Next Article in Journal
A Mycorrhiza-Associated Receptor-like Kinase Regulates Disease Resistance in Rice
Previous Article in Journal
Effect of Processing Solid Organic Municipal Wastes on Their Phosphorus Fertilizer Value
Previous Article in Special Issue
Visual Observation of Polystyrene Microplastics/Nanoplastics in Peanut Seedlings and Their Effects on Growth and the Antioxidant Defense System
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Agronomy
Volume 15
Issue 10
10.3390/agronomy15102297
agronomy-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

An Unsupervised and Supervised Machine Learning Approach to Evidence Tetranychus mexicanus (McGregor) Activity in Fluorescence and Thermal Response in Passion Fruit

by
Maria Alaíne da Cunha Lima
,
Eleazar Botta Ferret
,
Magaly Morgana Lopes da Costa
,
Mariana Tamires da Silva
,
Roberto Ítalo Lima da Silva
,
Shirley Santos Monteiro
,
Manoel Bandeira de Albuquerque
and
José Bruno Malaquias
*
Crop Production and Environmental Science Department, Federal University of Paraiba, Areia 58397000, Paraiba, Brazil
*
Author to whom correspondence should be addressed.
Agronomy 2025, 15(10), 2297; https://doi.org/10.3390/agronomy15102297
Submission received: 19 August 2025 / Revised: 19 September 2025 / Accepted: 26 September 2025 / Published: 28 September 2025
(This article belongs to the Collection Crop Physiology and Stress)
Browse Figures
Versions Notes

Abstract

Tetranychus mexicanus (McGregor, 1950) (Tetranychidae) is considered one of the primary phytosanitary problems in passion fruit crops, resulting in significant production losses. Understanding the impact of this mite species’ activity on the physiology of passion fruit plants can serve as a basis for developing sustainable management strategies. With this in mind, this research sought to analyze, using supervised and unsupervised machine learning models, how T. mexicanus mite infestation influences gas exchange, chlorophyll “a” and chlorophyll “b” levels, fluorescence, and thermal response of passion fruit plants. We tested the hypothesis that juvenile and adult mites alter the physiological and thermal response patterns of plants. Only the variables related to the fluorescent response (Fo, Fm, and Fv) had a significant relationship with mite infestation. In the joint comparison of multiple fluorescent variables, there were differences between the treatments of plants infested and not infested by T. mexicanus. The variables’ initial fluorescence (Fo), maximum fluorescence (Fm), and variable fluorescence (Fv) of chlorophyll a had a direct negative impact on both reproductive activity, as measured by the number of eggs and nymphs produced, and the total number of mites found. The unsupervised model based on multidimensional scaling with the k-means algorithm revealed a clear separation between the groups of infested passion fruit plants (Group 1) and healthy plants (Group 2). The Fo response was described with high accuracy for the reproductive rate (75%) and total infestation of eggs, nymphs, and adults of the mites (99.99%). Kappa values were moderate (Kappa = 0.50) and high (Kappa = 0.99) for reproductive and total rates of T. mexicanus, respectively. Additionally, the thermal response revealed that the infested passion fruit plants had a median temperature of 25.1 °C, compared to a median temperature of 25.7 °C, with notable differences between these medians. Therefore, the T. mexicanus mite altered both the fluorescent and thermal patterns of passion fruit plants. Our findings have implications for the development of early detection tools and the generation of future resistance breeding.

1. Introduction

Passion fruit cultivation plays a crucial role in generating income for medium- and small-scale farmers, as it is favored by its high market value and shorter production cycle compared to other tropical fruits [1,2]. In semiarid regions, soil and climate conditions can be favorable for cultivation. However, poor rainfall distribution and high evapotranspiration often limit productivity, promoting abiotic stresses that compromise plant development [3]. These stresses, combined with biotic stresses and attacks of mites, trigger complex physiological responses, including changes in gene expression, cellular metabolism, growth rates, and photosynthetic processes. Biotic stress, caused by living organisms, deprives the host of essential nutrients, directly affecting photosynthesis and potentially leading to plant death [4,5]. Due to their wide range of host plants and their occurrence in different ecosystems, phytophagous mites pose a significant threat to global fruit production [6].
Mites of the species Tetranychus mexicanus (McGregor, 1950) (Tetranychidae) reduce productivity and can act as vectors of various diseases in passion fruits [7], causing severe losses to production [8,9]. When present on leaves, they generally pierce plant cells to feed on internal fluids, promoting necrotic or chlorotic spots, retarding growth, and compromising productivity [10]. Evidence from other crops reinforces the potential impact of these pests on plant physiology. In coffee plants, there is a reduction of more than 50% in the photosynthetic rate of leaves with high levels of infestation [11]. In Phaseolus vulgaris, mite infestation results in a significant decrease in leaf chlorophyll content compared to control plants [12]. Meanwhile, in Arabidopsis thaliana, T. urticae infestation leads to a reduction in instantaneous fluorescence and photosynthetic quantum yield, indicating damage to the photosystems [13]. Despite recent advances in research on plant–mite interactions, there is still a scarcity of studies investigating the physiological impacts of these infestations on passion fruit. Most of the literature focuses on the taxonomy or control of these arthropods, without evaluating their direct effects on leaves and photosynthetic yield.
Machine learning algorithms have been increasingly applied to predictions in agricultural science, demonstrating high performance in classifying arthropod infestations [14,15]. Differentiating plants with arthropod damage from healthy crops and predicting future physiological damage in these plants are essential parts of plant health monitoring. These tools enable high-resolution forecasting of pest dynamics, which is crucial for timely interventions in Integrated Pest Management (IPM) strategies [16,17]. Despite recent advances in research on plant–mite interactions, there is still a scarcity of studies investigating the physiological impacts of these infestations on passion fruit plants. Most of the literature focuses on the taxonomy or control of these arthropods, without evaluating their direct effects on leaves and photosynthetic yield.
Therefore, understanding how phytophagous mite infestations affect the physiological processes of passion fruit plants is essential to developing more effective and sustainable management strategies. Based on this context, the main objective of this research was to analyze using supervised and unsupervised machine learning models how T. mexicanus infestation influences the gas exchange, chlorophyll “a” and chlorophyll “b” levels, fluorescence, and thermal response of passion fruit plants. The central hypothesis is that mites, in their juvenile and adult stages, alter the physiological and thermal response patterns of passion fruit plants.

