Multi-study inference of regulatory networks for more accurate models of gene regulation

Prediction of Composite Clinical Outcomes for Childhood Neuroblastoma Using Multi-Omics Data and Machine Learning

Neuroblastoma is a common malignant tumor in childhood that seriously endangers the health and lives of children, making it essential to find effective prognostic markers to accurately predict their clinical outcomes. The development of high-throughput technology in the biomedical field has made it possible to obtain multi-omics data, whose integration can compensate for missing or unreliable information in a single data source. In this study, we integrated clinical data and two omics data, i.e., gene expression and DNA methylation data, to study the prognosis of neuroblastoma. Since the features in omics data are redundant, it is crucial to conduct feature selection on them. We proposed a two-step feature selection (TSFS) method to quickly and accurately select the optimal features, where the first step aims at selecting candidate features and the second step is to remove redundant features among them using our proposed maximal association coefficient (MAC). Our goal is to predict composite clinical outcomes for neuroblastoma patients, i.e., their survival time and vital status at the last follow-up, which was validated to be two inter-correlated tasks. We conducted a series of experiments and evaluated the experimental results using accuracy and AUC (area under the ROC curve) evaluation metrics, which indicated that by the combination of the integration of the three types of data, our proposed TSFS method and a multi-task learning method can synergistically improve the reliability and accuracy of the prediction models.

Learning the cellular activity representation based on gene regulatory networks for prediction of tumor response to drugs

Predicting the response of tumor cells to anti-tumor drugs is critical to realizing cancer precision medicine. Currently, most existing methods ignore the regulatory relationships between genes and thus have unsatisfactory predictive performance. In this paper, we propose to predict anti-tumor drug efficacy via learning the activity representation of tumor cells based on a priori knowledge of gene regulation networks (GRNs). Specifically, the method simulates the cellular biosystem by synthesizing a cell-gene activity network and then infers a new low-dimensional activity representation for tumor cells from the raw high-dimensional expression profile. The simulated cell-gene network mainly comprises known gene regulatory networks collected from multiple resources and fuses tumor cells by linking them to hotspot genes that are over- or under-expressed in them. The resulting activity representation could not only reflect the shallow expression profile (hotspot genes) but also mines in-depth information of gene regulation activity in tumor cells before treatment. Finally, we build deep learning models on the activity representation for predicting drug efficacy in tumor cells. Experimental results on the benchmark GDSC dataset demonstrate the superior performance of the proposed method over SOTA methods with the highest AUC of 0.954 in the efficacy label prediction and the best R2 of 0.834 in the regression of half maximal inhibitory concentration (IC50) values, suggesting the potential value of the proposed method in practice.

Flexible modeling of regulatory networks improves transcription factor activity estimation

Transcriptional regulation plays a crucial role in determining cell fate and disease, yet inferring the key regulators from gene expression data remains a significant challenge. Existing methods for estimating transcription factor (TF) activity often rely on static TF-gene interaction databases and cannot adapt to changes in regulatory mechanisms across different cell types and disease conditions. Here, we present a new algorithm - Transcriptional Inference using Gene Expression and Regulatory data (TIGER) - that overcomes these limitations by flexibly modeling activation and inhibition events, up-weighting essential edges, shrinking irrelevant edges towards zero through a sparse Bayesian prior, and simultaneously estimating both TF activity levels and changes in the underlying regulatory network. When applied to yeast and cancer TF knock-out datasets, TIGER outperforms comparable methods in terms of prediction accuracy. Moreover, our application of TIGER to tissue- and cell-type-specific RNA-seq data demonstrates its ability to uncover differences in regulatory mechanisms. Collectively, our findings highlight the utility of modeling context-specific regulation when inferring transcription factor activities.

DeepGRNCS: deep learning-based framework for jointly inferring gene regulatory networks across cell subpopulations

Inferring gene regulatory networks (GRNs) allows us to obtain a deeper understanding of cellular function and disease pathogenesis. Recent advances in single-cell RNA sequencing (scRNA-seq) technology have improved the accuracy of GRN inference. However, many methods for inferring individual GRNs from scRNA-seq data are limited because they overlook intercellular heterogeneity and similarities between different cell subpopulations, which are often present in the data. Here, we propose a deep learning-based framework, DeepGRNCS, for jointly inferring GRNs across cell subpopulations. We follow the commonly accepted hypothesis that the expression of a target gene can be predicted based on the expression of transcription factors (TFs) due to underlying regulatory relationships. We initially processed scRNA-seq data by discretizing data scattering using the equal-width method. Then, we trained deep learning models to predict target gene expression from TFs. By individually removing each TF from the expression matrix, we used pre-trained deep model predictions to infer regulatory relationships between TFs and genes, thereby constructing the GRN. Our method outperforms existing GRN inference methods for various simulated and real scRNA-seq datasets. Finally, we applied DeepGRNCS to non-small cell lung cancer scRNA-seq data to identify key genes in each cell subpopulation and analyzed their biological relevance. In conclusion, DeepGRNCS effectively predicts cell subpopulation-specific GRNs. The source code is available at https://github.com/Nastume777/DeepGRNCS.

Structure-primed embedding on the transcription factor manifold enables transparent model architectures for gene regulatory network and latent activity inference

BACKGROUND

Modeling of gene regulatory networks (GRNs) is limited due to a lack of direct measurements of genome-wide transcription factor activity (TFA) making it difficult to separate covariance and regulatory interactions. Inference of regulatory interactions and TFA requires aggregation of complementary evidence. Estimating TFA explicitly is problematic as it disconnects GRN inference and TFA estimation and is unable to account for, for example, contextual transcription factor-transcription factor interactions, and other higher order features. Deep-learning offers a potential solution, as it can model complex interactions and higher-order latent features, although does not provide interpretable models and latent features.

RESULTS

We propose a novel autoencoder-based framework, StrUcture Primed Inference of Regulation using latent Factor ACTivity (SupirFactor) for modeling, and a metric, explained relative variance (ERV), for interpretation of GRNs. We evaluate SupirFactor with ERV in a wide set of contexts. Compared to current state-of-the-art GRN inference methods, SupirFactor performs favorably. We evaluate latent feature activity as an estimate of TFA and biological function in S. cerevisiae as well as in peripheral blood mononuclear cells (PBMC).

