Estimating genome-wide regulatory activity from multi-omics data sets using mathematical optimization

Gene essentiality in cancer is better predicted by mRNA abundance than by gene regulatory network-inferred activity

Gene regulatory networks (GRNs) are often deregulated in tumor cells, resulting in altered transcriptional programs that facilitate tumor growth. These altered networks may make tumor cells vulnerable to the inhibition of specific regulatory proteins. Consequently, the reconstruction of GRNs in tumors is often proposed as a means to identify therapeutic targets. While there are examples of individual targets identified using GRNs, the extent to which GRNs can be used to predict sensitivity to targeted intervention in general remains unknown. Here we use the results of genome-wide CRISPR screens to systematically assess the ability of GRNs to predict sensitivity to gene inhibition in cancer cell lines. Using GRNs derived from multiple sources, including GRNs reconstructed from tumor transcriptomes and from curated databases, we infer regulatory gene activity in cancer cell lines from ten cancer types. We then ask, in each cancer type, if the inferred regulatory activity of each gene is predictive of sensitivity to CRISPR perturbation of that gene. We observe slight variation in the correlation between gene regulatory activity and gene sensitivity depending on the source of the GRN and the activity estimation method used. However, we find that there is consistently a stronger relationship between mRNA abundance and gene sensitivity than there is between regulatory gene activity and gene sensitivity. This is true both when gene sensitivity is treated as a binary and a quantitative property. Overall, our results suggest that gene sensitivity is better predicted by measured expression than by GRN-inferred activity.

Identifying genetic regulatory variants that affect transcription factor activity

Genetic variants affecting gene expression levels in humans have been mapped in the Genotype-Tissue Expression (GTEx) project. Trans-acting variants impacting many genes simultaneously through a shared transcription factor (TF) are of particular interest. Here, we developed a generalized linear model (GLM) to estimate protein-level TF activity levels in an individual-specific manner from GTEx RNA sequencing (RNA-seq) profiles. It uses observed differential gene expression after TF perturbation as a predictor and, by analyzing differential expression within pairs of neighboring genes, controls for the confounding effect of variation in chromatin state along the genome. We inferred genotype-specific activities for 55 TFs across 49 tissues. Subsequently performing genome-wide association analysis on this virtual trait revealed TF activity quantitative trait loci (aQTLs) that, as a set, are enriched for functional features. Altogether, the set of tools we introduce here highlights the potential of genetic association studies for cellular endophenotypes based on a network-based multi-omics approach. The transparent peer review record is available.

Characterization of cell-fate decision landscapes by estimating transcription factor dynamics

Time-specific modulation of gene expression during differentiation by transcription factors promotes cell diversity. However, estimating their dynamic regulatory activity at the single-cell level and in a high-throughput manner remains challenging. We present FateCompass, an integrative approach that utilizes single-cell transcriptomics data to identify lineage-specific transcription factors throughout differentiation. By combining a probabilistic framework with RNA velocities or differentiation potential, we estimate transition probabilities, while a linear model of gene regulation is employed to compute transcription factor activities. Considering dynamic changes and correlations of expression and activities, FateCompass identifies lineage-specific regulators. Our validation using in silico data and application to pancreatic endocrine cell differentiation datasets highlight both known and potentially novel lineage-specific regulators. Notably, we uncovered undescribed transcription factors of an enterochromaffin-like population during in vitro differentiation toward ß-like cells. FateCompass provides a valuable framework for hypothesis generation, advancing our understanding of the gene regulatory networks driving cell-fate decisions.

