A functional module states framework reveals transcriptional states for drug and target prediction.

Guangrong Qin,Russell Moser,Theo A Knijnenburg,Raymond J Monnat,Ilya Shmulevich,Christopher J Kemp,David L Gibbs

doi:10.1016/j.celrep.2021.110269

Guangrong Qin, Russell Moser + Show 5 more

Open Access

https://doi.org/10.1016/j.celrep.2021.110269

Copy DOI

Abstract

Cells are complex systems in which many functions are performed by different genetically defined and encoded functional modules. To systematically understand how these modules respond to drug or genetic perturbations, we develop a functional module states framework. Using this framework, we (1) define the drug-induced transcriptional state space for breast cancer cell lines using large public gene expression datasets and reveal that the transcriptional states are associated with drug concentration and drug targets, (2) identify potential targetable vulnerabilities through integrative analysis of transcriptional states after drug treatment and gene knockdown-associated cancer dependency, and (3) use functional module states to predict transcriptional state-dependent drug sensitivity and build prediction models for drug response. This approach demonstrates a similar prediction performance as approaches using high-dimensional gene expression values, with the added advantage of more clearly revealing biologically relevant transcriptional states and key regulators.

Highlights

Cells are complex systems that have been investigated at multiple levels, including genomic, epigenomic, transcriptomic, proteomic, and metabolomic levels
We proposed that cell states or the average transcriptional states of samples can be defined by the overall activity profile of the functional module (FM), which we define as FM factors
In the remainder of this work, we demonstrate the utility of the newly established FM states framework by addressing the following questions: (1) are the transcriptomic states of a cancer cell line (MCF7) after drug treatment associated with drug concentration or drug targets? (2) Can we predict targetable vulnerabilities using the transcriptional cell states? (3) Can we use the functional states of cancer cell lines prior to drug treatment to predict the drug response? As an additional use demonstration, we applied it to samples from individuals with acute myeloid leukemia (AML) to understand how the gene alterations or clinical features are associated with transcriptional states and how these transcription states are associated with drug response

Summary

Introduction

Cells are complex systems that have been investigated at multiple levels, including genomic, epigenomic, transcriptomic, proteomic, and metabolomic levels. Many methods have been developed to measure the transcriptome at different levels of resolution, such as gene expression microarrays, bulk RNA sequencing (RNA-seq), the L1000 platform (Subramanian et al, 2017), and single-cell RNA-seq (Zheng et al, 2017). These high-throughput techniques have been used to capture transcriptomes from thousands of primary tumor samples and cell lines. The Cancer Genome Atlas (TCGA) (Hutter and Zenklusen, 2018; Knijnenburg et al, 2018; Thorsson et al, 2018) project measured gene expression on more than 10,000 tumor samples for 33 tumor types using

Methods

Results

Discussion

Conclusion