CONCLUSION

Here we present a framework for structure-primed inference and interpretation of GRNs, SupirFactor, demonstrating interpretability using ERV in multiple biological and experimental settings. SupirFactor enables TFA estimation and pathway analysis using latent factor activity, demonstrated here on two large-scale single-cell datasets, modeling S. cerevisiae and PBMC. We find that the SupirFactor model facilitates biological analysis acquiring novel functional and regulatory insight.

iLSGRN: inference of large-scale gene regulatory networks based on multi-model fusion

MOTIVATION

Gene regulatory networks (GRNs) are a way of describing the interaction between genes, which contribute to revealing the different biological mechanisms in the cell. Reconstructing GRNs based on gene expression data has been a central computational problem in systems biology. However, due to the high dimensionality and non-linearity of large-scale GRNs, accurately and efficiently inferring GRNs is still a challenging task.

RESULTS

In this article, we propose a new approach, iLSGRN, to reconstruct large-scale GRNs from steady-state and time-series gene expression data based on non-linear ordinary differential equations. Firstly, the regulatory gene recognition algorithm calculates the Maximal Information Coefficient between genes and excludes redundant regulatory relationships to achieve dimensionality reduction. Then, the feature fusion algorithm constructs a model leveraging the feature importance derived from XGBoost (eXtreme Gradient Boosting) and RF (Random Forest) models, which can effectively train the non-linear ordinary differential equations model of GRNs and improve the accuracy and stability of the inference algorithm. The extensive experiments on different scale datasets show that our method makes sensible improvement compared with the state-of-the-art methods. Furthermore, we perform cross-validation experiments on the real gene datasets to validate the robustness and effectiveness of the proposed method.

AVAILABILITY AND IMPLEMENTATION

The proposed method is written in the Python language, and is available at: https://github.com/lab319/iLSGRN.

MCPNet: a parallel maximum capacity-based genome-scale gene network construction framework

MOTIVATION

Gene network reconstruction from gene expression profiles is a compute- and data-intensive problem. Numerous methods based on diverse approaches including mutual information, random forests, Bayesian networks, correlation measures, as well as their transforms and filters such as data processing inequality, have been proposed. However, an effective gene network reconstruction method that performs well in all three aspects of computational efficiency, data size scalability, and output quality remains elusive. Simple techniques such as Pearson correlation are fast to compute but ignore indirect interactions, while more robust methods such as Bayesian networks are prohibitively time consuming to apply to tens of thousands of genes.

RESULTS

We developed maximum capacity path (MCP) score, a novel maximum-capacity-path-based metric to quantify the relative strengths of direct and indirect gene-gene interactions. We further present MCPNet, an efficient, parallelized gene network reconstruction software based on MCP score, to reverse engineer networks in unsupervised and ensemble manners. Using synthetic and real Saccharomyces cervisiae datasets as well as real Arabidopsis thaliana datasets, we demonstrate that MCPNet produces better quality networks as measured by AUPRC, is significantly faster than all other gene network reconstruction software, and also scales well to tens of thousands of genes and hundreds of CPU cores. Thus, MCPNet represents a new gene network reconstruction tool that simultaneously achieves quality, performance, and scalability requirements.

AVAILABILITY AND IMPLEMENTATION

Source code freely available for download at https://doi.org/10.5281/zenodo.6499747 and https://github.com/AluruLab/MCPNet, implemented in C++ and supported on Linux.

Inference of cell type-specific gene regulatory networks on cell lineages from single cell omic datasets

Cell type-specific gene expression patterns are outputs of transcriptional gene regulatory networks (GRNs) that connect transcription factors and signaling proteins to target genes. Single-cell technologies such as single cell RNA-sequencing (scRNA-seq) and single cell Assay for Transposase-Accessible Chromatin using sequencing (scATAC-seq), can examine cell-type specific gene regulation at unprecedented detail. However, current approaches to infer cell type-specific GRNs are limited in their ability to integrate scRNA-seq and scATAC-seq measurements and to model network dynamics on a cell lineage. To address this challenge, we have developed single-cell Multi-Task Network Inference (scMTNI), a multi-task learning framework to infer the GRN for each cell type on a lineage from scRNA-seq and scATAC-seq data. Using simulated and real datasets, we show that scMTNI is a broadly applicable framework for linear and branching lineages that accurately infers GRN dynamics and identifies key regulators of fate transitions for diverse processes such as cellular reprogramming and differentiation.

Identifying strengths and weaknesses of methods for computational network inference from single-cell RNA-seq data

Single-cell RNA-sequencing (scRNA-seq) offers unparalleled insight into the transcriptional programs of different cellular states by measuring the transcriptome of thousands of individual cells. An emerging problem in the analysis of scRNA-seq is the inference of transcriptional gene regulatory networks and a number of methods with different learning frameworks have been developed to address this problem. Here, we present an expanded benchmarking study of eleven recent network inference methods on seven published scRNA-seq datasets in human, mouse, and yeast considering different types of gold standard networks and evaluation metrics. We evaluate methods based on their computing requirements as well as on their ability to recover the network structure. We find that, while most methods have a modest recovery of experimentally derived interactions based on global metrics such as Area Under the Precision Recall curve, methods are able to capture targets of regulators that are relevant to the system under study. Among the top performing methods that use only expression were SCENIC, PIDC, MERLIN or Correlation. Addition of prior biological knowledge and the estimation of transcription factor activities resulted in the best overall performance with the Inferelator and MERLIN methods that use prior knowledge outperforming methods that use expression alone. We found that imputation for network inference did not improve network inference accuracy and could be detrimental. Comparisons of inferred networks for comparable bulk conditions showed that the networks inferred from scRNA-seq datasets are often better or at par with the networks inferred from bulk datasets. Our analysis should be beneficial in selecting methods for network inference. At the same time, this highlights the need for improved methods and better gold standards for regulatory network inference from scRNAseq datasets.

Structure primed embedding on the transcription factor manifold enables transparent model architectures for gene regulatory network and latent activity inference

The modeling of gene regulatory networks (GRNs) is limited due to a lack of direct measurements of regulatory features in genome-wide screens. Most GRN inference methods are therefore forced to model relationships between regulatory genes and their targets with expression as a proxy for the upstream independent features, complicating validation and predictions produced by modeling frameworks. Separating covariance and regulatory influence requires aggregation of independent and complementary sets of evidence, such as transcription factor (TF) binding and target gene expression. However, the complete regulatory state of the system, e.g. TF activity (TFA) is unknown due to a lack of experimental feasibility, making regulatory relations difficult to infer. Some methods attempt to account for this by modeling TFA as a latent feature, but these models often use linear frameworks that are unable to account for non-linearities such as saturation, TF-TF interactions, and other higher order features. Deep learning frameworks may offer a solution, as they are capable of modeling complex interactions and capturing higher-order latent features. However, these methods often discard central concepts in biological systems modeling, such as sparsity and latent feature interpretability, in favor of increased model complexity. We propose a novel deep learning autoencoder-based framework, StrUcture Primed Inference of Regulation using latent Factor ACTivity (SupirFactor), that scales to single cell genomic data and maintains interpretability to perform GRN inference and estimate TFA as a latent feature. We demonstrate that SupirFactor outperforms current leading GRN inference methods, predicts biologically relevant TFA and elucidates functional regulatory pathways through aggregation of TFs.

