects (genotypes) a and b in a multidimensional space, from Equation 4:

${\text{d}}_{\left(\text{a},\text{b}\right)}=\sqrt{{\displaystyle \sum }_{i=1}^n\left(Xai-Xbi\right)}\begin{array}{c}2\\ \end{array}$ (4)

where: d_{(a, b)}= distance between a and b; n = 1, 2,3 ...i; x_ai=value of variable i for genotype a; x_bi=value of genotype b.

Subsequently, the cluster analysis by the K-means method was performed using the following Equation 5:

$E^{}={\displaystyle \sum }_{k=1}^k{\displaystyle \sum }_{x1\in CK}^{}{d}^{}\left(x_i,x_{ok}{}^{}\right)$ (5)

where: x_ok= cluster centroid; ck e d (x_i x_ok)= distance between x_i and x_ok.

Clustering is a procedure where, from a number of previously defined clusters, the centers of these clusters are calculated seeking to minimize the distance between each point and its respective centroid (Hair et al., 2009).

All multivariate analyses were performed using the Statistica software, version 10 (StatSoft, 2010).

3. Results and Discussion

The threshold for economic damage set at 2 stink bugs 1 linear m⁻¹ was reached after February 3, 2015 (Figure 1). Most of the genotypes had reached the R5 development stage, being more susceptible to the disease and, consequently, when the greatest losses in seed quality were expected to occur due to the infestation (Corrêa-Ferreira et al., 2013). The highest occurrences were verified for the species E. heros (70%), followed by P. guildinii (23%) and N. viridula (4%), and 3% others. Similarly, Bueno et al. (2011, 2015) reported E. heros population of 99% and 90%, respectively. Although the P. guildinii population was smaller, it is noteworthy that this species damage potential is higher than that of E. heros and N. viridula (Corrêa-Ferreira and Azevedo, 2002), and even small populations can cause great losses.

Figure 1
Fluctuation of the soybean stink bug (Euschistus heros, Nezara viridula, and Piezodorus guildinii) population infesting soybean genotypes obtained from the crossbreeding ofIAC-100 × CD-215 between the R4 and R8 growth stages, in the 2014/15 harvest.

The correlation between the traits of agronomic interest, evaluated under the attack of the stink bug complex, is shown in Table 1. The traits PG (−0.57), PSB (−0.75), VA (−0.88), NDM (−0.57), APM (−0.79), AIV (−0.68) and NNR (−0.83) are directly correlated and inversely correlated with the RF trait (0.63). Therefore, it can be stated that as RF increases, PG, PSB, VA, NDM, APM, AIV and NNR decrease, and vice versa. Likewise, as both NV (−0.60) and NR (−0.64) decrease, PCS (0.61) increases. Hair et al. (2009) reported that traits with the same sign are positively correlated, while those with opposing signs are negatively correlated. These results corroborate partially those found by Vianna et al. (2013) and Dallastra et al. (2014), who reported a direct correlation between the traits PG, NR, and NV, which, in turn, correlated inversely with the APM and AIV traits.

Table 1
Correlation coefficients between traits and the three principal components that retained the largest amount of relevant information.

Traits	PC1	PC2	PC3
PG	−0.57	0.48	−0.46
PSB	−0.75	0.35	−0.37
VA	−0.88	−0.22	0.09
NDM	−0.57	0.49	−0.26
APM	−0.79	0.09	0.14
AIV	−0.68	0.12	0.32
NNR	−0.83	−0.17	0.00
RF	0,63	0,51	−0,11
PCS	0.48	0.61	−0.12
NV	−0.29	−0.60	−0.62
NR	0.50	−0.32	−0.64

• PG = grain productivity; PSB = good seed weight; VA = agronomic value; NDM = number of days to maturity; APM = plant height at maturity; AIV = insertion height of the first pod; RF = foliar retention; NNR= number of reproductive nodes; PCS = 100-seed weight; NV = number of pods; NR = number of branches; PC = principal component.

The direct correlation between PG and PSB and inverse with RF allows the indirect selection for resistance to the stink bug complex since Rocha et al. (2014) reported that it is possible to identify the resistant genotypes from the PG and PSB traits in environments with high population of the stink bug complex.