2. Materials and Methods

2.1. Place of Experiment

The experiment was carried out in a protected environment, between July and August 2025, at the Department of Plant Science and Environmental Sciences, Center of Agricultural Sciences of the Federal University of Paraíba (UFPB), Campus II, in Areia, Paraíba, Brazil. The geographic coordinates of the site are 06°58′59″ S and 35°42′57″ W, at an elevation of 503 m. According to the Köppen classification, the local climate is classified as type ‘As’, characterized by dry and hot summer periods and rainfall in winter [18].

2.2. Experimental Design and Experiment Conduction

Using a randomized block design with four replications, 200 seedlings were equally distributed among cages (25 seedlings per cage, per replication). The cage has the following dimensions: 2.5 meters high, 2 meters wide, and 2 meters long. A voile tissue covered the cages. Treatments consisted of mite-infested plants and healthy plants (control). The seedlings were grown in 3 L polyethylene bags. Seeds of the local yellow passion fruit variety (Passiflora edulis Sims f. flavicarpa D.), grown in the municipality of Nova Floresta, Paraíba, and popularly known as “Guinezinho,” were purchased from a commercial orchard. The substrate used was the commercial Mecplant® (Telêmaco Borba, PR, Brazil), composed of 60% pine bark, 15% fine vermiculite, 15% superfine vermiculite, and 10% humus.
The plants, 60 days after germination, were taken to the Entomology Laboratory, where they were selected for uniform size and the absence of insect pests and diseases. We inspected the plants following a prior assessment to ensure the absence of insect pests and diseases. They remained in a protected environment for 15 days, housed in isolated cages to prevent contact between treatments and to prevent the entry of other external organisms. The infestation was carried out with organisms from passion fruit cultivation areas with a high incidence of mites, located in the municipality of Areia, Paraíba. The transfer was carried out manually, using brush number 7/0, and directed to the aerial part of the plants, specifically the first fully expanded leaf. The mite-infested plants received twenty female individuals per plant. The female mites used were age-synchronized at 2–5 days old. The mites were kept in these 15 days in a state of feeding and reproduction activity without any interference, except for physiological assessments conducted every 5 days.

2.3. Physiological Assessments

Measurements were performed 15 days after infestation, on five plants per treatment. These 15 days are justified based on the sufficient time required to achieve differences in plant physiology. We analyzed the response every 5 days, but no difference was observed. We assessed five plants per treatment, selecting the third leaf from top to bottom.

2.3.1. Gas Exchange

Gas exchanges were measured in the third fully expanded leaf, using an infrared gas analyzer—IRGA (LI-6400XT LI-COR®, Nebraska, NE, USA), to determine the variables of photosynthesis (A) (µmol m-2 s-1), stomatal conductance (gs) (mol m-2 s-1), transpiration (E) (mmol m-2 s-1) and internal carbon concentration (Ci) (µmol mol-1). From these parameters will be calculated, instantaneous Water Use Efficiency (WUE, A/E) (µmol m-2 s-1) (mmol m-2 s-1)-1, intrinsic water use efficiency (A/gs) (µmol m-2 s-1) (mol m-2 s-1)-1, and instantaneous carboxylation efficiency (A/Ci) (µmol m-2 s-1) (µmol mol-1)-1 [19]. With an airflow of 300 mL/min and a coupled light source of 1200 mol/m2/s, measurements were taken from 8:00 a.m. to 11:00 a.m.

2.3.2. Leaf Chlorophyll Index and Chlorophyll “a” Fluorescence

Leaf chlorophyll a/chlorophyll b, chlorophyll b/a ratio, and total chlorophyll were measured in the middle third of each plant using a portable digital chlorophyll meter (ClorofiLOG, FalKer®, model CFL 1030, Porto Alegre, RS, Brazil). The values obtained refer to the product of photodiodes emitting at wavelengths of 635, 660, and 880 nm [19,20].
Chlorophyll fluorescence was measured on the third leaf from the apex of each plant, after dark pre-adaptation for 30 min, using a modulated fluorometer (OptiSciences Inc. model OS-30p, Hudson, NY, USA). The initial fluorescence (Fo), maximum fluorescence (Fm), and variable fluorescence (Fv) of chlorophyll a, quantum efficiency of photosystem II (Fv/Fm), and the ratio between quantum yields of competing photochemical and non-photochemical products in photosystem II (Fv/Fo) were determined.

