From In Silico to In Vitro: Evaluating Molecule Generative Models for Hit Generation

Hit identification is a critical yet resource-intensive step in the drug discovery pipeline, traditionally relying on high-throughput screening of large compound libraries. Despite advancements in virtual screening, these methods remain time-consuming and costly. Recent progress in deep learning has enabled the development of generative models capable of learning complex molecular representations and generating novel compounds \textit{de novo}. However, using ML to replace the entire drug-discovery pipeline is highly challenging. In this work, we rather investigate whether generative models can replace one step of the pipeline: \textit{hit-like} molecule generation. To the best of our knowledge, this is the first study to directly test this idea. Specifically, we investigate if such models can be trained to generate \textit{hit-like} molecules, enabling direct incorporation into, or even substitution of, traditional hit identification workflows. We propose an evaluation framework tailored to this task, integrating physicochemical, structural, and bioactivity-related criteria within a multi-stage filtering pipeline that defines the \textit{hit-like} chemical space. Two autoregressive and one diffusion-based generative models were benchmarked across various datasets and training settings, with outputs assessed using standard metrics and target-specific docking scores. Our results show that these models can generate valid, diverse, and biologically relevant compounds across multiple targets, with a few selected GSK-3$\beta$ hits synthesized and confirmed active in vitro. We also identify key limitations in current evaluation metrics and available training data.