A Bayesian Noisy Logic Model for Inference of Transcription Factor Activity from Single Cell and Bulk Transcriptomic Data

The advent of high-throughput sequencing has made it possible to measure the expression of genes at relatively low cost. However, direct measurement of regulatory mechanisms, such as Transcription Factor (TF) activity is still not readily feasible in a high-throughput manner. Consequently, there is a need for computational approaches that can reliably estimate regulator activity from observable gene expression data. In this work, we present a noisy Boolean logic Bayesian model for TF activity inference from differential gene expression data and causal graphs. Our approach provides a flexible framework to incorporate biologically motivated TF-gene regulation logic models. Using simulations and controlled over-expression experiments in cell cultures, we demonstrate that our method can accurately identify TF activity. Moreover, we apply our method to bulk and single cell transcriptomics measurements to investigate transcriptional regulation of fibroblast phenotypic plasticity. Finally, to facilitate usage, we provide user-friendly software packages and a web-interface to query TF activity from user input differential gene expression data: https://umbibio.math.umb.edu/nlbayes/.

Profiling transcription factor activity dynamics using intronic reads in time-series transcriptome data

Activities of transcription factors (TFs) are temporally modulated to regulate dynamic cellular processes, including development, homeostasis, and disease. Recent developments of bioinformatic tools have enabled the analysis of TF activities using transcriptome data. However, because these methods typically use exon-based target expression levels, the estimated TF activities have limited temporal accuracy. To address this, we proposed a TF activity measure based on intron-level information in time-series RNA-seq data, and implemented it to decode the temporal control of TF activities during dynamic processes. We showed that TF activities inferred from intronic reads can better recapitulate instantaneous TF activities compared to the exon-based measure. By analyzing public and our own time-series transcriptome data, we found that intron-based TF activities improve the characterization of temporal phasing of cycling TFs during circadian rhythm, and facilitate the discovery of two temporally opposing TF modules during T cell activation. Collectively, we anticipate that the proposed approach would be broadly applicable for decoding global transcriptional architecture during dynamic processes.

Genetic interactions derived from high-throughput phenotyping of 6589 yeast cell cycle mutants

Over the last 30 years, computational biologists have developed increasingly realistic mathematical models of the regulatory networks controlling the division of eukaryotic cells. These models capture data resulting from two complementary experimental approaches: low-throughput experiments aimed at extensively characterizing the functions of small numbers of genes, and large-scale genetic interaction screens that provide a systems-level perspective on the cell division process. The former is insufficient to capture the interconnectivity of the genetic control network, while the latter is fraught with irreproducibility issues. Here, we describe a hybrid approach in which the 630 genetic interactions between 36 cell-cycle genes are quantitatively estimated by high-throughput phenotyping with an unprecedented number of biological replicates. Using this approach, we identify a subset of high-confidence genetic interactions, which we use to refine a previously published mathematical model of the cell cycle. We also present a quantitative dataset of the growth rate of these mutants under six different media conditions in order to inform future cell cycle models.

Robustness and applicability of transcription factor and pathway analysis tools on single-cell RNA-seq data

BACKGROUND

Many functional analysis tools have been developed to extract functional and mechanistic insight from bulk transcriptome data. With the advent of single-cell RNA sequencing (scRNA-seq), it is in principle possible to do such an analysis for single cells. However, scRNA-seq data has characteristics such as drop-out events and low library sizes. It is thus not clear if functional TF and pathway analysis tools established for bulk sequencing can be applied to scRNA-seq in a meaningful way.

RESULTS

To address this question, we perform benchmark studies on simulated and real scRNA-seq data. We include the bulk-RNA tools PROGENy, GO enrichment, and DoRothEA that estimate pathway and transcription factor (TF) activities, respectively, and compare them against the tools SCENIC/AUCell and metaVIPER, designed for scRNA-seq. For the in silico study, we simulate single cells from TF/pathway perturbation bulk RNA-seq experiments. We complement the simulated data with real scRNA-seq data upon CRISPR-mediated knock-out. Our benchmarks on simulated and real data reveal comparable performance to the original bulk data. Additionally, we show that the TF and pathway activities preserve cell type-specific variability by analyzing a mixture sample sequenced with 13 scRNA-seq protocols. We also provide the benchmark data for further use by the community.

