Associating lncRNAs with small molecules via bilevel optimization reveals cancer-related lncRNAs

Long noncoding RNA (lncRNA) transcripts have emerging impacts in cancer studies, which suggests their potential as novel therapeutic agents. However, the molecular mechanism behind their treatment effects is still unclear. Here, we designed a computational model to Associate LncRNAs with Anti-Cancer...

Full description

Saved in:
Bibliographic Details
Published in:PLoS computational biology 2019-12, Vol.15 (12), p.e1007540-e1007540
Main Authors: Wang, Yongcui, Chen, Shilong, Chen, Luonan, Wang, Yong
Format: Article
Language:English
Subjects:
Algorithms
Animals
Biology
Biology and life sciences
Biomarkers
Biomarkers, Tumor - genetics
Biomarkers, Tumor - metabolism
Breast cancer
Breast Neoplasms - drug therapy
Breast Neoplasms - genetics
Breast Neoplasms - metabolism
Cancer
Chemical compounds
Computational Biology
Computer applications
Databases, Genetic
Drug development
Drugs
Engineering and Technology
Epigenetics
Evolution
Female
Gene expression
Gene Expression Profiling
Genetic engineering
Genomes
Genomics
Humans
Identification
Kaplan-Meier Estimate
Laboratories
Medicine and Health Sciences
Metastasis
Mice
Models, Genetic
Neoplasms - drug therapy
Neoplasms - genetics
Neoplasms - metabolism
Optimization
Pharmacology
Physical Sciences
RNA, Long Noncoding - genetics
RNA, Long Noncoding - metabolism
Signatures
Supervised Machine Learning
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Long noncoding RNA (lncRNA) transcripts have emerging impacts in cancer studies, which suggests their potential as novel therapeutic agents. However, the molecular mechanism behind their treatment effects is still unclear. Here, we designed a computational model to Associate LncRNAs with Anti-Cancer Drugs (ALACD) based on a bilevel optimization model, which optimized the gene signature overlap in the upper level and imputed the missing lncRNA-gene association in the lower level. ALACD predicts genes coexpressed with lncRNAs mean while matching drug's gene signatures. This model allows us to borrow the target gene information of small molecules to understand the mechanisms of action of lncRNAs and their roles in cancer. The ALACD model was systematically applied to the 10 cancer types in The Cancer Genome Atlas (TCGA) that had matched lncRNA and mRNA expression data. Cancer type-specific lncRNAs and associated drugs were identified. These lncRNAs show significantly different expression levels in cancer patients. Follow-up functional and molecular pathway analysis suggest the gene signatures bridging drugs and lncRNAs are closely related to cancer development. Importantly, patient survival information and evidence from the literature suggest that the lncRNAs and drug-lncRNA associations identified by the ALACD model can provide an alternative choice for cancer targeting treatment and potential cancer pognostic biomarkers. The ALACD model is freely available at https://github.com/wangyc82/ALACD-v1.
ISSN:1553-7358
1553-734X
1553-7358
DOI:10.1371/journal.pcbi.1007540