# Can Winged Aphid Abundance Be a Predictor of Cucurbit Aphid-Borne Yellows Virus Epidemics in Melon Crop?

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Materials and Methods

#### 2.1. Field Experiments

#### 2.2. Plant Sampling and Virus Monitoring

#### 2.3. Insect Sampling and Aphid Monitoring

#### 2.4. Computation of Variables Related to Virus Epidemics

_{i}represents CABYV incidence, expressed as a percentage, at date D

_{i}.

_{56}was divided by the total virus monitoring duration (56 days) to calculate the mean incidence over the epidemic. On the basis of their mean incidence, epidemics were categorized as mild (0–20%), intermediate (21–40%), severe (41–60%) or extreme (61–100%) [7].

_{t,k}is the incidence, expressed as a percentage, at time t ($t\in \u27e61;56\u27e7$) and for trial k ($k\in \u27e61;11\u27e7$); µ

_{k}is the abscissa of the inflection point for trial k, i.e., the date of the epidemic peak. Low values of µ indicate precocious epidemics while high values are associated with late epidemics; γ

_{k}is the plateau, i.e., the carrying capacity, for trial k. High values of γ indicate global epidemics (high incidence at the end of the season) whereas low values mean limited epidemics; α

_{k}is related to the slope at the inflection point for trial k, it reflects the speed of epidemic around the peak. Roughly, high values of α mean fast epidemics and low values mean slow epidemics; µ and α are positive parameters; and γ is bounded 0 and 1.

_{56}, µ and γ.

#### 2.5. Computation of Variables Related to Aphid Abundance

_{1}(${t}_{1}\in \u27e61;55\u27e7$) to time t

_{2}(${t}_{2}\in \u27e6{\mathrm{t}}_{1};55\u27e7$), resulting in 1540 different variables for each aphid species. Secondly, daily abundance was aggregated on periods relative to the date of epidemic peak (estimated with parameter µ of the logistic curve), by calculating the sum from time t

_{1}= µ − Δt

_{1}($\mathsf{\Delta}{t}_{1}\in \u27e61;\mathsf{\mu}\u27e7$) to time t

_{2}= µ − Δt

_{1}+ Δt

_{2}($\mathsf{\Delta}{t}_{2}\in \u27e61;55-\text{\xb5}+\mathsf{\Delta}\mathrm{t}1\u27e7$). Depending on the value of µ, this resulted in a maximum of 3025 additional variables.

#### 2.6. Relationship between Aphid and Virus Variables

_{56}, µ and γ), a relationship with one or several aphid variables was established in three steps. In a first step, we used the Spearman test with a maximal type-1 error of 1% to identify aphid variables that were significantly correlated to the virus variable under consideration. In a second step, for each remaining aphid variable, we modelled the relationship between the virus variable (dependent variable) and the aphid variable (explanatory variable). For AUDPC

_{56}, we used the following linear regression (3):

_{k}—the value of the virus variable (i.e., AUDPC

_{56}, µ or γ) for trial k ($k\in \u27e61;11\u27e7$);

_{k}—the value of the aphid variable for trial k;

_{0}and A

_{1}—the parameters of the linear model for AUPDC

_{56};

_{0}, B

_{1}and B

_{2}—the parameters of the exponential model for µ and γ, such as z

_{k}(0) = B

_{0}and z

_{k}(∞) = B

_{0}+B

_{1}.

#### 2.7. Data & Software

## 3. Results

#### 3.1. Virus Epidemics

#### 3.2. Vector Abundances

#### 3.3. Correlations between Virus and Aphid Variables

_{56}, µ and γ; parameter α was not included because its influence on virus incidence was negligible) and more than 9000 aphid variables. These aphid variables were computed by aggregating abundances on periods relative either to the planting date, or to the date of the epidemic peak (i.e., µ, the abscissa of the inflection point of the logistic curve). Depending on the virus variable under consideration, the Spearman test yielded a diverse number of significant correlations with one or several aphid variables. For AUDPC

_{56}, 413 significant correlations were obtained with abundances of A. gossypii or the total aphid population aggregated on periods relative to the planting date (Table S1), and one correlation was obtained with A. gossypii abundance aggregated on a period relative to the date of epidemic peak (four consecutive days starting from 11 days before the epidemic peak) (Table S2). Parameter µ was correlated to aphid variables involving either A. gossypii or the total aphid population aggregated on periods of 1 to 10 consecutive days within the two first weeks of cropping (Table S1). For parameter γ, 10 significant correlations were obtained with abundances of A. gossypii or M. persicae aggregated on periods of 1 to 9 consecutive days within the three first weeks of cropping (Table S1, Figure S2A). Ten supplementary significant correlations were found with A. gossypii abundances aggregated on periods of 1 to 12 consecutive days before or around the inflection point (Table S2, Figure S2B).

