Identification of drug responsive enhancers by predicting chromatin accessibility change from perturbed gene expression profiles.
NPJ Syst Biol Appl
; 10(1): 62, 2024 May 30.
Article
en En
| MEDLINE
| ID: mdl-38816426
ABSTRACT
Individual may response to drug treatment differently due to their genetic variants located in enhancers. These variants can alter transcription factor's (TF) binding strength, affect enhancer's chromatin activity or interaction, and eventually change expression level of downstream gene. Here, we propose a computational framework, PERD, to Predict the Enhancers Responsive to Drug. A machine learning model was trained to predict the genome-wide chromatin accessibility from transcriptome data using the paired expression and chromatin accessibility data collected from ENCODE and ROADMAP. Then the model was applied to the perturbed gene expression data from Connectivity Map (CMAP) and Cancer Drug-induced gene expression Signature DataBase (CDS-DB) and identify drug responsive enhancers with significantly altered chromatin accessibility. Furthermore, the drug responsive enhancers were related to the pharmacogenomics genome-wide association studies (PGx GWAS). Stepping on the traditional drug-associated gene signatures, PERD holds the promise to enhance the causality of drug perturbation by providing candidate regulatory element of those drug associated genes.
Texto completo:
1
Colección:
01-internacional
Banco de datos:
MEDLINE
Asunto principal:
Cromatina
/
Estudio de Asociación del Genoma Completo
/
Aprendizaje Automático
Límite:
Humans
Idioma:
En
Revista:
NPJ Syst Biol Appl
/
NPJ systems biology and applications
/
Npj syst. biol. appl
Año:
2024
Tipo del documento:
Article
País de afiliación:
China