Self-supervised learning for analysis of temporal and morphological drug effects in cancer cell imaging data

Abstract

In this work, we propose two novel methodologies to study temporal and morphological phenotypic effects caused by different experimental conditions using imaging data. As a proof of concept, we apply them to analyze drug effects in 2D cancer cell cultures. We train a convolutional autoencoder on 1M images dataset with random augmentations and multi-crops to use as feature extractor. We systematically compare it to the pretrained state-of-the-art models. We further use the feature extractor in two ways. First, we apply distance-based analysis and dynamic time warping to cluster temporal patterns of 31 drugs. We identify clusters allowing annotation of drugs as having cytotoxic, cytostatic, mixed or no effect. Second, we implement an adversarial/regularized learning setup to improve classification of 31 drugs and visualize image regions that contribute to the improvement. We increase top-3 classification accuracy by 8% on average and mine examples of morphological feature importance maps. We provide the feature extractor and the weights to foster transfer learning applications in biology. We also discuss utility of other pretrained models and applicability of our methods to other types of biomedical data.

CellDeathPred: a deep learning framework for ferroptosis and apoptosis prediction based on cell painting

“…Another study proposed a U-Net architecture to synthesize AT8-pTau image given two DAPI and YFP-tau image channels [15]. With the potential of DL architectures in extracting meaningful features directly from microscopic images, recent studies proposed selfsupervised learning frameworks, including a framework for studying the temporal drug effect on cancer cell images, or a framework to learn phenotypic embeddings of HCS images using self-supervised triplet network [16,17]. While these advancements in DL application to HCS images offer the potential to accelerate drug discovery, so far there is only very little work about the analysis and prediction of regulated cell death.…”

CellDeathPred: A Deep Learning framework for Ferroptosis and Apoptosis prediction based on cell painting

“…Another study proposed a U-Net architecture to synthesize AT8-pTau image given two DAPI and YFP-tau image channels (15). With the potential of DL architectures in extracting meaningful features directly from microscopic images, recent studies proposed self-supervised learning frameworks, including a framework for studying the temporal drug effect on cancer cell images, or a framework to learn phenotypic embeddings of HCS images using self-supervised triplet network (16,17). While these advancements in DL application to HCS images offer the potential to accelerate drug discovery, so far there is only very little work about the analysis and prediction of regulated cell death.…”