The first three principal components (PC) explained 71.59% of the variance retained in the 11 original traits since only the first three components had eigenvalues ​​higher than one (Kaiser, 1958). The first principal component (PC1) retained 42.91% of all the original variance (Figure 2). The traits VA, NNR, APM, PSB, AIV, RF, NDM and PG explained most of the variance retained in this component (Table 1). PC2 retained 16.16% of the remaining original variance, which was mainly explained by the PCS and NV traits, whereas PC3 retained 12.52%, explained by the NR and NV traits. According to Ferraudo (2010), the most important traits are those with the highest correlations, regardless of sign. In addition, Valladares et al. (2008) stated that the principal component analysis is useful for clustering individuals with similar traits and studying their correlations.

Figure 2
Biplot showing the dispersion of 256 soybean genotypes as a function of the principal components PC1 × PC2, with the projection of the vectors of the agronomic traits: PG = grain productivity; PSB = good seed weight; VA = agronomic value; NDM = number of days to maturity; APM = plant height at maturity; AIV = height insertion of the first pod; RF = foliar retention; NNR = number of reproductive nodes; PCS = 100-seed weight; NV = number of pods; NR = number of branches.

The two-dimensional plane formed by the components PC1 and PC2 retained 59.07% of the original variance (Figure 2) and allowed identifying the genotypes 202, 22, 48, 150, 110, 121 and 241 as having specific properties for the PG, PSB and NDM traits being, therefore, the most productive, performing better for good seed weight, however, it also had a longer cycle. Similarly, genotypes 2, 71, 69, 133, 25, 132, IAC-100 (stink bug-resistant genitor), 18, 178, 73, 181, 219, 217, 214, 171, 38, 149 and 242 had the highest values for the traits AIV, APM, NNR and VA, being considered resistant to the attack of the stink bug complex, since they are close to the genotype IAC-100 (resistant genitor). In addition, due the cluster of genotypes 202, 22, 48, 150, 110, 121 and 241 are close to the IAC-100 genotype group, it is possible to conclude that these two clusters share similar traits and, therefore, may present resistance to the stink bug complex. However, further and more detailed studies are needed to confirm this hypothesis.

Another fact that allows hypothesizing that the aforementioned genotypes present resistance to the stink bug complex is that the CD-215 (susceptible) genotype is located in the quadrant opposite the IAC-100 genotype (Figure 2). Similar to the CD-215 genotype, the 135, 246, 43, 207, 106, 106, 33, 16, 206, 208, 166, 12, 154 and 144 genotypes displayed specific characteristics for RF, an undesirable trait when aiming at resistance to the stink bug complex. Likewise, these genotypes also presented specific characteristics for PCS. The strong relationship between RF and PCS allows stating that as the plant produces larger seeds, the foliar retention also increases and, consequently, these genotypes are more susceptible to the attack of the stink bug complex. On the other hand, plants that produce smaller seeds such as the IAC-100 genotype have lower PCS and RF and increased resistance to the stink bug complex, but Rossetto et al. (1995) characterized the smaller seeds produced by the IAC-100 genotype and higher NV as a pseudo-resistance since the damage caused by the insect is diluted. However, these authors reported that the IAC-100ability to replace damaged pods with new pods characterizes a tolerant typeresistance, and due to its genetic composition, the plant is also able to withstand the attack of the stink bug complex without compromising productivity and seed quality.

The NNR and NV yield components had a direct and low magnitude correlation with PG, whereas NR was inversely correlated with PG (Figure 2). On the other hand, Vianna et al. (2013) and Dallastra et al. (2014) reported a direct and high magnitude correlation between yield components and PG. In addition, Alcantara Neto et al. (2011) reported that NV and NR contributed effectively to the increase of PG since they are direct production components. The inverse correlation between PG and NR in this work can be explained by the fact that the NR, although weakly, was directly correlated with RF and PCS. Thus, the genotypes most susceptible to the disease also had high NR and were the most attacked by the insects so that the number of damaged seeds (shriveled and malformed) was higher, which reduced PG and PSB.

The two-dimensional plane formed by the PC1 and PC3 components retained 55.43% of the original variance (Figure 3), allowing to identify the 25, 132, IAC-100, 69, 133, 18, 181, 168, 73, 171, 22, 146, and 217 genotypes as having specific characteristics for NNR, VA, APM and AIV traits, as well as resistance to the stink bug complex since they clustered with the IAC-100 cluster genotypes. In addition, the genotypes 202, 71, 2, 85, 142, 178, 167, 149, 219, 93, 38 and 214 also showed specific properties for the PG, PSB and NDM and were considered as the most productive as well as possibly resistant to the stink bug complex due to the proximity to the previous cluster. On the other hand, genotypes 33, 189, 21, 203, 24, 207, 246, 97, 67, 225, 23, 135, 122, 154 and 144 were the most susceptible to the attack of the stink bug complex, being inferior to CD-215 with specific characteristics for RF and PCS. The three controls used did not have specific characteristics and the variables were ​​close to the mean population values.