#### 3.4. Selection of the Best Aphid Variables Based on Their Potential to Explain Virus Variables

_{56}and exponential models to relate µ and γ with aphid variables used as single explanatory variables. Among these models, we selected those associated with the lowest mean square error (MSE). The best linear model to explain the variability of AUDPC

_{56}was obtained with the abundance of A. gossypii aggregated between D11 and D17 (Figure 3). The variability of parameter µ was best explained by a negative exponential model involving the abundance of A. gossypii aggregated between D1 and D10. With regard to parameter γ, the best exponential model involved the abundance of A. gossypii aggregated between D12 and D14.

#### 3.5. Prediction of CABYV Epidemics

## 4. Discussion

_{56}could be explained by the abundance of A. gossypii aggregated between D11 and D17 using a simple linear model. The variability of µ and γ were respectively explained by the abundance of A. gossypii aggregated between D1 and D10, and between D12 and D14 using exponential models. It is noteworthy that these two parameters can be predicted as early as two weeks after planting/fleece removal. Thereby, using these predicted parameter values in the logistic equation, it is possible to have an early insight into the probable CABYV dynamic.

## Supplementary Materials

_{k}= A

_{0}+ A

_{1}· x

_{k}) for the area under the disease progress curve (AUDPC

_{56}) and an exponential model (${z}_{k}={B}_{0}+{B}_{1}\left(1-{e}^{-{B}_{2}.{x}_{k}}\right)$) for the parameters of the logistic equation (µ and γ). For AUDPC

_{56}, only the 15 correlations having the lowest mean square errors are presented (total of 413), Table S2: Significant correlations (with a maximal type-1 error of 1%) between virus variables and aphid variables computed on periods relative to the date of epidemic peak (µ). Significant correlations involve abundances of Aphis gossypii (RIS-181) aggregated between t

_{1}= µ − Δt

_{1}and t

_{2}= µ − Δt

_{1}+ Δt

_{2}(in days from planting date). The relationship between the virus variable and the aphid variable was modelled with a linear model (${z}_{k}={A}_{0}+{A}_{1}.{x}_{k}$) for the area under the disease progress curve (AUDPC

_{56}) and an exponential model (${z}_{k}={B}_{0}+{B}_{1}\left(1-{e}^{-{B}_{2}.{x}_{k}}\right)$) for the parameter γ of the logistic equation. No aphid variable was found significantly correlated to µ.

