Latent World Models of Cell Painting Data for In Silico Phenotypic Screening

David Scott Lewis, Enrique Zueco, Junhan Wang

Submission Type: I want my submission to be considered for both oral and poster presentation.

Keywords: Cell Painting, phenotypic screening, AI for biology, AI for chemistry, latent world models, in silico phenotypic screening, morphology profiling, high-throughput screening, drug discovery, vision transformers, transition models, molecular fingerprints, phenotype classification, JUMP Cell Painting, self-driving labs, agents and autonomous AI researchers, virtual screening, denoising latent dynamics, microscopy imaging, representation learning, multimodal biology, benchmark datasets, software pipelines, causal perturbation modeling

TL;DR: A latent world model for Cell Painting predicts how compounds shift cellular morphology, denoises noisy image-level variation, classifies phenotypes on unseen compounds, and points toward ultrafast, low-cost in silico phenotypic screening.

Confirmation Of Submission Requirements: I submit an abstract. It uses the template provided on the submission page and is no longer than 2 pages.

Submission Number: 403