DeepPaint: A deep-learning package for Cell Painting Image Classification
Luna, D.; Johnson, E.; Dunphy, L.; McQuillen, R. J.
Show abstract
Recent developments in the high-content imaging (HCI) space have allowed for the production of large Cell Painting datasets. These datasets are typically derived from cells exposed to a set of biological perturbants including proteins, small molecules, or even pathogens. While the method of Cell Painting has shown utility for drug discovery and hazard evaluation purposes, traditional analyses pipelines applied Cell Painting datasets typically require the segmentation of single cells from thousands to millions of images, a process that is time-consuming and subject to noise and experimental variability. Here we present DeepPaint, a Python package that uses a deep learning framework to perform image analysis of cell painting images including treatment classification and latent space analysis, circumventing the need for image segmentation. DeepPaint is easily tunable to different HCI setups and datasets and can be applied to classify broad types of biological perturbations. Here we demonstrate that DeepPaint can generate highly accurate neural networks for binary and multiclass classification of cell painting images. The DeepPaint package and example notebooks are freely available at https://github.com/jhuapl-bio/DeepPaint.
Matching journals
The top 8 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Capturing cell heterogeneity in representations of cell populations for image-based profiling using contrastive learning 94%
- Unsupervised discovery of dynamic cell phenotypic states from transmitted light movies 93%
- Predicting drug polypharmacology from cell morphology readouts using variational autoencoder latent space arithmetic 93%
Similar papers in this journal
- Fully unsupervised deep mode of action learning for phenotyping high-content cellular images 93%
- Improving Deconvolution Methods in Biology through Open Innovation Competitions: An Application to the Connectivity Map 93%
- Domain-invariant features for mechanism of action prediction in a multi-cell-line drug screen 92%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.