## Author Contributions

## Funding

## Acknowledgments

## Conflicts of Interest

## References

- Lecoq, H.; Bourdin, D.; Wipfscheibel, C.; Bon, M.; Lot, H.; Lemaire, O.; Herrbach, E. A new yellowing disease of cucurbits caused by a luteovirus, cucurbit aphid-borne yellows virus. Plant Pathol.
**1992**, 41, 749–761. [Google Scholar] [CrossRef] - CABI. Cucurbit aphid-borne yellows virus (original text by Lecoq, H.; Desbiez, C.; and Schoeny, A.). In Invasive Species Compendium; CABI: Wallingford, UK, 2020; Available online: http://www.cabi.org/isc/datasheet/110067 (accessed on 17 June 2020).
- Desbiez, C.; Wipf-Scheibel, C.; Millot, P.; Berthier, K.; Girardot, G.; Gognalons, P.; Hirsch, J.; Moury, B.; Nozeran, K.; Piry, S.; et al. Distribution and evolution of the major viruses infecting cucurbitaceous and solanaceous crops in the French Mediterranean area. Virus Res.
**2020**, 286, 198042. [Google Scholar] [CrossRef] [PubMed] - Juarez, M.; Legua, P.; Mengual, C.M.; Kassem, M.A.; Sempere, R.N.; Gomez, P.; Truniger, V.; Aranda, M.A. Relative incidence, spatial distribution and genetic diversity of cucurbit viruses in eastern Spain. Ann. Appl. Biol.
**2013**, 162, 362–370. [Google Scholar] [CrossRef] - Lecoq, H.; Desbiez, C. Viruses of Cucurbit Crops in the Mediterranean Region: An Ever-Changing Picture. In Viruses and Virus Diseases of Vegetables in the Mediterranean Basin; Elsevier Academic Press Inc.: San Diego, CA, USA, 2012; Volume 84, pp. 67–126. [Google Scholar]
- Kassem, M.A.; Juarez, M.; Gomez, P.; Mengual, C.M.; Sempere, R.N.; Plaza, M.; Elena, S.F.; Moreno, A.; Fereres, A.; Aranda, M.A. Genetic Diversity and Potential Vectors and Reservoirs of Cucurbit aphid-borne yellows virus in Southeastern Spain. Phytopathology
**2013**, 103, 1188–1197. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Schoeny, A.; Desbiez, C.; Millot, P.; Wipf-Scheibel, C.; Nozeran, K.; Gognalons, P.; Lecoq, H.; Boissot, N. Impact of Vat resistance in melon on viral epidemics and genetic structure of virus populations. Virus Res.
**2017**, 241, 105–115. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Lecoq, H.; Katis, N. Control of Cucurbit Viruses. In Control of Plant Virus Diseases—Seed-Propagated Crops; Elsevier Academic Press Inc.: San Diego, CA, USA, 2014; Volume 90, pp. 255–296. [Google Scholar]
- Dogimont, C.; Bussemakers, A.; Martin, J.; Slama, S.; Lecoq, H.; Pitrat, M. Two complementary recessive genes conferring resistance to Cucurbit Aphid Borne Yellows Luteovirus in an Indian melon line (cucumis melo L.). Euphytica
**1997**, 96, 391–395. [Google Scholar] [CrossRef] - Kassem, M.A.; Gosalvez, B.; Garzo, E.; Fereres, A.; Gómez-Guillamón, M.L.; Aranda, M.A. Resistance to Cucurbit aphid-borne yellows virus in Melon Accession TGR-1551. Phytopathology
**2015**, 105, 1389–1396. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Boissot, N.; Schoeny, A.; Vanlerberghe-Masutti, F. Vat, an Amazing Gene Conferring Resistance to Aphids and Viruses They Carry: From Molecular Structure to Field Effects. Front. Plant Sci.
**2016**, 7, 1420. [Google Scholar] [CrossRef] [PubMed] [Green Version] - Schoeny, A.; Gognalons, P. Data on winged insect dynamics in melon crops in southeastern France. Data Brief
**2020**, 29, 105132. [Google Scholar] [CrossRef] [PubMed] - Saltelli, A.; Annoni, P.; Azzini, I.; Campolongo, F.; Ratto, M.; Tarantola, S. Variance based sensitivity analysis of model output. Design and estimator for the total sensitivity index. Comput. Phys. Commun.
**2010**, 181, 259–270. [Google Scholar] [CrossRef] - Team, R.C. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2012. [Google Scholar]
- Pujol, G.; Iooss, B.; Janon, A.; Khalid, B.; Da Veiga, S.; Delage, T.; Fruth, J.; Gilquin, L.; Guillaume, J.; Le Gratiet, L.; et al. Sensitivity: Global Sensitivity Analysis of Model Outputs. R Package Version 1.11.0. 2017. Available online: https://CRAN.R-project.org/package=sensitivity (accessed on 2 March 2020).
- Batista, L.; Velazquez, K.; Estevez, I.; Pena, I.; Lopez, D.; Reyes, M.L.; Rodriguez, D.; Laranjeira, F.F. Spatiotemporal dynamics of Citrus tristeza virus in Cuba. Plant Pathol.
**2008**, 57, 427–437. [Google Scholar] [CrossRef] - Moreira, A.S.; Bergamin, A.; Rezende, J.A.M. Comparative Epidemiology of Three Virus Diseases on Zucchini Squash. J. Phytopathol.
**2015**, 163, 386–394. [Google Scholar] [CrossRef] - Nutter, F.W.J. Quantifying the temporal dynamics of plant virus epidemics: A review. Crop Prot.
**1997**, 16, 603–618. [Google Scholar] [CrossRef] - Salvalaggio, A.E.; Lopez Lambertini, P.M.; Cendoya, G.; Huarte, M.A. Temporal and spatial dynamics of Tomato spotted wilt virus and its vector in a potato crop in Argentina. Ann. Appl. Biol.
**2017**, 171, 5–14. [Google Scholar] [CrossRef] - Werker, A.R.; Dewar, A.; Harrington, R. Modelling the incidence of virus yellows in sugar beet in the UK in relation to numbers of migrating Myzus persicae. J. Appl. Ecol.
**1998**, 35, 811–818. [Google Scholar] [CrossRef] [Green Version] - Macaulay, E.D.M.; Tatchell, G.M.; Taylor, L.R. The Rothamsted Insect Survey ‘12-metre’suction trap. Bull. Entomol. Res.
**1988**, 78, 121–128. [Google Scholar] [CrossRef] - Congdon, B.S.; Coutts, B.A.; Jones, R.A.C.; Renton, M. Forecasting model for Pea seed-borne mosaic virus epidemics in field pea crops in a Mediterranean-type environment. Virus Res.
**2017**, 241, 163–171. [Google Scholar] [CrossRef] [PubMed] - Steinger, T.; Goy, G.; Gilliand, H.; Hebeisen, T.; Derron, J. Forecasting virus disease in seed potatoes using flight activity data of aphid vectors. Ann. Appl. Biol.
**2015**, 166, 410–419. [Google Scholar] [CrossRef] - Lecoq, H.; Pitrat, M. Effects of resistance on the epidemiology of virus diseases of cucurbits. In Proceedings of Cucurbitaceae 89: Evaluation and Enhancement of Cucurbit Germplasm; USDA Vegetable Lab.: Charleston, SC, USA, 1989; pp. 404–408. [Google Scholar]
- Schoeny, A.; Lauvernay, A.; Lambion, J.; Mazzia, C.; Capowiez, Y. The beauties and the bugs: A scenario for designing flower strips adapted to aphid management in melon crops. Biol. Control
**2019**, 136, 103986. [Google Scholar] [CrossRef] [Green Version] - Thomas, S.; Vanlerberghe-Masutti, F.; Mistral, P.; Loiseau, A.; Boissot, N. Insight into the durability of plant resistance to aphids from a demo-genetic study of Aphis gossypii in melon crops. Evol. Appl.
**2016**, 9, 756–768. [Google Scholar] [CrossRef] [PubMed]