Learning Cell Annotation under Multiple Reference Datasets by Multisource Domain Adaptation

Accurate and efficient cell type annotation is essential for single-cell sequence analysis. Currently, cell type annotation using well-annotated reference datasets with powerful models has become increasingly popular. However, with the increasing amount of single-cell data, there is an urgent need to develop a novel annotation method that can integrate multiple reference datasets to improve cell type annotation performance. Since the unwanted batch effects between individual reference datasets, integrating multiple reference datasets is still an open challenge. To address this, we proposed scMDR and scMultiR, respectively, using multisource domain adaptation to learn cell type-specific information from multiple reference datasets and query cells. Based on the learned cell type-specific information, scMDR and scMultiR provide the most likely cell types for the query cells. Benchmark experiments demonstrated their state-of-the-art effectiveness for integrative single-cell assignment with multiple reference datasets.

Improving gene regulatory network inference and assessment: The importance of using network structure

Gene regulatory networks are graph models representing cellular transcription events. Networks are far from complete due to time and resource consumption for experimental validation and curation of the interactions. Previous assessments have shown the modest performance of the available network inference methods based on gene expression data. Here, we study several caveats on the inference of regulatory networks and methods assessment through the quality of the input data and gold standard, and the assessment approach with a focus on the global structure of the network. We used synthetic and biological data for the predictions and experimentally-validated biological networks as the gold standard (ground truth). Standard performance metrics and graph structural properties suggest that methods inferring co-expression networks should no longer be assessed equally with those inferring regulatory interactions. While methods inferring regulatory interactions perform better in global regulatory network inference than co-expression-based methods, the latter is better suited to infer function-specific regulons and co-regulation networks. When merging expression data, the size increase should outweigh the noise inclusion and graph structure should be considered when integrating the inferences. We conclude with guidelines to take advantage of inference methods and their assessment based on the applications and available expression datasets.

Coordinated control of neuronal differentiation and wiring by sustained transcription factors

The large diversity of cell types in nervous systems presents a challenge in identifying the genetic mechanisms that encode it. Here, we report that nearly 200 distinct neurons in the Drosophila visual system can each be defined by unique combinations of on average 10 continuously expressed transcription factors. We show that targeted modifications of this terminal selector code induce predictable conversions of neuronal fates that appear morphologically and transcriptionally complete. Cis-regulatory analysis of open chromatin links one of these genes to an upstream patterning factor that specifies neuronal fates in stem cells. Experimentally validated network models describe the synergistic regulation of downstream effectors by terminal selectors and ecdysone signaling during brain wiring. Our results provide a generalizable framework of how specific fates are implemented in postmitotic neurons.

Network-based approaches for modeling disease regulation and progression

Molecular interaction networks lay the foundation for studying how biological functions are controlled by the complex interplay of genes and proteins. Investigating perturbed processes using biological networks has been instrumental in uncovering mechanisms that underlie complex disease phenotypes. Rapid advances in omics technologies have prompted the generation of high-throughput datasets, enabling large-scale, network-based analyses. Consequently, various modeling techniques, including network enrichment, differential network extraction, and network inference, have proven to be useful for gaining new mechanistic insights. We provide an overview of recent network-based methods and their core ideas to facilitate the discovery of disease modules or candidate mechanisms. Knowledge generated from these computational efforts will benefit biomedical research, especially drug development and precision medicine. We further discuss current challenges and provide perspectives in the field, highlighting the need for more integrative and dynamic network approaches to model disease development and progression.

Open problems in mathematical biology

Biology is data-rich, and it is equally rich in concepts and hypotheses. Part of trying to understand biological processes and systems is therefore to confront our ideas and hypotheses with data using statistical methods to determine the extent to which our hypotheses agree with reality. But doing so in a systematic way is becoming increasingly challenging as our hypotheses become more detailed, and our data becomes more complex. Mathematical methods are therefore gaining in importance across the life- and biomedical sciences. Mathematical models allow us to test our understanding, make testable predictions about future behaviour, and gain insights into how we can control the behaviour of biological systems. It has been argued that mathematical methods can be of great benefit to biologists to make sense of data. But mathematics and mathematicians are set to benefit equally from considering the often bewildering complexity inherent to living systems. Here we present a small selection of open problems and challenges in mathematical biology. We have chosen these open problems because they are of both biological and mathematical interest.

Inferring Gene Regulatory Networks From Single-Cell Transcriptomic Data Using Bidirectional RNN

Accurate inference of gene regulatory rules is critical to understanding cellular processes. Existing computational methods usually decompose the inference of gene regulatory networks (GRNs) into multiple subproblems, rather than detecting potential causal relationships simultaneously, which limits the application to data with a small number of genes. Here, we propose BiRGRN, a novel computational algorithm for inferring GRNs from time-series single-cell RNA-seq (scRNA-seq) data. BiRGRN utilizes a bidirectional recurrent neural network to infer GRNs. The recurrent neural network is a complex deep neural network that can capture complex, non-linear, and dynamic relationships among variables. It maps neurons to genes, and maps the connections between neural network layers to the regulatory relationship between genes, providing an intuitive solution to model GRNs with biological closeness and mathematical flexibility. Based on the deep network, we transform the inference of GRNs into a regression problem, using the gene expression data at previous time points to predict the gene expression data at the later time point. Furthermore, we adopt two strategies to improve the accuracy and stability of the algorithm. Specifically, we utilize a bidirectional structure to integrate the forward and reverse inference results and exploit an incomplete set of prior knowledge to filter out some candidate inferences of low confidence. BiRGRN is applied to four simulated datasets and three real scRNA-seq datasets to verify the proposed method. We perform comprehensive comparisons between our proposed method with other state-of-the-art techniques. These experimental results indicate that BiRGRN is capable of inferring GRN simultaneously from time-series scRNA-seq data. Our method BiRGRN is implemented in Python using the TensorFlow machine-learning library, and it is freely available at https://gitee.com/DHUDBLab/bi-rgrn.