2.4. Leaf Temperature

To record leaf surface temperature, thermographic images were obtained 15 days after infestation, during the period of physiological evaluations (7:00 a.m. to 11:00 a.m.), to evaluate variations in leaf temperature associated with biotic stress, using a T2 Pro Xinfrared infrared thermographic camera (Xinfrared, Hong Kong, China). The camera has the following functionalities: (1) thermal image observation through the thermal camera; (2) thermal temperature measurement and analysis; (3) photographing; video recording, and reticle; and (4) motion control and parameter setting. The images were captured in the following color palettes: (a) white-hot; (b) rainbow; and (c) red.

2.5. Survival Rate and Individual Reproduction

The survival rate, measured by the number of remaining individuals, and reproduction, measured by the number of eggs and nymphs, of mites on infested plants, were determined 15 days after infestation, by directly counting the living individuals using a bench magnifying glass.

2.6. Statistical Analysis

All analyses were conducted in R Core Team [21]. The data were grouped into reproductive rates, consisting of the number of eggs, nymphs (larval, protonymph, and deutonymph stages), adults, and the total, which corresponded to the sum of the number of individuals in all stages. The data were first analyzed using multivariate permutation analysis (PERMANOVA), with 10,000 permutations, to determine whether there was evidence of differences between the multiple variables between infested and healthy plants. We used the PERMDISP2 procedure R Core Team [21] for the analysis of multivariate homogeneity of group dispersions (variances). Secondly, the hypothesis of a correlation between the physiological variables of passion fruit plants and the reproductive rate, as well as the number of adults and the total, was tested. An unsupervised model, based on multidimensional scaling with the k-means algorithm, for variables related to fluorescence, reproductive rate, and the number of T. mexicanus adults. Additionally, based on a supervised model, statistics from the confusion matrix related to modeling fluorescence as a function of the mite infestation were used, and the following metric statistics were captured: Accuracy, Sensitivity, and Specificity of the models. The parameter initial fluorescence (Fo) was retained for supervised modeling due to the high correlations between Fo and the number of eggs and total individuals [22]. The last statistical test was a sign test to compare the differences between the medians from the thermal responses of infested and non-infested leaves by mites.

3. Results

Based on Permanova, there is evidence of differences between treatments of plants infested and not infested by T. mexicanus (F1,8 = 51.73; p = 0.01). Therefore, at least two physiological variables of passion fruit plants are significantly affected by the attack of T. mexicanus (Table 1).
Variables with correlation indices below 0.50 were excluded from further statistical analyses. Therefore, the following variables Fo, Fv, and Fm remained in the subsequent analyses, along with their respective Fv/Fm and Fv/Fo relationships. Based on the correlation matrix, it is clear to observe the impact of both reproductive activity and the total number of mites found on the directly negative relationship between the Fo, Fv, and Fm variables, with Spearman’s correlation values ranging from −0.56 (reproductive rate vs. Fv relationship) to −0.87 (reproductive rate vs. Fo relationship). The only positive and moderate relationship found was between the number of T. mexicanus adults and Fv/Fo (ρ = 0.50) (Figure 1).
The unsupervised model, based on multidimensional scaling with the k-means algorithm, for variables related to fluorescence, reproductive rate, and the number of T. mexicanus adults, revealed a clear separation between infested passion fruit plants (Group 01) and non-infested plants (Group 02). Only one of the observations from the infested condition was incorrectly allocated (Figure 2).
Since the correlations with initial fluorescence (Fo) were the highest, at −0.82 and −0.87 for eggs and total individuals, respectively, the machine learning analysis was conducted only for the relationships between Fo and reproductive rate, as well as between Fo and adults. The results showed that the Fo response was described with high accuracy for reproductive rate (75%) and for the total number of eggs, nymphs, and adults (99.99%). Kappa values were moderate (Kappa = 0.50) and high (Kappa = 0.99) for reproductive rate and total T. mexicanus (Table 2). The sensitivity and specificity metrics were 100% for both cases (Table 2). Therefore, this supervised machine learning model reinforces previous findings, revealing that a clear relationship exists between mite infestation and the fluorescence of passion fruit plants.
Based on the thermal response, it was found that the infested passion fruit plants presented a median of 25.1 °C, in contrast to the median temperature of 25.7 °C. The applied sign test reveals evidence of differences between these medians (χ2 = 3.60; df = 1; p = 0.0477) (Figure 3). In addition, there was no Confidence Interval (CI95%) overlapping between Control (CI95% = 25.38–25.44 °C) and Treatment (CI95% = 24.06–25.08 °C). The differences in the thermal response pattern of the passion fruit plants are revealed in Figure 4.