**Figure 1.**CABYV disease progress and aphid abundance assessed in melon crops in eleven field trials conducted in Avignon between 2010 and 2019. Black dots represent observed cumulative incidences (proportion of infected plants expressed as a ratio). Black solid lines are fitted curves (logistic model). Red dashed lines represent daily abundances of the pool of CABYV aphid vectors (Aphis gossypii, Myzus persicae, Macrosiphum euphorbiae).

**Figure 2.**First-order and total Sobol’s sensitivity indices of the three parameters of the logistic equation on the average virus incidence. µ is the abscissa of the inflection point (i.e., the date of the epidemic peak); γ is the plateau (i.e., the carrying capacity); α is related to the slope at the inflection point (i.e., the speed of epidemic around the peak). The first-order index indicates the main influence of a parameter, whereas the total index includes its interactions with other parameters.

**Figure 3.**Best models obtained between the virus variables (dependent) and the aphid variables (explanatory). A linear model (${z}_{k}={A}_{0}+{A}_{1}.{x}_{k}$) was used for the area under the disease progress curve (AUDPC) and an exponential model (${z}_{k}={B}_{0}+{B}_{1}\left(1-{e}^{-{B}_{2}.{x}_{k}}\right)$) was used for two parameters of the logistic equation (µ, γ).

**Figure 4.**Observed and modelled CABYV epidemic dynamics in melon crops for eleven field trials conducted in Avignon between 2010 and 2019. Black dots represent observed cumulative incidences (proportion of infected plants expressed as a ratio). Black solid lines are fitted curves (logistic model). Red dashed lines represent rebuilt dynamics from the best predictive aphid variables.

**Table 1.**Melon crop and sampling details for field trials conducted in Avignon between 2010 and 2019.

Trial Code | Experimental Site | Planting Date | Number of Plants | Number of Rows | Number of Plants Per Row | Row Spacing (m) | Plant Spacing (m) | Number of Plants Sampled Per Date (Sampling Effort %) |
---|---|---|---|---|---|---|---|---|