CONCLUSIONS

Our analyses suggest that bulk-based functional analysis tools that use manually curated footprint gene sets can be applied to scRNA-seq data, partially outperforming dedicated single-cell tools. Furthermore, we find that the performance of functional analysis tools is more sensitive to the gene sets than to the statistic used.

Integration of multi-omics data to mine cancer-related gene modules

The identification of cancer-related genes is a major research goal, with implications for determining the pathogenesis of cancer and identifying biomarkers for early diagnosis and treatment. In this study, by integrating multi-omics data, including gene expression, DNA copy number variation, DNA methylation, transcription factors, miRNA, and lncRNA data, we propose a method for mining cancer-related genes based on network models. First, using random forest-based feature selection method multi-omics data are integrated to identify key regulatory factors that affect gene expression, and then genome-wide regulatory networks are constructed. Next, by comparing the regulatory networks of key candidate genes in variant samples and non-variant samples, a differential expression regulatory network is generated. The differential network contains a collection of abnormal regulatory genes of key candidate genes. Then, by introducing the functional similarity as a distance metric for gene sets, a density-based clustering method is used to mine gene modules related to cancer. We applied this method to LUSC (lung squamous cell carcinoma) and mined cancer-related gene modules composed of 20 genes. GO function and KEGG pathway analyses indicated that the modules were closely related to cancer. A survival analysis was used to verify that the excavated gene modules can effectively distinguish between high- and low-risk groups. Overall, these results suggest that the proposed method can be used to identify cancer-related gene modules, providing a basis for the development of biomarkers for diagnosis and treatment.

Benchmark and integration of resources for the estimation of human transcription factor activities

The prediction of transcription factor (TF) activities from the gene expression of their targets (i.e., TF regulon) is becoming a widely used approach to characterize the functional status of transcriptional regulatory circuits. Several strategies and data sets have been proposed to link the target genes likely regulated by a TF, each one providing a different level of evidence. The most established ones are (1) manually curated repositories, (2) interactions derived from ChIP-seq binding data, (3) in silico prediction of TF binding on gene promoters, and (4) reverse-engineered regulons from large gene expression data sets. However, it is not known how these different sources of regulons affect the TF activity estimations and, thereby, downstream analysis and interpretation. Here we compared the accuracy and biases of these strategies to define human TF regulons by means of their ability to predict changes in TF activities in three reference benchmark data sets. We assembled a collection of TF-target interactions for 1541 human TFs and evaluated how different molecular and regulatory properties of the TFs, such as the DNA-binding domain, specificities, or mode of interaction with the chromatin, affect the predictions of TF activity. We assessed their coverage and found little overlap on the regulons derived from each strategy and better performance by literature-curated information followed by ChIP-seq data. We provide an integrated resource of all TF-target interactions derived through these strategies, with confidence scores, as a resource for enhanced prediction of TF activities.

Estimation of Transcription Factor Activity in Knockdown Studies

Numerous methods have been developed trying to infer actual regulatory events in a sample. A prominent class of methods model genome-wide gene expression as linear equations derived from a transcription factor (TF) – gene network and optimizes parameters to fit the measured expression intensities. We apply four such methods on experiments with a TF-knockdown (KD) in human and E. coli. The transcriptome data provides clear expression signals and thus represents an extremely favorable test setting. The methods estimate activity changes of all TFs, which we expect to be highest in the KD TF. However, only in 15 out of 54 cases, the KD TFs ranked in the top 5%. We show that this poor overall performance cannot be attributed to a low effectiveness of the knockdown or the specific regulatory network provided as background knowledge. Further, the ranks of regulators related to the KD TF by the network or pathway are not significantly different from a random selection. In general, the result overlaps of different methods are small, indicating that they draw very different conclusions when presented with the same, presumably simple, inference problem. These results show that the investigated methods cannot yield robust TF activity estimates in knockdown schemes.

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

Background

Diverse 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.

Results

Using 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.

Conclusions

We 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 material

The online version of this article (doi:10.1186/s13059-017-1366-0) contains supplementary material, which is available to authorized users.