4. Discussion

Our analysis revealed that herbivory by T. mexicanus was the determining factor in the fluorescence changes observed in passion fruit plants. The machine learning metric values showed agreement between the observed and predicted data. Furthermore, sensitivity and specificity showed that all cases were correctly classified, with no false positives or negatives [22,23]. These results are consistent with those reported in modeling studies applied to precision agriculture, in which high metrics are considered fundamental to ensure the reliable detection of biotic stresses [24,25]. Moreover, we emphasize the risk of overfitting in machine learning models, particularly given the small sample size used in the current research.
There was a distinct separation between infested and non-infested plants, suggesting that mite damage alters the physiological response in a multivariate perspective. This result is consistent with studies in tomato, where fluorescence imaging techniques allowed differentiation between herbivore-attacked and non-attacked plants. Thus, fluorescence can act as a sensitive marker of physiological responses to insect pests in different hosts [23,25].
Chlorophyll a fluorescence is a widely used method for understanding the functioning of the photosynthetic apparatus in plants under stress conditions [23,26]. It is based on the ability of chlorophyll a to emit photons when excited by light, providing rapid and non-destructive information on the photochemical efficiency of photosystem II (PSII) [23,27]. The parameters analyzed in this study are fundamental for this interpretation. Initial fluorescence (Fo) represents the emission when the PSII reaction centers are open. At the same time, maximum fluorescence (Fm) corresponds to the maximum emission level after light saturation, when all centers are closed. Variable fluorescence (Fv = Fm − Fo) expresses the potential capacity to convert light energy into photochemical energy [28]. Reductions in these values indicate compromised chloroplast integrity and photochemical efficiency.
Significant results from the experiment showed a strong negative correlation between the population density of T. mexicanus and the values of Fo, Fm, and Fv, indicating that herbivory by mites compromised the photochemical efficiency of passion fruit plants. This effect can be explained by the mite’s feeding pattern, which punctures epidermal cells and sucks their contents, destroying chloroplasts and damaging cell membranes, directly affecting the initial stage of photosynthesis [29,30]. It was also observed that the young phase from the reproduction activity was more closely associated with reduced Fo and Fv, suggesting that the early stages of development, characterized by high feeding rates, exert a greater impact on the photosynthetic apparatus. Previous studies in other species corroborate these findings. In soybeans infested with T. urticae, accumulation of reactive oxygen species, chlorophyll degradation, and metabolic changes were reported a few days after infestation [30]. Infested Arabidopsis thaliana plants presented lower Fo values compared to the control, indicating a reduction in photochemical activity [17]. Similarly, in Jatropha curcas, herbivory promoted significant reductions in Fm and Fv, compromising electron transport and photosynthetic efficiency [31]. These results reinforce that the decrease in fluorescence parameters is a consistent physiological pattern across different hosts under attack by spider mites.
Studies with spider mites have shown that T. urticae attacks compromise parameters related to chlorophyll fluorescence and leaf structural integrity in different crops. In strawberry, it was revealed that the progression of infestation promoted characteristic shifts in visible and near-infrared spectrum bands, associated with the collapse of leaf macrostructure, degradation of cell wall polysaccharides, and reorganization of chlorophyll fluorescence [32]. Fluorescence analyses in rice revealed that spider mite infestation triggers distinct changes in the Kautsky curve, which measures chlorophyll fluorescence over time to assess the functional state of Photosystem II (PSII) in plants, reflecting different levels of susceptibility between species and cultivars [33]. Conversely, infested bamboo leaves showed significantly reduced chlorophyll fluorescence values (30.10% loss in the Fv/Fm ratio), indicating a decline in photosynthetic efficiency [34]. In lima beans, effects on pigments were observed, with reductions in chlorophyll a and b levels, increased pheophytin concentrations, and changes in the Chlorophyll a/Chlorophyll b ratio, characterizing the disruption of the photosynthetic apparatus [35].
In our study, herbivory by T. mexicanus resulted in alterations in chloroplast integrity and reduced photochemical efficiency, typical effects of biotic stress processes. In addition to the aforementioned direct damage caused by the deleterious action of mites on plant cells, several species, including those of the genus Tetranychus, can act as transmitters of important viruses that reduce the productive potential of crops essential for human consumption [36]. This further highlights the importance of accurately monitoring these individuals in the field. Furthermore, recent advances have demonstrated the potential of using remote sensing techniques, such as the integration of hyperspectral imaging and machine learning algorithms, for the early detection of T. urticae infestation in other crops, including cotton, with an accuracy of up to 100% [37]. These results confirm that both fluorescence and thermal response can be explored as sensitive indicators that can be integrated into precision monitoring systems, contributing to more effective mite management strategies in different crops.
Regarding leaf temperature, the present study identified minor thermal differences between infested and control plants. In experiments conducted with apple leaves infested with T. urtice, it was observed that, at an air temperature of 25 °C, the range of leaf surface temperatures encompassed the mite’s optimal developmental temperature, which is around 30 °C. Phytomites buffer approximately 29% of leaf surface heating through thermoregulation under moderate environmental conditions. Thus, two-spotted spider mites select leaf areas with surface temperatures close to their optimum for developmental rate or other temperature-dependent life cycle traits, potentially leading to high population growth rates [33]. In contrast to what was observed in vineyards under water stress conditions, where increased leaf temperature favors the proliferation of T. pacificus [34]. In our study, leaves infested with T. mexicanus showed a slight reduction in mean temperature (25.1 °C) compared to control plants (25.7 °C). This modest cooling suggests that mite feeding can sometimes enhance transpiration and evaporative cooling, even though Tetranychus is often associated with stomatal closure and reduced transpiration that is typically found in warm leaves. Previous studies highlight that T. urticae, as a highly generalist herbivore, has evolved strategies to suppress or circumvent plant defenses, including jasmonate signaling, in ways that may alter stomatal behavior and water balance [38]. Herbivore salivary effectors are also known to reprogram host defense responses, further influencing physiological traits such as transpiration and leaf temperature [39]. Recent evidence reinforces the importance of stomatal traits in shaping these outcomes. Rosa-Diaz et al. [40] demonstrated that Arabidopsis mutants with higher stomatal density exhibited stronger cooling. At the same time, those with fewer stomata warmed more during mite infestation, confirming that both stomatal number and aperture are key to determining thermal responses. Our results suggest a scenario in which mite feeding promoted enough transpiration to reduce leaf temperature. This finding indicates that plant thermal responses to the herbivory of Tetranychus species are context-dependent and conditioned by the interaction between host physiology, stomatal regulation, and mite strategies.

