Machine-learning assisted phenotyping: From fungal morphology to mode of action hypothesis - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2019

Machine-learning assisted phenotyping: From fungal morphology to mode of action hypothesis

Résumé

Beyond growth inhibition, fungicides can also trigger specific morphological modifications visualized under transmitted light microscopy. These morphological changes result from the activity of a given compound via the inhibition of a molecular target, commonly named as its mode of action (MoA). We are hence able to classify different molecules into their respective MoA by observing their phenotypic signature, and even to detect new MoA with unknown phenotypic effect for further deconvolution. The aim of the presented work is to develop a robust method for automated recognition and classification of these phenotypic signatures in order to lead to a Mode of Action hypothesis. We compare two machine-learning methods (Random forest and Convolutional Neural Network) for direct processing of images generated on the grey mold Botrytis cinerea subjected to different antifungal molecules. © Bayer | Abteilung | Verfasser | Datum
Fichier principal
Vignette du fichier
IUPAC_2019_Vernay_AI and FungalPhenotyping.pdf (1.25 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-02403936 , version 1 (11-12-2019)

Identifiants

  • HAL Id : hal-02403936 , version 1

Citer

Sarah Laroui, Eric Debreuve, Xavier Descombes, François Villalba, Florent Villiers, et al.. Machine-learning assisted phenotyping: From fungal morphology to mode of action hypothesis. IUPAC - 14th International Congress of Crop Protection Chemistry, May 2019, Ghent, Belgium. ⟨hal-02403936⟩
218 Consultations
59 Téléchargements

Partager

Gmail Facebook X LinkedIn More