Abstract:Deciphering cellular responses to genetic perturbations is fundamental for a wide array of biomedical applications, ranging from uncovering gene roles and interactions to unraveling effective therapeutics. Accurately predicting the transcriptional outcomes of genetic perturbations is indispensable for optimizing experimental perturbations and deciphering cellular response mechanisms; however, three scenarios present principal challenges, i.e., predicting single genetic perturbation outcomes, predicting multiple genetic perturbation outcomes and predicting genetic outcomes across cell lines. In this study, we introduce ub sk decomposition odeling for genetic erturbation prediction (STAMP), a conceptually novel computational strategy for genetic perturbation outcome prediction and downstream applications. STAMP innovatively formulates genetic perturbation prediction as a subtask decomposition (STD) problem by resolving three progressive subtasks in a divide-and-conquer manner, i.e., identifying differentially expressed gene (DEG) postperturbations, determining the regulatory directions of DEGs and finally estimating the magnitudes of gene expression changes. In addition to facilitating perturbation prediction, STAMP also serves as a robust and generalizable benchmark guide for evaluating various genetic perturbation prediction models. As a result, STAMP exhibits a substantial improvement in terms of its genetic perturbation prediction ability over the existing approaches on three subtasks and beyond, including revealing the ability to identify key regulatory genes and pathways on small samples and to reveal precise genetic interactions. Overall, STAMP serves as a fundamentally novel and effective prediction and generalizable benchmarking strategy that can facilitate genetic perturbation prediction, guide the design of perturbation experiments, and broaden the understanding of perturbation mechanisms.

Toward subtask-decomposition-based learning and benchmarking for predicting genetic perturbation outcomes and beyond

Toward subtask decomposition-based learning and benchmarking for genetic perturbation outcome prediction and beyond

Predicting cell morphological responses to perturbations using generative modeling

Predicting Cellular Responses to Complex Perturbations in High-Throughput Screens.

Learning interpretable cellular responses to complex perturbations in high-throughput screens

AttentionPert: Accurately Modeling Multiplexed Genetic Perturbations with Multi-scale Effects

Modeling and predicting single-cell multi-gene perturbation responses with scLAMBDA

Scgen Predicts Single-Cell Perturbation Responses

Machine learning for perturbational single-cell omics

Compositional perturbation autoencoder for single-cell response modeling

Benchmarking Transcriptomics Foundation Models for Perturbation Analysis : one PCA still rules them all

PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis

A genome-scale deep learning model to predict gene expression changes of genetic perturbations from multiplex biological networks

Interpretable spatially aware dimension reduction of spatial transcriptomics with STAMP

Inference of heterogeneous effects in single-cell genetic perturbation screens

Refining computational inference of gene regulatory networks: integrating knockout data within a multi-task framework

Representing high throughput expression profiles via perturbation barcodes reveals compound targets

Improving drug response prediction via integrating gene relationships with deep learning

GenePert: Leveraging GenePT Embeddings for Gene Perturbation Prediction