5. Conclusions

Tetranhychus mexicanus infestation compromises chlorophyll fluorescence (Fo, Fv, and Fm), impacting photosystem II and reducing photochemical efficiency, with a strong correlation observed between the total number of mites and the infestation. The results reveal that machine learning models confirm the sensitivity of initial fluorescence (Fo) to the density of eggs and adults. Furthermore, thermal analysis shows a reduction in leaf temperature in infested passion fruit plants. These results demonstrate a sensitive indicator of the maximum quantum efficiency of PSII, revealing the impact of mite infestation on photosynthetic performance and, consequently, plant physiology.

Author Contributions

Conceptualization, M.A.d.C.L. and E.B.F.; methodology, M.M.L.d.C. and M.T.d.S.; validation, R.Í.L.d.S., S.S.M. and M.B.d.A.; formal analysis, J.B.M. All authors have read and agreed to the published version of the manuscript.

Funding

JBM is supported by a grant from the Brazilian National Council for Scientific and Technological Development (CNPq) (Process numbers: 420064/2023-0 and 308296/2025-7). The research received financial assistance from the Coordination for the Improvement of Higher Education Personnel (CAPES), Brazil.

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author(s).

Acknowledgments

The authors would like to thank the Graduate Program in Agronomy at the Federal University of Paraiba.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Araújo, D.L.; Souto, A.G.L.; Cavalcante, A.G.; Cavalcante, L.F.; Pereira, W.E.; Melo, A.S. Physiological aspects of yellow passion fruit with use of hydrogel and mulching. Rev. Caatinga 2022, 35, 382–392. [Google Scholar] [CrossRef]
  2. Antunes, L.F.S.; Vaz, A.F.S.; Martelleto, L.A.P.; Leal, M.A.A.; Alves, R.S.; Ferreira, T.S.; Rumjanek, N.G.; Correia, M.E.F.; Rosa, R.C.C.; Guerra, J.G.M. Sustainable organic substrate production using millicompost in combination with different plant residues for the cultivation of Passiflora edulis seedlings. Environ. Technol. Innov. 2022, 28, 102612. [Google Scholar] [CrossRef]
  3. Souza, G.L.; Nascimento, A.P.; Silva, J.D.A.; Bezerra, F.T.; da Silva, R.Í.; Cavalcante, L.F.; Mendonça, R. Growth of wild passion fruit (Passiflora foetida L.) rootstock under irrigation water salinity. Rev. Bras. Eng. Agríc. Ambient. 2023, 27, 114–120. [Google Scholar] [CrossRef]
  4. Gull, A.; Lone, A.A.; Wani, N.U.I. Abiotic and Biotic Stress in Plants. In Abiotic and Biotic Stress in Plants; IntechOpen: London, UK, 2019; pp. 1–19. [Google Scholar] [CrossRef]
  5. Du, B.; Haensch, R.; Alfarraj, S.; Rennenberg, H. Strategies of plants to overcome abiotic and biotic stresses. Biol. Rev. 2024, 99, 1524–1536. [Google Scholar] [CrossRef]
  6. Mahmood, S.U.; Mao, Y.; Inzamam, U.L.; Fang, X. Eriophyid mites in fruit crops: Biology, ecology, molecular aspects, and innovative control strategies. Not. Bot. Horti Agrobot. Cluj-Napoca 2024, 52, 13781. [Google Scholar] [CrossRef]
  7. Shukla, A. Mites. In Polyphagous Pests of Crops; Springer: Singapore, 2021; pp. 409–455. [Google Scholar] [CrossRef]
  8. Neves, L.G.; Bruckner, C.H.; Picanço, M.C.; Araújo, K.L.; Luz, P.B.; Serafim, M.E. Selection of full and half brothers families of yellow passionfruit resistant to Tetranychus mexicanus. Semin. Ciênc. Agrár. 2015, 36, 2507–2514. [Google Scholar] [CrossRef]
  9. Barrancas, J.F.; Silva, N.M.; Vasconcelos, G.J.N. Biological parameters of Tetranychus mexicanus (McGregor) (Acari: Tetranychidae) in papaya and passion fruit. Pesqui. Agropecu. Trop. 2022, 52, e72154. [Google Scholar] [CrossRef]
  10. Golan, K.; Kot, I.; Górska-Drabik, E.; Jurado, I.G.; Kmieć, K.; Łagowska, B. Physiological response of basil plants to twospotted spider mite (Acari: Tetranychidae) infestation. J. Econ. Entomol. 2019, 112, 948–956. [Google Scholar] [CrossRef] [PubMed]
  11. Franco, R.A.; Reis, P.R.; Zacarias, M.S.; Altoé, B.F.; Barbosa, J.P.R.A.D. Influência da infestação de Oligonychus ilicis (McGregor, 1917) (Acari: Tetranychidae) sobre a taxa de fotossíntese potencial de folhas de cafeeiro. Arq. Inst. Biol. 2009, 76, 205–210. [Google Scholar] [CrossRef]
  12. Akyazi, R.; Soysal, M. Effect of Tetranychus urticae and Polyphagotarsonemus latus (Acari: Tetranychidae, Tarsonemidae) at different infestation levels and feeding durations on chlorophyll content of bean plants. Plant Prot. Bulletin. 2024, 64, 5–13. [Google Scholar] [CrossRef]
  13. Garcia, A.; Talavera-Mateo, L.; Petrik, I.; Oklestkova, J.; Novak, O.; Santamaria, M.E. Spider mite infestation triggers coordinated hormonal trade-offs enabling plant survival with a fitness cost. Physiol. Plant. 2024, 176, e14479. [Google Scholar] [CrossRef] [PubMed]
  14. Palma, G.R.; Mello, R.F.; Godoy, W.A.; Engel, E.; Lau, D.; Markham, C.; Moral, R.A. Forecasting insect abundance using time series embedding and machine learning. Ecol. Inform. 2025, 85, 102934. [Google Scholar] [CrossRef]
  15. Mittal, M.; Gupta, V.; Aamash, M.; Upadhyay, T. Machine learning for pest detection and infestation prediction: A comprehensive review. Wiley Interdiscip. Rev. Data Min. Knowl. Discov. 2024, 14, e1551. [Google Scholar] [CrossRef]
  16. Tramboo, H.A.; Sherwani, A.; Ahmad, M.M.; Khan, A.A.; Malik, A.A.; Jaiswal, R.; Shah, M.A. Canopy stratification, seasonal dynamics, and machine learning-based forecasting of Tetranychus urticae Koch on cucumber under polyhouse conditions. J. Plant Dis. Prot. 2025, 132, 1–16. [Google Scholar] [CrossRef]
