Color traits of soybean conventional and transgenic cultivars are affected by glyphosate concentrations using digital imaging

Soybean is the most cultivated oilseed crop in the world, with much of the merit obtained in recent years due to genetic improvement in which greater genetic progress can be obtained by improving physiological characteristics, which result in a greater impact on yield. Color spectra can be good indicators of the physiological quality of plants by quantifying the intensity of colors in the RGB spectra, being a non-destructive method that allows optimizing the collection and the number of data. We aimed to identify differences in spectral reflection between conventional, transgenic RR1 and RR2 soybean cultivars under the effect of glyphosate herbicide. The plants were cultivated in five-liter pots in a greenhouse in a randomized block design, following cultural treatments recommended for the soybean crop. Leaves of plants in vegetative stages V1 or V2 were collected and duly identified to compose a second experiment, being then submitted to incubation in plastic trays in which the treatments were organized in a completely randomized design in a 6x6 factorial scheme, with three replications. Glyphosate herbicide doses (0.0, 0.03%, 0.06%, 0.12% and 0.24% AE/ha) were added to the leaf petiole on a cotton pad in order to maintain constant contact with the respective dose. The trays were kept under ambient conditions for 14 days after incubation under 16h-light and 8h-dark of artificial light. Images were collected with a smartphone camera on the 13th day after leaves collection (DAC) in an appropriate studio to maintain adequate lighting, and on the 14th DAC SPAD index data were collected in three locations on the leaf, avoiding the midrib. The images were cropped and then segmented using manual thresholding in which the average values for the red, green and blue channels were extracted. The Excess Red Index (ExR) was calculated using the red and green channels data. The data obtained were analyzed and the significant effects of the model were analyzed by the Skott-Knott test for the cultivar factor and regression models were adjusted for the dose factor. Spearmans correlation was used to verify the relationship between the studied variables. In view of our results, glyphosate affects the chlorophyll of resistant plants when subjected to continuous exposure and at high doses, leading to senescence and that the red channel information can be used to infer the level of interference in the photosynthetic activity of plants subjected to the herbicide.