System Principles Governing the Organization, Architecture, Dynamics, and Evolution of Gene Regulatory Networks

Synthetic biology aims to apply engineering principles for the rational, systematical design and construction of biological systems displaying functions that do not exist in nature or even building a cell from scratch. Understanding how molecular entities interconnect, work, and evolve in an organism is pivotal to this aim. Here, we summarize and discuss some historical organizing principles identified in bacterial gene regulatory networks. We propose a new layer, the concilion, which is the group of structural genes and their local regulators responsible for a single function that, organized hierarchically, coordinate a response in a way reminiscent of the deliberation and negotiation that take place in a council. We then highlight the importance that the network structure has, and discuss that the natural decomposition approach has unveiled the system-level elements shaping a common functional architecture governing bacterial regulatory networks. We discuss the incompleteness of gene regulatory networks and the need for network inference and benchmarking standardization. We point out the importance that using the network structural properties showed to improve network inference. We discuss the advances and controversies regarding the consistency between reconstructions of regulatory networks and expression data. We then discuss some perspectives on the necessity of studying regulatory networks, considering the interactions’ strength distribution, the challenges to studying these interactions’ strength, and the corresponding effects on network structure and dynamics. Finally, we explore the ability of evolutionary systems biology studies to provide insights into how evolution shapes functional architecture despite the high evolutionary plasticity of regulatory networks.

RSNET: inferring gene regulatory networks by a redundancy silencing and network enhancement technique

Background

Current gene regulatory network (GRN) inference methods are notorious for a great number of indirect interactions hidden in the predictions. Filtering out the indirect interactions from direct ones remains an important challenge in the reconstruction of GRNs. To address this issue, we developed a redundancy silencing and network enhancement technique (RSNET) for inferring GRNs.

Results

To assess the performance of RSNET method, we implemented the experiments on several gold-standard networks by using simulation study, DREAM challenge dataset and Escherichia coli network. The results show that RSNET method performed better than the compared methods in sensitivity and accuracy. As a case of study, we used RSNET to construct functional GRN for apple fruit ripening from gene expression data.

Conclusions

In the proposed method, the redundant interactions including weak and indirect connections are silenced by recursive optimization adaptively, and the highly dependent nodes are constrained in the model to keep the real interactions. This study provides a useful tool for inferring clean networks.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12859-022-04696-w.

High-performance single-cell gene regulatory network inference at scale: the Inferelator 3.0

MOTIVATION

Gene regulatory networks define regulatory relationships between transcription factors and target genes within a biological system, and reconstructing them is essential for understanding cellular growth and function. Methods for inferring and reconstructing networks from genomics data have evolved rapidly over the last decade in response to advances in sequencing technology and machine learning. The scale of data collection has increased dramatically; the largest genome-wide gene expression datasets have grown from thousands of measurements to millions of single cells, and new technologies are on the horizon to increase to tens of millions of cells and above.

RESULTS

In this work, we present the Inferelator 3.0, which has been significantly updated to integrate data from distinct cell types to learn context-specific regulatory networks and aggregate them into a shared regulatory network, while retaining the functionality of the previous versions. The Inferelator is able to integrate the largest single-cell datasets and learn cell-type-specific gene regulatory networks. Compared to other network inference methods, the Inferelator learns new and informative Saccharomyces cerevisiae networks from single-cell gene expression data, measured by recovery of a known gold standard. We demonstrate its scaling capabilities by learning networks for multiple distinct neuronal and glial cell types in the developing Mus musculus brain at E18 from a large (1.3 million) single-cell gene expression dataset with paired single-cell chromatin accessibility data.

AVAILABILITY AND IMPLEMENTATION

The inferelator software is available on GitHub (https://github.com/flatironinstitute/inferelator) under the MIT license and has been released as python packages with associated documentation (https://inferelator.readthedocs.io/).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A 1p/19q Codeletion-Associated Immune Signature for Predicting Lower Grade Glioma Prognosis

Lower grade gliomas (LGGs) with codeletion of chromosomal arms 1p and 19q (1p/19 codeletion) have a favorable outcome. However, its overall survival (OS) varies. Here, we established an immune signature associated with 1p/19q codeletion for accurate prediction of prognosis of LGGs. The Chinese Glioma Genome Atlas (CGGA) and The Cancer Genome Atlas (TCGA) databases with RNA sequencing and corresponding clinical data were dichotomized into training group and testing group. The immune-related differentially expressed genes (DEGs) associated with 1p/19q codeletion were screened using Cox proportional hazards regression analyses. A prognostic signature was established using dataset from CGGA and tested in TCGA database. Subsequently, we explored the correlation between the prognostic signature and immune response. Thirteen immune genes associated with 1p/19q codeletion were used to construct a prognostic signature. The 1-, 3-, 5-year survival rates of the low-risk group were approximately 97%, 89%, and 79%, while those of the high-risk group were 81%, 50% and 34%, respectively, in the training group. The nomogram which comprised age, WHO grade, primary or recurrent types, 1p/19q codeletion status and risk score provided accurate prediction for the survival rate of glioma. DEGs that were highly expressed in the high-risk group clustered with many immune-related pathways. Immune checkpoints including TIM3, PD1, PDL1, CTLA4, TIGIT, MIR155HG, and CD48 were correlated with the risk score. VAV3 and TNFRFSF11B were found to be candidate immune checkpoints associated with prognosis. The 1p/19q codeletion-associated immune signature provides accurate prediction of OS. VAV3 and TNFRFSF11B are novel immune checkpoints.

Deep neural network prediction of genome-wide transcriptome signatures - beyond the Black-box

Prediction algorithms for protein or gene structures, including transcription factor binding from sequence information, have been transformative in understanding gene regulation. Here we ask whether human transcriptomic profiles can be predicted solely from the expression of transcription factors (TFs). We find that the expression of 1600 TFs can explain >95% of the variance in 25,000 genes. Using the light-up technique to inspect the trained NN, we find an over-representation of known TF-gene regulations. Furthermore, the learned prediction network has a hierarchical organization. A smaller set of around 125 core TFs could explain close to 80% of the variance. Interestingly, reducing the number of TFs below 500 induces a rapid decline in prediction performance. Next, we evaluated the prediction model using transcriptional data from 22 human diseases. The TFs were sufficient to predict the dysregulation of the target genes (rho = 0.61, P < 10⁻²¹⁶). By inspecting the model, key causative TFs could be extracted for subsequent validation using disease-associated genetic variants. We demonstrate a methodology for constructing an interpretable neural network predictor, where analyses of the predictors identified key TFs that were inducing transcriptional changes during disease.