Figure 3
Biplot showing the dispersion of the 256 soybean genotypes as a function of the principal components PC1 × PC3, with the projection of the trait vectors: PG = grain productivity; PSB = good seed weight; VA = agronomic value; NDM = number of days to maturity; APM = plant height at maturity; AIV = insertion height of the first pod; RF = foliar retention; NNR = number of reproductive nodes; PCS = 100-seed weight; NV = number of pods; NR = number of branches.

The K-means clustering analysis (Figure 4) grouped in cluster eight the genotypes with the best mean PG and PCS, and small RF and PCS, extremely important traits regarding resistance to the stink bug complex. Cluster 9 with the IAC-100 genotype had the lowest values ​​of RF and PCS and the highest NDM, APM, AIV, NNR, VA, and NV. The data in these clusters corroborates the principal component analysis since most of the genotypes in these clusters had the same traits in both analyses. Therefore, the genotypes in clusters eight and nine can be classified as resistant to the stink bug complex.

Figure 4
Distribution profile of the centroids of the clusters formed by the K-means clustering analysis based on the agronomic traits: PG = grain productivity; PSB = good seed weight; VA = agronomic value; NDM = number of days to maturity; APM = plant height at maturity; AIV = insertion height of the first pod; RF = foliar retention; NNR = number of reproductive nodes; PCS = 100-seed weight; NV = number of pods; NR = number of branches.

On the other hand, the genotypes in clusters three, including CD-215, and six were susceptible to the disease as indicated by the principal component analysis. This result was confirmed by the K-means clustering analysis since the genotypes in these clusters had the highest RF and PCS and lowest VA, while the remaining traits ​​were close to the mean.

4. Conclusions

Genotypes 22, 48, 110, 121, 150, 202 and 241 had the best performance for grain yield and good seed weight but also had the longest cycle. These genotypes can be considered resistant to the stink bug complex; The genotypes IAC-100, 2, 18, 25, 38, 69, 71, 73, 132, 133, 149, 171, 178, 181, 214, 217, 219 and 242 have specific characteristics for the AIV, APM, NNR, and VA traits, showing resistance to the stink bug complex; The traits PSB, PG, RF, and PCS can be used for selecting soybean genotypes resistant to the stink bug complex; with good response in the selection process in soybean breeding programs. In conclusion, further and more detailed studies are required to confirm the resistance of the selected genotypes since the principal component and clustering analysis by the K-means method are exploratory analyses.

Data Availability Statement

The datasets generated and analyzed during the current study are not publicly available but are available from the corresponding author upon reasonable request.