  17. Tongkaw, A.; Tongkaw, S. A comparative analysis of machine learning algorithms for allergy scale prediction. In Proceedings of the 2025 11th International Conference on Engineering, Applied Sciences, and Technology (ICEAST), Phuket, Thailand, 2–5 May 2025; IEEE: New York, NY, USA; pp. 69–72. [Google Scholar] [CrossRef]
  18. Alvares, C.A.; Stape, J.L.; Sentelhas, P.C.; Gonçalves, J.L.M.; Sparovek, G. Köppen’s climate classification map for Brazil. Meteorol. Z. 2013, 22, 711–728. [Google Scholar] [CrossRef]
  19. Freire, J.L.O.; Dias, T.J.; Cavalcante, L.F.; Fernandes, P.D.; Lima Neto, A.J. Quantum yield and gas exchange in yellow passion fruit under salinity water, biofertilization and mulch. Rev. Ciênc. Agron. 2014, 45, 82–91. [Google Scholar] [CrossRef]
  20. Jesie, R.S.; Premi, M.G. Improved Tunicate Swarm Optimization Based Hybrid Convolutional Neural Network for Classification of Leaf Diseases and Nutrient Deficiencies in Rice (Oryza). Agronomy 2024, 14, 1851. [Google Scholar] [CrossRef]
  21. R Core Team. R Language Definition; R Foundation for Statistical Computing: Vienna, Austria, 2024; Volume 3, p. 116. [Google Scholar]
  22. Więckowska, B.; Kubiak, K.B.; Jóźwiak, P.; Moryson, W.; Stawińska-Witoszyńska, B. Cohen’s kappa coefficient as a measure to assess classification improvement following the addition of a new marker to a regression model. Int. J. Environ. Res. Public Health 2022, 19, 10213. [Google Scholar] [CrossRef]
  23. Blumenthal, J.; Megherbi, D.B.; Lussier, R. Unsupervised machine learning via hidden Markov models for accurate clustering of plant stress levels based on imaged chlorophyll fluorescence profiles and their rate of change in time. Comput. Electron. Agric. 2020, 174, 105064. [Google Scholar] [CrossRef]
  24. Yang, R.; Qi, Y.; Zhang, H.; Wang, H.; Zhang, J.; Ma, X.; Ma, C. A study on the object-based high-resolution remote sensing image classification of crop planting structures in the Loess Plateau of eastern Gansu Province. Remote Sens. 2024, 16, 2479. [Google Scholar] [CrossRef]
  25. Moustaka, J.; Meyling, N.V.; Hauser, T.P. Root-associated entomopathogenic fungi modulate their host plant’s photosystem II photochemistry and response to herbivorous insects. Molecules 2021, 26, 207. [Google Scholar] [CrossRef]
  26. Goltsev, V.N.; Kalaji, H.M.; Paunov, M.; Bąba, W.; Horaczek, T.; Mojski, J.; Allakhverdiev, S.I. Variable chlorophyll fluorescence and its use for assessing physiological condition of plant photosynthetic apparatus. Russ. J. Plant Physiol. 2016, 63, 869–893. [Google Scholar] [CrossRef]
  27. Zhou, L.; Wu, H.; Jing, T.; Li, T.; Li, J.; Kong, L.; Zhu, F. Application of chlorophyll fluorescence analysis technique in studying the response of lettuce (Lactuca sativa L.) to cadmium stress. Sensors 2024, 24, 1501. [Google Scholar] [CrossRef] [PubMed]
  28. Pszczółkowski, P.; Sawicka, B.; Skiba, D.; Barbaś, P.; Noaema, A.H. The use of chlorophyll fluorescence as an indicator of predicting potato yield, its dry matter and starch in the conditions of using microbiological preparations. Appl. Sci. 2023, 13, 10764. [Google Scholar] [CrossRef]
  29. Silva, P.R.; Santos, C.E.M.; Silva, R.S.; Silva, É.M.; Lopes, M.C.; Paes, J.S.; Picanço, M.C. Assessing the resistance of passion fruit genotypes as control method to Tetranychus mexicanus. Crop Prot. 2022, 128, 104990. [Google Scholar] [CrossRef]
  30. Ruffatto, K.; Silva, L.C.O.; Neves, C.D.O.; Kuntzler, S.G.; Lima, J.C.; Almeida, F.A.; Sperotto, R.A. Unravelling soybean responses to early and late Tetranychus urticae (Acari: Tetranychidae) infestation. Plant Biol. 2024, 26, 1223–1239. [Google Scholar] [CrossRef]
  31. Hsu, M.H.; Chen, C.C.; Lin, K.H.; Huang, M.Y.; Yang, C.M.; Huang, W.D. Photosynthetic responses of Jatropha curcas to spider mite injury. Photosynthetica 2015, 53, 349–355. [Google Scholar] [CrossRef]
  32. Bueno, A.D.F.; Bueno, R.C.O.D.F.; Nabity, P.D.; Higley, L.G.; Fernandes, O.A. Photosynthetic response of soybean to twospotted spider mite (Acari: Tetranychidae) injury. Braz. Arch. Biol. Technol. 2009, 52, 825–834. [Google Scholar] [CrossRef]
  33. Caillon, R.; Suppo, C.; Casas, J.; Woods, H.A.; Pincebourde, S. Warming decreases thermal heterogeneity of leaf surfaces: Implications for behavioural thermoregulation by arthropods. Funct. Ecol. 2014, 28, 1449–1458. [Google Scholar] [CrossRef]
  34. Stavrinides, M.C.; Daane, K.M.; Lampinen, B.D.; Mills, N.J. Plant water stress, leaf temperature, and spider mite (Acari: Tetranychidae) outbreaks in California vineyards. Environ. Entomol. 2010, 39, 1232–1241. [Google Scholar] [CrossRef]
  35. Parmagnani, A.S.; Mannino, G.; Brillada, C.; Novero, M.; Dall’Osto, L.; Maffei, M.E. Biology of Two-Spotted Spider Mite (Tetranychus urticae): Ultrastructure, Photosynthesis, Guanine Transcriptomics, Carotenoids and Chlorophylls Metabolism, and Decoyinine as a Potential Acaricide. Int. J. Mol. Sci. 2023, 24, 1715. [Google Scholar] [CrossRef]
  36. Ferreira, L.Y.M.; Sousa, A.G.; Silva, J.L.; Santos, J.P.N.; Souza, D.G.N.; Orellana, L.C.B.; Santana, S.F.; Vasconcelos, L.B.C.M.; Oliveira, A.R.; Aguiar, E.R.G.R. Characterization of the Virome Associated with the Two-Spotted Spider Mite, Tetranychus urticae. Viruses 2024, 16, 1532. [Google Scholar] [CrossRef]