M10 | St Paul | 28/05/2010 | 160 | 8 | 20 | 2 | 0.8 | 26 (16%) |

V11 | St Paul | 09/05/2011 ^{a} | 120 | 6 | 20 | 2 | 0.5 | 24 (20%) |

V12 | St Paul | 11/05/2012 ^{a} | 150 | 6 | 25 | 2 | 0.5 | 24 (16%) |

V13 | St Paul | 06/05/2013 ^{a} | 150 | 6 | 25 | 2 | 0.5 | 24 (16%) |

P11 | St Paul | 24/05/2011 | 208 | 16 | 13 | 1.5 | 0.5 | 40 (19%) |

P12 | St Paul | 31/05/2012 | 240 | 16 | 15 | 1.5 | 0.5 | 40 (17%) |

P13 | St Paul | 24/05/2013 | 240 | 16 | 15 | 1.5 | 0.5 | 32 (13%) |

P14 | St Paul | 27/05/2014 | 240 | 16 | 15 | 1.5 | 0.5 | 40 (17%) |

P15 | St Paul | 28/05/2015 | 240 | 16 | 15 | 1.5 | 0.5 | 40 (17%) |

M18 | St Maurice | 25/05/2018 | 160 | 8 | 20 | 1.5 | 0.5 | 96 (60%) |

M19 | St Maurice | 28/05/2019 | 160 | 8 | 20 | 1.5 | 0.5 | 96 (60%) |

^{a}Agryl P17 fleece removal; fleece optimizes plant growth by increasing both air and soil temperatures and reducing wind damage.

**Table 2.**CABYV epidemics and winged aphid abundances in melon crops in eleven field trials conducted in Avignon between 2010 and 2019. Epidemics are summarized by their area under the disease progress curve calculated over 56 days (AUDPC

_{56}), mean incidence (AUDPC

_{56}/56), epidemic category and parameter estimates of the logistic models (µ, γ and α) fitted to cumulative incidences. Aphis gossypii (RIS-181), Myzus persicae (RIS-322), Macrosiphum euphorbiae (RIS-410) and total aphid abundances were monitored with suction traps between 1 and 55 days after planting.

Trial | AUDPC_{56} | Mean Incidence (%) | Epidemic Category ^{a} | µ | γ | α | RIS-181 | RIS-322 | RIS-410 | Total Aphids |
---|---|---|---|---|---|---|---|---|---|---|

M10 | 3635 | 65 | Extreme | 20 | 1.00 | 0.141 | 1693 | 110 | 0 | 3468 |

M18 | 2443 | 44 | Severe | 21 | 0.71 | 0.046 | 90 | 5 | 0 | 810 |

M19 | 697 | 12 | Mild | 26 | 0.24 | 0.078 | 76 | 7 | 3 | 841 |

P11 | 3540 | 63 | Extreme | 21 | 1.00 | 0.066 | 776 | 72 | 0 | 3113 |

P12 | 2878 | 51 | Severe | 22 | 0.86 | 0.051 | 207 | 24 | 0 | 4004 |

P13 | 1330 | 24 | Intermediate | 40 | 0.86 | 0.045 | 506 | 49 | 0 | 1772 |

P14 | 1617 | 29 | Intermediate | 32 | 0.70 | 0.037 | 277 | 139 | 0 | 1382 |

P15 | 2834 | 51 | Severe | 27 | 0.99 | 0.044 | 407 | 53 | 0 | 2271 |

V11 | 2502 | 45 | Severe | 32 | 1.00 | 0.067 | 256 | 379 | 1 | 2488 |

V12 | 1379 | 25 | Intermediate | 36 | 0.71 | 0.028 | 317 | 118 | 2 | 4097 |

V13 | 671 | 12 | Mild | 51 | 1.00 | 0.066 | 246 | 46 | 3 | 1468 |

^{a}On the basis of their mean incidence, epidemics were categorized as mild (0–20%), intermediate (21–40%), severe (41–60%) or extreme (61–100%) [7].

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Schoeny, A.; Rimbaud, L.; Gognalons, P.; Girardot, G.; Millot, P.; Nozeran, K.; Wipf-Scheibel, C.; Lecoq, H.
Can Winged Aphid Abundance Be a Predictor of Cucurbit Aphid-Borne Yellows Virus Epidemics in Melon Crop? *Viruses* **2020**, *12*, 911.
https://doi.org/10.3390/v12090911

**AMA Style**

Schoeny A, Rimbaud L, Gognalons P, Girardot G, Millot P, Nozeran K, Wipf-Scheibel C, Lecoq H.
Can Winged Aphid Abundance Be a Predictor of Cucurbit Aphid-Borne Yellows Virus Epidemics in Melon Crop? *Viruses*. 2020; 12(9):911.
https://doi.org/10.3390/v12090911

**Chicago/Turabian Style**

Schoeny, Alexandra, Loup Rimbaud, Patrick Gognalons, Grégory Girardot, Pauline Millot, Karine Nozeran, Catherine Wipf-Scheibel, and Hervé Lecoq.
2020. "Can Winged Aphid Abundance Be a Predictor of Cucurbit Aphid-Borne Yellows Virus Epidemics in Melon Crop?" *Viruses* 12, no. 9: 911.
https://doi.org/10.3390/v12090911