References

• ALCANTARA NETO, F., GRAVINA, G.A., MONTEIRO, M.M.S., MORAIS, F.B., PETTER, F.A. and ALBUQUERQUE, J.A.A., 2011. Path analysis of soybean yield in the micro region of Alto Médio Gurguéia. Comunicata Scientiae, vol. 2, pp. 107-112.
• BRASIL. Companhia Nacional de Abastecimento – CONAB, 2022 [viewed 10 February 2022]. 12° levantamento da safra de grãos 2020/21 [online]. Brasília. Available from: http://www.conab.gov.br/conteudos.php?a=1252&
  » http://www.conab.gov.br/conteudos.php?a=1252&
• BUENO, A., BATISTELA, M.J., BUENO, R.C.O.F., FRANÇA-NETO, J.B., NAIME NISHIKAWA, M.A. and LIBÉRIO FILHO, A., 2011. Effects of integrated pest management, biological control and prophylactic use of insecticides on the management and sustainability of soybean. Crop Protection, vol. 30, no. 7, pp. 937-945. https://doi.org/10.1016/j.cropro.2011.02.021
  » https://doi.org/10.1016/j.cropro.2011.02.021
• BUENO, A.F., BORTOLOTTO, O.C., POMARI-FERNANDES, A. and FRANÇA-NETO, J.B., 2015. Assessment of a more conservative stink bug economic threshold for managing stink bugs in Brasilian soybean production. Crop Protection, vol. 71, pp. 132-137. https://doi.org/10.1016/j.cropro.2015.02.012
  » https://doi.org/10.1016/j.cropro.2015.02.012
• CORRÊA-FERREIRA, B.S. and AZEVEDO, J., 2002. Soybean seed damage by different species of stink bugs. Agricultural and Forest Entomology, vol. 4, no. 2, pp. 145-150. https://doi.org/10.1046/j.1461-9563.2002.00136.x
  » https://doi.org/10.1046/j.1461-9563.2002.00136.x
• CORRÊA-FERREIRA, B.S., CASTRO, L.C., ROGGIA, S., CESCONETTO, N.L., COSTA, J.M. and OLIVEIRA, M.C.N., 2013 [viewed 10 February 2022]. MIP-soja: resultados de uma tecnologia eficiente e sustentável no manejo de percevejos no atual sistema produtivo de soja [online]. Londrina: Embrapa Soja, 55 p. Embrapa Soja. Documentos, no. 341. Available from: http://ainfo.cnptia.embrapa.br/digital/bitstream/item/87596/1/Doc-341.pdf
  » http://ainfo.cnptia.embrapa.br/digital/bitstream/item/87596/1/Doc-341.pdf
• DALLASTRA, A., UNÊDA-TREVISOLI, S.H., FERRAUDO, A.S. and DI MAURO, A.O., 2014. Multivariate approach in the selection of superior soybean progeny which carry the RR gene. Revista Ciência Agronômica, vol. 45, no. 3, pp. 588-597. https://doi.org/10.1590/S1806-66902014000300021
  » https://doi.org/10.1590/S1806-66902014000300021
• DEPIERI, R.A. and PANIZZI, A.R., 2011. Duration of feeding and superficial and In-depth damage to soybean seed by selected species of stink bugs (Heteroptera: Pentatomidae). Neotropical Entomology, vol. 40, no. 2, pp. 197-203. https://doi.org/10.1590/S1519-566X2011000200007 PMid:21584400.
  » https://doi.org/10.1590/S1519-566X2011000200007
• EMPRESA BRASILEIRA DE PESQUISA AGROPECUÁRIA – EMBRAPA, 2013. Tecnologias de produção de soja Londrina: Embrapa Soja, 265 p.
• FERRAUDO, A.S., 2010. Técnicas de análise multivariada. São Caetano: Statsoft South América.
• GODOI, C.R.C. and PINHEIRO, J.B., 2009. Genetic parameters and selection strategies of soybean genotypes resistant to stink bug complex. Genetics and Molecular Biology, vol. 32, no. 2, pp. 328-336. https://doi.org/10.1590/S1415-47572009000200020 PMid:21637688.
  » https://doi.org/10.1590/S1415-47572009000200020
• HAIR, J.F., BLACK, W.C., BABIN, B.J., ANDERSON, R.E. and TATHAM, R.L., 2009. Análise multivariada de dados 6. ed. Porto Alegre: Bookman, 688 p.
• KAISER, H.F., 1958. The varimax criterion for analytic rotation in fator analysis. Psychometrika, vol. 23, no. 3, pp. 187-200. https://doi.org/10.1007/BF02289233
  » https://doi.org/10.1007/BF02289233
• KÖPPEN, W., 1948. Climatologia México: Fondo de Cultura Econômica, 478 p.
• LOURENÇÃO, A.L., PEREIRA, C.V.N.A., MIRANDA, M.A.C. and AMBROSANO, M., 2000. Avaliação de danos causados por percevejos e por lagartas em genótipos de soja de ciclos precoce e semiprecoce. Pesquisa Agropecuária Brasileira, vol. 35, no. 5, pp. 879-886. https://doi.org/10.1590/S0100-204X2000000500003
  » https://doi.org/10.1590/S0100-204X2000000500003
