Cell-painting phenotyping in filamentous fungi: deep learning for image-based drug discovery

Polina Simkina, Bruno Leggio, Gerard Sanroma-Guell, Yoann Huet

Keywords: Drug discovery, Biological Image Analysis, Self-supervised models

Abstract: The advancement of Cell Painting, a relatively new morphological profiling assay, requires adequate and effective image analysis methods suitable for high-content data. That is why, developing deep learning methods for biological image analysis is extremely relevant. In this paper, we present feature extraction results on Cell Painting images of phytopathogenic fungi using deep learning techniques. We give an overview of the current method, which is based on a supervised convolutional neural network, and present the prospects and advantages of leveraging a self-supervised model based on vision transformers in the given problem setting.

Submission Number: 35