Network inference with Granger causality ensembles on single-cell transcriptomics

Cellular gene expression changes throughout a dynamic biological process, such as differentiation. Pseudotimes estimate cells' progress along a dynamic process based on their individual gene expression states. Ordering the expression data by pseudotime provides information about the underlying regulator-gene interactions. Because the pseudotime distribution is not uniform, many standard mathematical methods are inapplicable for analyzing the ordered gene expression states. Here we present single-cell inference of networks using Granger ensembles (SINGE), an algorithm for gene regulatory network inference from ordered single-cell gene expression data. SINGE uses kernel-based Granger causality regression to smooth irregular pseudotimes and missing expression values. It aggregates predictions from an ensemble of regression analyses to compile a ranked list of candidate interactions between transcriptional regulators and target genes. In two mouse embryonic stem cell differentiation datasets, SINGE outperforms other contemporary algorithms. However, a more detailed examination reveals caveats about poor performance for individual regulators and uninformative pseudotimes.

The Promise of Patient-Derived Colon Organoids to Model Ulcerative Colitis

Physiologic, molecular, and genetic findings all point to impaired intestinal epithelial function as a key element in the multifactorial pathogenesis of ulcerative colitis (UC). The lack of epithelial-directed therapies is a conspicuous weakness of our UC therapeutic armamentarium. However, a critical barrier to new drug discovery is the lack of preclinical human models of UC. Patient tissue-derived colon epithelial organoids (colonoids) are primary epithelial stem cell-derived in vitro structures capable of self-organization and self-renewal that hold great promise as a human preclinical model for UC drug development. Several single and multi-tissue systems for colonoid culture have been developed, including 3-dimensional colonoids grown in a gelatinous extracellular matrix, 2-dimensional polarized monolayers, and colonoids on a chip that model luminal and blood flow and nutrient delivery. A small number of pioneering studies suggest that colonoids derived from UC patients retain some disease-related transcriptional and epigenetic changes, but they also raise questions regarding the persistence of inflammatory transcriptional programs in culture over time. Additional research is needed to fully characterize the extent to which and under what conditions colonoids accurately model disease-associated epithelial molecular and functional aberrations. With further advancement and standardization of colonoid culture methodology, colonoids will likely become an important tool for realizing precision medicine in UC.

Genetic and epigenetic coordination of cortical interneuron development

One of the hallmarks of the cerebral cortex is the extreme diversity of interneurons^1-3. The two largest subtypes of cortical interneurons, parvalbumin- and somatostatin-positive cells, are morphologically and functionally distinct in adulthood but arise from common lineages within the medial ganglionic eminence^4-11. This makes them an attractive model for studying the generation of cell diversity. Here we examine how developmental changes in transcription and chromatin structure enable these cells to acquire distinct identities in the mouse cortex. Generic interneuron features are first detected upon cell cycle exit through the opening of chromatin at distal elements. By constructing cell-type-specific gene regulatory networks, we observed that parvalbumin- and somatostatin-positive cells initiate distinct programs upon settling within the cortex. We used these networks to model the differential transcriptional requirement of a shared regulator, Mef2c, and confirmed the accuracy of our predictions through experimental loss-of-function experiments. We therefore reveal how a common molecular program diverges to enable these neuronal subtypes to acquire highly specialized properties by adulthood. Our methods provide a framework for examining the emergence of cellular diversity, as well as for quantifying and predicting the effect of candidate genes on cell-type-specific development.

Bayesian information sharing enhances detection of regulatory associations in rare cell types

MOTIVATION

Recent advances in single-cell RNA-sequencing (scRNA-seq) technologies promise to enable the study of gene regulatory associations at unprecedented resolution in diverse cellular contexts. However, identifying unique regulatory associations observed only in specific cell types or conditions remains a key challenge; this is particularly so for rare transcriptional states whose sample sizes are too small for existing gene regulatory network inference methods to be effective.

RESULTS

We present ShareNet, a Bayesian framework for boosting the accuracy of cell type-specific gene regulatory networks by propagating information across related cell types via an information sharing structure that is adaptively optimized for a given single-cell dataset. The techniques we introduce can be used with a range of general network inference algorithms to enhance the output for each cell type. We demonstrate the enhanced accuracy of our approach on three benchmark scRNA-seq datasets. We find that our inferred cell type-specific networks also uncover key changes in gene associations that underpin the complex rewiring of regulatory networks across cell types, tissues and dynamic biological processes. Our work presents a path toward extracting deeper insights about cell type-specific gene regulation in the rapidly growing compendium of scRNA-seq datasets.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

AVAILABILITY AND IMPLEMENTATION

The code for ShareNet is available at http://sharenet.csail.mit.edu and https://github.com/alexw16/sharenet.

Single cell gene regulatory networks in plants: Opportunities for enhancing climate change stress resilience

Global warming poses major challenges for plant survival and agricultural productivity. Thus, efforts to enhance stress resilience in plants are key strategies for protecting food security. Gene regulatory networks (GRNs) are a critical mechanism conferring stress resilience. Until recently, predicting GRNs of the individual cells that make up plants and other multicellular organisms was impeded by aggregate population scale measurements of transcriptome and other genome-scale features. With the advancement of high-throughput single cell RNA-seq and other single cell assays, learning GRNs for individual cells is now possible, in principle. In this article, we report on recent advances in experimental and analytical methodologies for single cell sequencing assays especially as they have been applied to the study of plants. We highlight recent advances and ongoing challenges for scGRN prediction, and finally, we highlight the opportunity to use scGRN discovery for studying and ultimately enhancing abiotic stress resilience in plants.

Forecasting cellular states: from descriptive to predictive biology via single-cell multiomics

As the single cell field races to characterize each cell type, state, and behavior, the complexity of the computational analysis approaches the complexity of the biological systems. Single cell and imaging technologies now enable unprecedented measurements of state transitions in biological systems, providing high-throughput data that capture tens-of-thousands of measurements on hundreds-of-thousands of samples. Thus, the definition of cell type and state is evolving to encompass the broad range of biological questions now attainable. To answer these questions requires the development of computational tools for integrated multi-omics analysis. Merged with mathematical models, these algorithms will be able to forecast future states of biological systems, going from statistical inferences of phenotypes to time course predictions of the biological systems with dynamic maps analogous to weather systems. Thus, systems biology for forecasting biological system dynamics from multi-omic data represents the future of cell biology empowering a new generation of technology-driven predictive medicine.