  37. Yamada, M.; Thiesen, L.V.; Iost Filho, F.H.; Yamamoto, P.T. Hyperspectral imaging and machine learning: A promising tool for the early detection of Tetranychus urticae Koch infestation in cotton. Agriculture 2024, 14, 1573. [Google Scholar] [CrossRef]
  38. Rioja, C.; Zhurov, V.; Bruinsma, K.; Grbic, M.; Grbic, V. Plant–herbivore interactions: A case of an extreme generalist, the two-spotted spider mite Tetranychus urticae. Mol. Plant Microbe Interact. 2017, 30, 935–945. [Google Scholar] [CrossRef]
  39. Acevedo, F.E.; Smith, P.; Peiffer, M.; Helms, A.; Tooker, J.; Felton, G.W. Phytohormones in Fall Armyworm Saliva Modulate Defense Responses in Plants. J. Chem. Ecol. 2019, 45, 598–609. [Google Scholar] [CrossRef]
  40. Rosa-Diaz, I.; Rowe, J.; Cayuela-Lopez, A.; Arbona, V.; Díaz, I.; Jones, A.M. Spider mite herbivory induces an ABA-driven stomatal defense. Plant Physiol. 2024, 195, 2970–2984. [Google Scholar] [CrossRef]
Agronomy 15 02297 g001
Figure 1. Correlation matrix involving physiological variables, reproductive rate, and number of adults of Tetranychus mexicanus in passion fruit plants. Positive correlations are displayed in blue, and negative correlations are displayed in red. Color intensity and the inclination of the circle are proportional to the correlation coefficients. On the right side of the correlogram, the legend colors indicate the correlation coefficients and their corresponding values.
Figure 1. Correlation matrix involving physiological variables, reproductive rate, and number of adults of Tetranychus mexicanus in passion fruit plants. Positive correlations are displayed in blue, and negative correlations are displayed in red. Color intensity and the inclination of the circle are proportional to the correlation coefficients. On the right side of the correlogram, the legend colors indicate the correlation coefficients and their corresponding values.
Agronomy 15 02297 g001
Agronomy 15 02297 g002
Figure 2. Multidimensional scaling involving physiological variables, reproductive rate, and number of Tetranychus mexicanus adults in passion fruit plants (Group 01) and non-infested plants (Group 02). Outputs from the analysis: Dimension 2: Stress: 4.7.26931 × 10−5; Stress type 1, weak ties. The best solution was not repeated after 100 tries; thus, the best solution was from try 75 (random start). Additional information: Scaling: centering, PC rotation, and half-change scaling.
Figure 2. Multidimensional scaling involving physiological variables, reproductive rate, and number of Tetranychus mexicanus adults in passion fruit plants (Group 01) and non-infested plants (Group 02). Outputs from the analysis: Dimension 2: Stress: 4.7.26931 × 10−5; Stress type 1, weak ties. The best solution was not repeated after 100 tries; thus, the best solution was from try 75 (random start). Additional information: Scaling: centering, PC rotation, and half-change scaling.
Agronomy 15 02297 g002
Agronomy 15 02297 g003
Figure 3. Thermal response of passion fruit plants as a function of Tetranychus mexicanus infestation. Outputs from descriptive analysis for control: Mean = 25.52, Standard Error (SE) = 0.028; Confidence Interval Lower (CIL) = 25.44; Confidence Interval Upper (CIU) = 25.59; Minimum value (Min) = 24.50; Maximum value (Max) = 26.20; Quantile 25 (Q25) = 25.5; Quantile 75 (Q75) = 27.50. Outputs from descriptive analysis for treatment: Mean = 25.02, SE = 0.022; CIL = 24.06; CIU = 25.08; Min = 24.50; Max = 25.06; Q25 = 25.00; Q75 = 25.20.
Figure 3. Thermal response of passion fruit plants as a function of Tetranychus mexicanus infestation. Outputs from descriptive analysis for control: Mean = 25.52, Standard Error (SE) = 0.028; Confidence Interval Lower (CIL) = 25.44; Confidence Interval Upper (CIU) = 25.59; Minimum value (Min) = 24.50; Maximum value (Max) = 26.20; Quantile 25 (Q25) = 25.5; Quantile 75 (Q75) = 27.50. Outputs from descriptive analysis for treatment: Mean = 25.02, SE = 0.022; CIL = 24.06; CIU = 25.08; Min = 24.50; Max = 25.06; Q25 = 25.00; Q75 = 25.20.
Agronomy 15 02297 g003
Agronomy 15 02297 g004
Figure 4. Thermal response with images captured in the following color palettes: White-hot, Rainbow, and Red, by the infrared thermographic camera T2 Pro Xinfrared, Xinfrared®, Hong Kong, China, in control passion fruit plants (without infestation) and with infestation of Tetranychus mexicanus.
Figure 4. Thermal response with images captured in the following color palettes: White-hot, Rainbow, and Red, by the infrared thermographic camera T2 Pro Xinfrared, Xinfrared®, Hong Kong, China, in control passion fruit plants (without infestation) and with infestation of Tetranychus mexicanus.
Agronomy 15 02297 g004
Table 1. Summary of multivariate permutation analysis using the Bray method for a reduced model.
Table 1. Summary of multivariate permutation analysis using the Bray method for a reduced model.
SourceR2Fp > F
Model0.866051.730.01
Residual effect0.1339
Total1.0000
Table 2. Confusion matrix for modeling the relationship between initial fluorescence and the number of eggs or adults of Tetranychus mexicanus in passion fruit plants.
Table 2. Confusion matrix for modeling the relationship between initial fluorescence and the number of eggs or adults of Tetranychus mexicanus in passion fruit plants.
Mite VariableAccuracyKappaSensitivitySpecificitySamples (No.)
Reproduction rate75.00%0.50100.00%100.00%100
Total99.99%0.99100.00%100.00%100
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