• MAIA, M.C.C., VELLO, N.A., ROCHA, M.M., FONSECA JUNIOR, N.S., LAVORANTE, O.J., PINHEIRO, J.B., DIAS, C.T.S. and ASSIS, G.M.L., 2009. Seleção de linhagens experimentais de soja para características agronômicas e tolerância a insetos. Bragantia, vol. 68, no. 1, pp. 85-97. https://doi.org/10.1590/S0006-87052009000100010
  » https://doi.org/10.1590/S0006-87052009000100010
• MCPHERSON, R.M., BUSS, G.R. and ROBERTS, P.M., 2007. Assessing stink bug resistance in soybean breeding lines containing genes from germplasm IAC-100. Journal of Economic Entomology, vol. 100, no. 4, pp. 1456-1463. https://doi.org/10.1093/jee/100.4.1456 PMid:17849902.
  » https://doi.org/10.1093/jee/100.4.1456
• MUSSER, F.R., CATCHOT, A.L., GIBSON, B.K. and KNIGHTEN, K.S., 2011. Economic injury levels for southern green stink bugs (Hemiptera: Pentatomidae) in R7 growth stage soybeans. Crop Protection, vol. 30, no. 1, pp. 63-69. https://doi.org/10.1016/j.cropro.2010.09.011
  » https://doi.org/10.1016/j.cropro.2010.09.011
• PANIZZI, A.R. and SLANSKY, F., 1985. Review of phytophagous pentatomids (Hemiptera: Pentatomidae) associated with soybean in the Americas. The Florida Entomologist, vol. 68, no. 1, pp. 184-214. https://doi.org/10.2307/3494344
  » https://doi.org/10.2307/3494344
• PINHEIRO, J.B., VELLO, N.A., ROSSETTO, C.J. and ZUCCHI, M.I., 2005. Potential of soybean genotypes as insect resistance sources. Crop Breeding and Applied Biotechnology, vol. 5, no. 3, pp. 294-301. https://doi.org/10.12702/1984-7033.v05n03a06
  » https://doi.org/10.12702/1984-7033.v05n03a06
• QUIRINO, R.J., 2012 [viewed 10 February 2022]. Consequências na armazenagem da soja devido ao ataque de percevejos nas lavouras de soja. In: Anais do X Congresso Brasileiro de Soja [online], 2012, Cuiabá. Londrina: Embrapa Soja. Available from: http://www.cbsoja.com.br/anais/palestras/JoseRonaldoQuirino.pdf
  » http://www.cbsoja.com.br/anais/palestras/JoseRonaldoQuirino.pdf
• ROCHA, F., BERMUDEZ, F., FERREIRA, M.C., OLIVEIRA, K.C. and PINHEIRO, J.B., 2014. Effectiveselectioncriteria for assessing the resistance of stink bugs complex in soybean. Crop Breeding and Applied Biotechnology, vol. 14, no. 3, pp. 174-179. https://doi.org/10.1590/1984-70332014v14n3a26
  » https://doi.org/10.1590/1984-70332014v14n3a26
• ROSSETTO, C.J., GALLO, P.B., RAZERA, L.F., BORTOLETTO, N., IGUE, T., MEDINA, P.F., TISSELI FILHO, O., AQUILERA, V., VEIGA, R.F.A. and PINHEIRO, J.B., 1995. Mechanisms of resistance to stink bug complex in the soybean cultivar IAC-100. Anais da Sociedade Entomológica do Brasil, vol. 24, pp. 517-522. https://doi.org/10.37486/0301-8059.v24i3.1061
  » https://doi.org/10.37486/0301-8059.v24i3.1061
• SOSA-GÓMÉZ, D.R. and SILVA, J.J., 2010. Neotropical brown stink bug (Euschistus heros) resistance to methamidophos in Paraná, Brasil. Pesquisa Agropecuária Brasileira, vol. 45, no. 7, pp. 767-769. https://doi.org/10.1590/S0100-204X2010000700019
  » https://doi.org/10.1590/S0100-204X2010000700019
• STATSOFT, 2010. Statistica (data analysis software system) version 10.0. Tulsa: StatSoft, Inc.
• STÜRMER, G.R., CARGNELUTTI FILHO, A., STEFANELO, L. and GUEDES, J.V.C., 2012. Eficiência de métodos de amostragem de lagartas e de percevejos na cultura da soja. Ciência Rural, vol. 42, no. 12, pp. 2105-2111. https://doi.org/10.1590/S0103-84782012005000145
  » https://doi.org/10.1590/S0103-84782012005000145
• VALLADARES, G.S., GOMES, E.G., MELLO, C.C.B.S., PEREIRA, M.G., ANJOS, H.C., EBELING, G. and BENITES, V.M., 2008. Análise de componentes principais e métodos multicritério ordinais no estudo de organossolos e solos afins. Revista Brasileira de Ciência do Solo, vol. 32, no. 1, pp. 285-296. https://doi.org/10.1590/S0100-06832008000100027
  » https://doi.org/10.1590/S0100-06832008000100027
• VIANNA, V.F., UNÊDA-TREVISOLI, S.H., DESIDÉRIO, J.A., SANTIAGO, S., CHARNAI, K., FERREIRA JUNIOR, J.A., FERRAUDO, A.S. and DI MAURO, A.O., 2013. The multivariate approach and influence of characters in selecting superior soybean genotypes. African Journal of Agricultural Research, vol. 8, pp. 4162-4169.

Edited by

• Editor:
  Takako Matsumura Tundisi

Publication Dates

• Publication in this collection
  10 Apr 2026
• Date of issue
  2026

History

• Received
  04 Nov 2025
• Accepted
  10 Dec 2025