Integrating multiple references for single-cell assignment

Efficient single-cell assignment is essential for single-cell sequencing data analysis. With the explosive growth of single-cell sequencing data, multiple single-cell sequencing data sources are available for the same kind of tissue, which can be integrated to further improve single-cell assignment; however, an efficient integration strategy is still lacking due to the great challenges of data heterogeneity existing in multiple references. To this end, we present mtSC, a flexible single-cell assignment framework that integrates multiple references based on multitask deep metric learning designed specifically for cell type identification within tissues with multiple single-cell sequencing data as references. We evaluated mtSC on a comprehensive set of publicly available benchmark datasets and demonstrated its state-of-the-art effectiveness for integrative single-cell assignment with multiple references.

Inferring gene regulatory network from single-cell transcriptomic data by integrating multiple prior networks

Gene regulatory network models the interactions between transcription factors and target genes. Reconstructing gene regulation network is critically important to understand gene function in a particular cellular context, providing key insights into complex biological systems. We develop a new computational method, named iMPRN, which integrates multiple prior networks to infer regulatory network. Based on the network component analysis model, iMPRN adopts linear regression, graph embedding, and elastic networks to optimize each prior network in line with specific biological context. For each rewired prior networks, iMPRN evaluate the confidence of the regulatory edges in each network based on B scores and finally integrated these optimized networks. We validate the effectiveness of iMPRN by comparing it with four widely-used gene regulatory network reconstruction algorithms on a simulation data set. The results show that iMPRN can infer the gene regulatory network more accurately. Further, on a real scRNA-seq dataset, iMPRN is respectively applied to reconstruct gene regulatory networks for malignant and nonmalignant head and neck tumor cells, demonstrating distinctive differences in their corresponding regulatory networks.

Network Analysis of Gene Transcriptions of Arabidopsis thaliana in Spaceflight Microgravity

The transcriptomic datasets of the plant model organism Arabidopsis thaliana grown in the International Space Station provided by GeneLab have been mined to isolate the impact of spaceflight microgravity on gene expressions related to root growth. A set of computational tools is used to identify the hub genes that respond differently in spaceflight with controlled lighting compared to on the ground. These computational tools based on graph-theoretic approaches are used to infer gene regulatory networks from the transcriptomic datasets. The three main algorithms used for network analyses are Least Absolute Shrinkage and Selection Operator (LASSO), Pearson correlation, and the Hyperlink-Induced Topic Search (HITS) algorithm. Graph-based spectral analyses reveal distinct properties of the spaceflight microgravity networks for the Wassilewskija (WS), Columbia (Col)-0, and mutant phytochromeD (phyD) ecotypes. The set of hub genes that are significantly altered in spaceflight microgravity are mainly involved in cell wall synthesis, protein transport, response to auxin, stress responses, and catabolic processes. Network analysis highlights five important root growth-regulating hub genes that have the highest outdegree distribution in spaceflight microgravity networks. These concerned genes coding for proteins are identified from the Gene Regulatory Networks (GRNs) corresponding to spaceflight total light environment. Furthermore, network analysis uncovers genes that encode nucleotide-diphospho-sugar interconversion enzymes that have higher transcriptional regulation in spaceflight microgravity and are involved in cell wall biosynthesis.

Biomarker-based evaluation of cytogenotoxic potential of glyphosate in Vigna mungo (L.) Hepper genotypes

Herbicides have proven to be a boon for agricultural fields. Their inherent property to kill weeds and unwanted vegetation makes them an essential biological tool for farmers and agricultural systems. Besides being capable of destroying weeds, they also exhibit certain effects on non-target crop plants. In the present study, a laboratory experiment was performed to assess the effect of glyphosate on Vigna mungo root meristem cells. Seeds of five different genotypes of V. mungo were treated with a series of concentrations of glyphosate ranging from 1 to 10 mM, and their effects on mitotic cell division were studied. Healthy and uniform-sized seeds were selected and were allowed to grow in Petri plates for 3 days, and all the doses were maintained in triplicates. Roots were fixed at day 3 after treatment (DAT) for cytological microscopic slide preparation. The results obtained indicate the dose-dependent reduction in the mitotic index in all the genotypes and an increase in the percentage of chromosomal aberrations (CAs) and relative abnormality rate (RAR). Most commonly observed chromosome aberrations at lower doses (< 6 mM) were fragments, stickiness, and disoriented metaphase, while at higher doses (6 to 10 mM) bridges, laggards, spindle disorientation, and clumping were obvious. The increase in the percentage of CAs and RAR indicates the inhibitory effect of glyphosate on cell cycle progression at various stages in root tip cells. The present study is a fine example of a biomarker-based genotoxic assessment of mitotic damage caused by glyphosate.

Tools for time-course simulation in systems biology: a brief overview

Dynamic modeling of biological systems is essential for understanding all properties of a given organism as it allows us to look not only at the static picture of an organism but also at its behavior under various conditions. With the increasing amount of experimental data, the number of tools that enable dynamic analysis also grows. However, various tools are based on different approaches, use different types of data and offer different functions for analyses; so it can be difficult to choose the most suitable tool for a selected type of model. Here, we bring a brief overview containing descriptions of 50 tools for the reconstruction of biological models, their time-course simulation and dynamic analysis. We examined each tool using test data and divided them based on the qualitative and quantitative nature of the mathematical apparatus they use.

SPONGEdb: a pan-cancer resource for competing endogenous RNA interactions

microRNAs (miRNAs) are post-transcriptional regulators involved in many biological processes and human diseases, including cancer. The majority of transcripts compete over a limited pool of miRNAs, giving rise to a complex network of competing endogenous RNA (ceRNA) interactions. Currently, gene-regulatory networks focus mostly on transcription factor-mediated regulation, and dedicated efforts for charting ceRNA regulatory networks are scarce. Recently, it became possible to infer ceRNA interactions genome-wide from matched gene and miRNA expression data. Here, we inferred ceRNA regulatory networks for 22 cancer types and a pan-cancer ceRNA network based on data from The Cancer Genome Atlas. To make these networks accessible to the biomedical community, we present SPONGEdb, a database offering a user-friendly web interface to browse and visualize ceRNA interactions and an application programming interface accessible by accompanying R and Python packages. SPONGEdb allows researchers to identify potent ceRNA regulators via network centrality measures and to assess their potential as cancer biomarkers through survival, cancer hallmark and gene set enrichment analysis. In summary, SPONGEdb is a feature-rich web resource supporting the community in studying ceRNA regulation within and across cancer types.