da Cunha Lima, M.A.; Ferret, E.B.; da Costa, M.M.L.; da Silva, M.T.; da Silva, R.Í.L.; Monteiro, S.S.; de Albuquerque, M.B.; Malaquias, J.B. An Unsupervised and Supervised Machine Learning Approach to Evidence Tetranychus mexicanus (McGregor) Activity in Fluorescence and Thermal Response in Passion Fruit. Agronomy 2025, 15, 2297. https://doi.org/10.3390/agronomy15102297

AMA Style

da Cunha Lima MA, Ferret EB, da Costa MML, da Silva MT, da Silva RÍL, Monteiro SS, de Albuquerque MB, Malaquias JB. An Unsupervised and Supervised Machine Learning Approach to Evidence Tetranychus mexicanus (McGregor) Activity in Fluorescence and Thermal Response in Passion Fruit. Agronomy. 2025; 15(10):2297. https://doi.org/10.3390/agronomy15102297

Chicago/Turabian Style

da Cunha Lima, Maria Alaíne, Eleazar Botta Ferret, Magaly Morgana Lopes da Costa, Mariana Tamires da Silva, Roberto Ítalo Lima da Silva, Shirley Santos Monteiro, Manoel Bandeira de Albuquerque, and José Bruno Malaquias. 2025. "An Unsupervised and Supervised Machine Learning Approach to Evidence Tetranychus mexicanus (McGregor) Activity in Fluorescence and Thermal Response in Passion Fruit" Agronomy 15, no. 10: 2297. https://doi.org/10.3390/agronomy15102297

APA Style

da Cunha Lima, M. A., Ferret, E. B., da Costa, M. M. L., da Silva, M. T., da Silva, R. Í. L., Monteiro, S. S., de Albuquerque, M. B., & Malaquias, J. B. (2025). An Unsupervised and Supervised Machine Learning Approach to Evidence Tetranychus mexicanus (McGregor) Activity in Fluorescence and Thermal Response in Passion Fruit. Agronomy, 15(10), 2297. https://doi.org/10.3390/agronomy15102297

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop