Systems-epigenomics inference of transcription factor activity implicates aryl-hydrocarbon-receptor inactivation as a key event in lung cancer development

Abstract

BackgroundDiverse molecular alterations associated with smoking in normal and precursor lung cancer cells have been reported, yet their role in lung cancer etiology remains unclear. A prominent example is hypomethylation of the aryl hydrocarbon-receptor repressor (AHRR) locus, which is observed in blood and squamous epithelial cells of smokers, but not in lung cancer.ResultsUsing a novel systems-epigenomics algorithm, called SEPIRA, which leverages the power of a large RNA-sequencing expression compendium to infer regulatory activity from messenger RNA expression or DNA methylation (DNAm) profiles, we infer the landscape of binding activity of lung-specific transcription factors (TFs) in lung carcinogenesis. We show that lung-specific TFs become preferentially inactivated in lung cancer and precursor lung cancer lesions and further demonstrate that these results can be derived using only DNAm data. We identify subsets of TFs which become inactivated in precursor cells. Among these regulatory factors, we identify AHR, the aryl hydrocarbon-receptor which controls a healthy immune response in the lung epithelium and whose repressor, AHRR, has recently been implicated in smoking-mediated lung cancer. In addition, we identify FOXJ1, a TF which promotes growth of airway cilia and effective clearance of the lung airway epithelium from carcinogens.ConclusionsWe identify TFs, such as AHR, which become inactivated in the earliest stages of lung cancer and which, unlike AHRR hypomethylation, are also inactivated in lung cancer itself. The novel systems-epigenomics algorithm SEPIRA will be useful to the wider epigenome-wide association study community as a means of inferring regulatory activity.Electronic supplementary materialThe online version of this article (doi:10.1186/s13059-017-1366-0) contains supplementary material, which is available to authorized users.

Ch IP ‐Atlas: a data‐mining suite powered by full integration of public Ch IP ‐seq data

“…In addition to the above‐mentioned analyses, ChIP‐Atlas Enrichment Analysis (formerly designated “ in silico ChIP”) has been used for various other purposes. For example, this tool has been applied to analyze TR enrichment at genomic ROIs such as expression quantitative trait loci (eQTLs), a user's own ChIP‐seq peak‐call data, and evolutionarily accelerated regions as well as genes whose expression levels are changed by drug exposure, aging, or cancer development (see http://chip-atlas.org/publications for the full list of publications citing ChIP‐Atlas). The results generated by ChIP‐Atlas are all assigned unique URLs (see ChIP‐Atlas document in https://github.com/inutano/chip-atlas/wiki for details) and are publicly available, and they are thus ready for sharing seamlessly among researchers for subsequent analysis in command lines and for interconnecting with other biodatabases such as the DeepBlue Epigenomic Data Server and RegulatorTrail , where ChIP‐Atlas data have been imported and subjected to analyses.…”

Integrative analysis of transcription factor occupancy at enhancers and disease risk loci in noncoding genomic regions

“…This tool is also able to accept a batch of gene symbols and can therefore discover TFs that collectively regulate the gene set of interest (e.g. Chatterjee et al 2018;Onodera et al 2017;Chen et al 2017).…”

Leveraging high-powered RNA-Seq datasets to improve inference of regulatory activity in single-cell RNA-Seq data

“…Because scRNA-Seq is noisy and subject to a high technical dropout rate 5 , and given a recent report which concluded that inferring regulatory relationships from scRNA-Seq is a highly suboptimal procedure 33 , we decided to approach this task differently, by first using a large-scale bulk RNA-Seq dataset to construct an approximate regulatory network. In particular, we here use the GTEX dataset 36 , which consists of bulk RNA-Seq profiles for 8555 samples encompassing 30 tissue types ( Methods ), to infer a regulatory network with our SEPIRA algorithm 37 ( Methods, Fig.1A ). SEPIRA uses a greedy partial correlation framework, similar in spirit to the state-of-the-art GENIE3 algorithm 35, 38 , to infer direct dependencies between regulators and downstream targets.…”

“…Using SEPIRA on the GTEX data, we constructed a lung-specific regulatory network, termed “LungNet”, which consists of 38 lung-specific TFs and a total of 1145 gene targets (range of targets per TF=10 to 152, mean=39.8) ( SI table.S2 , Methods ). We note that LungNet was already constructed and extensively validated by us previously 37 . Since by construction these 38 TFs are more highly expressed in lung tissue compared to all other tissue types and given their established role in lung tissue development 37 , most if not all of these ought to exhibit increased activation in the scRNA-Seq lung development study.…”