Multi-task learning for the simultaneous reconstruction of the human and mouse gene regulatory networks

The reconstruction of Gene Regulatory Networks (GRNs) from gene expression data, supported by machine learning approaches, has received increasing attention in recent years. The task at hand is to identify regulatory links between genes in a network. However, existing methods often suffer when the number of labeled examples is low or when no negative examples are available. In this paper we propose a multi-task method that is able to simultaneously reconstruct the human and the mouse GRNs using the similarities between the two. This is done by exploiting, in a transfer learning approach, possible dependencies that may exist among them. Simultaneously, we solve the issues arising from the limited availability of examples of links by relying on a novel clustering-based approach, able to estimate the degree of certainty of unlabeled examples of links, so that they can be exploited during the training together with the labeled examples. Our experiments show that the proposed method can reconstruct both the human and the mouse GRNs more effectively compared to reconstructing each network separately. Moreover, it significantly outperforms three state-of-the-art transfer learning approaches that, analogously to our method, can exploit the knowledge coming from both organisms. Finally, a specific robustness analysis reveals that, even when the number of labeled examples is very low with respect to the number of unlabeled examples, the proposed method is almost always able to outperform its single-task counterpart.

Data integration for inferring context-specific gene regulatory networks

Transcriptional regulatory networks control context-specific gene expression patterns and play important roles in normal and disease processes. Advances in genomics are rapidly increasing our ability to measure different components of the regulation machinery at the single-cell and bulk population level. An important challenge is to combine different types of regulatory genomic measurements to construct a more complete picture of gene regulatory networks across different disease, environmental, and developmental contexts. In this review, we focus on recent computational methods that integrate regulatory genomic data sets to infer context specificity and dynamics in regulatory networks.

Joint learning of multiple gene networks from single-cell gene expression data

Inferring gene networks from gene expression data is important for understanding functional organizations within cells. With the accumulation of single-cell RNA sequencing (scRNA-seq) data, it is possible to infer gene networks at single cell level. However, due to the characteristics of scRNA-seq data, such as cellular heterogeneity and high sparsity caused by dropout events, traditional network inference methods may not be suitable for scRNA-seq data. In this study, we introduce a novel joint Gaussian copula graphical model (JGCGM) to jointly estimate multiple gene networks for multiple cell subgroups from scRNA-seq data. Our model can deal with non-Gaussian data with missing values, and identify the common and unique network structures of multiple cell subgroups, which is suitable for scRNA-seq data. Extensive experiments on synthetic data demonstrate that our proposed model outperforms other compared state-of-the-art network inference models. We apply our model to real scRNA-seq data sets to infer gene networks of different cell subgroups. Hub genes in the estimated gene networks are found to be biological significance.

Perturbation-based gene regulatory network inference to unravel oncogenic mechanisms

The gene regulatory network (GRN) of human cells encodes mechanisms to ensure proper functioning. However, if this GRN is dysregulated, the cell may enter into a disease state such as cancer. Understanding the GRN as a system can therefore help identify novel mechanisms underlying disease, which can lead to new therapies. To deduce regulatory interactions relevant to cancer, we applied a recent computational inference framework to data from perturbation experiments in squamous carcinoma cell line A431. GRNs were inferred using several methods, and the false discovery rate was controlled by the NestBoot framework. We developed a novel approach to assess the predictiveness of inferred GRNs against validation data, despite the lack of a gold standard. The best GRN was significantly more predictive than the null model, both in cross-validated benchmarks and for an independent dataset of the same genes under a different perturbation design. The inferred GRN captures many known regulatory interactions central to cancer-relevant processes in addition to predicting many novel interactions, some of which were experimentally validated, thus providing mechanistic insights that are useful for future cancer research.

SCGRNs: Novel supervised inference of single-cell gene regulatory networks of complex diseases

Single-cell gene regulatory network (SCGRN) inference refers to the process of inferring gene regulatory networks from single-cell data, which are generated via single-cell RNA-sequencing (scRNA-seq) technologies. Although scRNA-seq leads to the generation of data pertaining to cells of particular interest, the single-cell data are noisy and highly sparse, which makes the analysis of such data a challenging task. In this study, we model an SCGRN as a directed graph where an edge from a source node (also called transcription factor (TF)) to a target node (also called target gene) indicates that a TF regulates a target gene. Inferring the SCGRN via predicting TF-target gene regulations would help biologists better understand various diseases in terms of networks. Following the modeling step, we propose three machine learning approaches. The first approach considers feature vectors encoding regulatory relationships of expressed TFs-target genes as input. The resulting model is then used to predict unseen TF-target gene regulations. The second machine learning approach constructs new feature vectors via incorporating features obtained from stacked autoencoders, which are provided to a machine learning algorithm to induce a model and predict unseen regulations of TFs-target genes. The third approach extends the second approach via including topological features extracted from an SCGRN. We perform an experimental study comparing our approaches against adapted unsupervised approaches. Experimental results on SCGRNs pertaining to healthy and type 2 pancreatic diabetes demonstrate the clinical importance and the accurate prediction performance of the proposed approaches.

Dual threshold optimization and network inference reveal convergent evidence from TF binding locations and TF perturbation responses

A high-confidence map of the direct, functional targets of each transcription factor (TF) requires convergent evidence from independent sources. Two significant sources of evidence are TF binding locations and the transcriptional responses to direct TF perturbations. Systematic data sets of both types exist for yeast and human, but they rarely converge on a common set of direct, functional targets for a TF. Even the few genes that are both bound and responsive may not be direct functional targets. Our analysis shows that when there are many nonfunctional binding sites and many indirect targets, nonfunctional sites are expected to occur in the cis-regulatory DNA of indirect targets by chance. To address this problem, we introduce dual threshold optimization (DTO), a new method for setting significance thresholds on binding and perturbation-response data, and show that it improves convergence. It also enables comparison of binding data to perturbation-response data that have been processed by network inference algorithms, which further improves convergence. The combination of dual threshold optimization and network inference greatly expands the high-confidence TF network map in both yeast and human. Next, we analyze a comprehensive new data set measuring the transcriptional response shortly after inducing overexpression of a yeast TF. We also present a new yeast binding location data set obtained by transposon calling cards and compare it to recent ChIP-exo data. These new data sets improve convergence and expand the high-confidence network synergistically.

