Cell Morphology-Guided Small Molecule Generation with GFlowNets

TL;DR

This paper introduces a novel method for HCI-guided molecular design using GFlowNets and unsupervised multimodal embeddings, enabling the generation of molecules with phenotypic effects similar to target images without extensive labeled data.

Contribution

It proposes an innovative approach combining GFlowNets with unsupervised multimodal embeddings for phenotype-guided molecule generation, bypassing the need for labeled phenotypic data.

Findings

Generated molecules show high morphological similarity to targets

The method increases likelihood of similar biological activity

Effective in low-data, high-dimensional phenotypic spaces

Abstract

High-content phenotypic screening, including high-content imaging (HCI), has gained popularity in the last few years for its ability to characterize novel therapeutics without prior knowledge of the protein target. When combined with deep learning techniques to predict and represent molecular-phenotype interactions, these advancements hold the potential to significantly accelerate and enhance drug discovery applications. This work focuses on the novel task of HCI-guided molecular design. Generative models for molecule design could be guided by HCI data, for example with a supervised model that links molecules to phenotypes of interest as a reward function. However, limited labeled data, combined with the high-dimensional readouts, can make training these methods challenging and impractical. We consider an alternative approach in which we leverage an unsupervised multimodal joint embedding to define a latent similarity as a reward for GFlowNets. The proposed model learns to generate new molecules that could produce phenotypic effects similar to those of the given image target, without relying on pre-annotated phenotypic labels. We demonstrate that the proposed method generates molecules with high morphological and structural similarity to the target, increasing the likelihood of similar biological activity, as confirmed by an independent oracle model.