Gene regulatory network reconstruction using single-cell RNA sequencing of barcoded genotypes in diverse environments

Understanding how gene expression programs are controlled requires identifying regulatory relationships between transcription factors and target genes. Gene regulatory networks are typically constructed from gene expression data acquired following genetic perturbation or environmental stimulus. Single-cell RNA sequencing (scRNAseq) captures the gene expression state of thousands of individual cells in a single experiment, offering advantages in combinatorial experimental design, large numbers of independent measurements, and accessing the interaction between the cell cycle and environmental responses that is hidden by population-level analysis of gene expression. To leverage these advantages, we developed a method for scRNAseq in budding yeast (Saccharomyces cerevisiae). We pooled diverse transcriptionally barcoded gene deletion mutants in 11 different environmental conditions and determined their expression state by sequencing 38,285 individual cells. We benchmarked a framework for learning gene regulatory networks from scRNAseq data that incorporates multitask learning and constructed a global gene regulatory network comprising 12,228 interactions.

Centralization Within Sub-Experiments Enhances the Biological Relevance of Gene Co-expression Networks: A Plant Mitochondrial Case Study

Gene co-expression networks (GCNs) can be prepared using a variety of mathematical approaches based on data sampled across diverse developmental processes, tissue types, pathologies, mutant backgrounds, and stress conditions. These networks are used to identify genes with similar expression dynamics but are prone to introducing false-positive and false-negative relationships, especially in the instance of large and heterogenous datasets. With the aim of optimizing the relevance of edges in GCNs and enhancing global biological insight, we propose a novel approach that involves a data-centering step performed simultaneously per gene and per sub-experiment, called centralization within sub-experiments (CSE). Using a gene set encoding the plant mitochondrial proteome as a case study, our results show that all CSE-based GCNs assessed had significantly more edges within the majority of the considered functional sub-networks, such as the mitochondrial electron transport chain and its complexes, than GCNs not using CSE; thus demonstrating that CSE-based GCNs are efficient at predicting canonical functions and associated pathways, here referred to as the core gene network. Furthermore, we show that correlation analyses using CSE-processed data can be used to fine-tune prediction of the function of uncharacterized genes; while its use in combination with analyses based on non-CSE data can augment conventional stress analyses with the innate connections underpinning the dynamic system being examined. Therefore, CSE is an effective alternative method to conventional batch correction approaches, particularly when dealing with large and heterogenous datasets. The method is easy to implement into a pre-existing GCN analysis pipeline and can provide enhanced biological relevance to conventional GCNs by allowing users to delineate a core gene network.

AUTHOR SUMMARY

Gene co-expression networks (GCNs) are the product of a variety of mathematical approaches that identify causal relationships in gene expression dynamics but are prone to the misdiagnoses of false-positives and false-negatives, especially in the instance of large and heterogenous datasets. In light of the burgeoning output of next-generation sequencing projects performed on a variety of species, and developmental or clinical conditions; the statistical power and complexity of these networks will undoubtedly increase, while their biological relevance will be fiercely challenged. Here, we propose a novel approach to generate a "core" GCN with enhanced biological relevance. Our method involves a data-centering step that effectively removes all primary treatment/tissue effects, which is simple to employ and can be easily implemented into pre-existing GCN analysis pipelines. The gain in biological relevance resulting from the adoption of this approach was assessed using a plant mitochondrial case study.

Network modeling of single-cell omics data: challenges, opportunities, and progresses

Single-cell multi-omics technologies are rapidly evolving, prompting both methodological advances and biological discoveries at an unprecedented speed. Gene regulatory network modeling has been used as a powerful approach to elucidate the complex molecular interactions underlying biological processes and systems, yet its application in single-cell omics data modeling has been met with unique challenges and opportunities. In this review, we discuss these challenges and opportunities, and offer an overview of the recent development of network modeling approaches designed to capture dynamic networks, within-cell networks, and cell-cell interaction or communication networks. Finally, we outline the remaining gaps in single-cell gene network modeling and the outlooks of the field moving forward.

Inferring Regulatory Programs Governing Region Specificity of Neuroepithelial Stem Cells during Early Hindbrain and Spinal Cord Development

Neuroepithelial stem cells (NSC) from different anatomical regions of the embryonic neural tube's rostrocaudal axis can differentiate into diverse central nervous system tissues, but the transcriptional regulatory networks governing these processes are incompletely understood. Here, we measure region-specific NSC gene expression along the rostrocaudal axis in a human pluripotent stem cell model of early central nervous system development over a 72-h time course, spanning the hindbrain to cervical spinal cord. We introduce Escarole, a probabilistic clustering algorithm for non-stationary time series, and combine it with prior-based regulatory network inference to identify genes that are regulated dynamically and predict their upstream regulators. We identify known regulators of patterning and neural development, including the HOX genes, and predict a direct regulatory connection between the transcription factor POU3F2 and target gene STMN2. We demonstrate that POU3F2 is required for expression of STMN2, suggesting that this regulatory connection is important for region specificity of NSCs.

Leveraging chromatin accessibility for transcriptional regulatory network inference in T Helper 17 Cells

Transcriptional regulatory networks (TRNs) provide insight into cellular behavior by describing interactions between transcription factors (TFs) and their gene targets. The assay for transposase-accessible chromatin (ATAC)–seq, coupled with TF motif analysis, provides indirect evidence of chromatin binding for hundreds of TFs genome-wide. Here, we propose methods for TRN inference in a mammalian setting, using ATAC-seq data to improve gene expression modeling. We test our methods in the context of T Helper Cell Type 17 (Th17) differentiation, generating new ATAC-seq data to complement existing Th17 genomic resources. In this resource-rich mammalian setting, our extensive benchmarking provides quantitative, genome-scale evaluation of TRN inference, combining ATAC-seq and RNA-seq data. We refine and extend our previous Th17 TRN, using our new TRN inference methods to integrate all Th17 data (gene expression, ATAC-seq, TF knockouts, and ChIP-seq). We highlight newly discovered roles for individual TFs and groups of TFs (“TF–TF modules”) in Th17 gene regulation. Given the popularity of ATAC-seq, which provides high-resolution with low sample input requirements, we anticipate that our methods will improve TRN inference in new mammalian systems, especially in vivo, for cells directly from humans and animal models.