Hierarchical Graph Transformer With Contrastive Learning for Gene Regulatory Network Inference.

Gene Regulatory Networks (GRNs) inference from gene expression data is a hard task and a widely addressed challenge. GRNs can be represented as Boolean models similarly to digital circuits. Cartesian Genetic Programming (CGP), often used for designing circuits, can thus be adopted in the inference of GRNs. The main CGP operator for generating candidate designs is mutation, making its choice important for obtaining good results. Although there are many mutation operators for CGP, to the best of our knowledge, there is no analysis of them in the GRN inference problem. An evaluation of the Single ActiveMutation (SAM) and the Semantically-Oriented Mutation Operator (SOMO)is performed here for GRNs inference. Also, a combination of both operators is proposed. We use a benchmark single-cell RNA-Sequencingtime series data and its evaluation pipeline to measure the performance of the approaches. The experiments indicate that (i) combining SOMOand SAM provides the best results, and (ii) the results obtained by the proposal are competitive with those from state-of-the-art methods.

Single-cell RNA-sequencing (scRNA-seq) has emerged as a powerful technique for studying gene expression patterns at the single-cell level. Inferring gene regulatory networks (GRNs) from scRNA-seq data provides insight into cellular phenotypes from the genomic level. However, the high sparsity, noise and dropout events inherent in scRNA-seq data present challenges for GRN inference. In recent years, the dramatic increase in data on experimentally validated transcription factors binding to DNA has made it possible to infer GRNs by supervised methods. In this study, we address the problem of GRN inference by framing it as a graph link prediction task. In this paper, we propose a novel framework called GNNLink, which leverages known GRNs to deduce the potential regulatory interdependencies between genes. First, we preprocess the raw scRNA-seq data. Then, we introduce a graph convolutional network-based interaction graph encoder to effectively refine gene features by capturing interdependencies between nodes in the network. Finally, the inference of GRN is obtained by performing matrix completion operation on node features. The features obtained from model training can be applied to downstream tasks such as measuring similarity and inferring causality between gene pairs. To evaluate the performance of GNNLink, we compare it with six existing GRN reconstruction methods using seven scRNA-seq datasets. These datasets encompass diverse ground truth networks, including functional interaction networks, Loss of Function/Gain of Function data, non-specific ChIP-seq data and cell-type-specific ChIP-seq data. Our experimental results demonstrate that GNNLink achieves comparable or superior performance across these datasets, showcasing its robustness and accuracy. Furthermore, we observe consistent performance across datasets of varying scales. For reproducibility, we provide the data and source code of GNNLink on our GitHub repository: https://github.com/sdesignates/GNNLink.

A survey of the methodological process of modeling, inference, and evaluation of gene regulatory networks using scRNA-Seq data.

To address one of the most challenging ecosystems issues at the cellular level, this paper surveys the fuzzy methods used in gene regulatory networks (GRNs) inference. GRNs represent causal relationships between genes that have a direct influence on the life and the development of living organisms, and provide a useful contribution to the understanding of the cellular functions as well as the mechanisms of diseases. The ecosystems impacted by GRN inference span various levels from cell to society -- globally.

Genes encode proteins, some of which in turn regulate other genes. Such interactions make up a gene regulatory network. The understanding and unraveling of gene regulatory networks have been proven very useful in disease diagnosis and genomic drug design. Due to the complexity of gene regulatory networks, the completely understanding of their dynamics is difficult to achieve only through biological experiments without any computational aids. As a consequence, computational models for gene regulatory networks are indispensable. Recently a wide variety of different computational models have been proposed for inferring gene regulatory networks. This paper surveys some of computational models for inferring large gene regulatory networks, in particular, Boolean network model, differential/ difference equation models, and state-space models. Some advantages and disadvantages of these models are commented on. Some criteria for validating the inferred gene regulatory networks are also discussed from the bioinformatics perspective. Finally, several directions of the future work for modeling gene regulatory networks are proposed.

Recent advances in measurement technologies, particularly single-cell RNA sequencing (scRNA-seq), have revolutionized our ability to acquire large amounts of omics-level data on cellular states. As measurement techniques evolve, there has been an increasing need for data analysis methodologies, especially those focused on cell-type identification and inference of gene regulatory networks (GRNs). We have developed a new method named BootCellNet, which employs smoothing and resampling to infer GRNs. Using the inferred GRNs, BootCellNet further infers the minimum dominating set (MDS), a set of genes that determines the dynamics of the entire network. We have demonstrated that BootCellNet robustly infers GRNs and their MDSs from scRNA-seq data and facilitates unsupervised identification of cell clusters using scRNA-seq datasets of peripheral blood mononuclear cells and hematopoiesis. It has also identified COVID-19 patient-specific cells and their potential regulatory transcription factors. BootCellNet not only identifies cell types in an unsupervised and explainable way but also provides insights into the characteristics of identified cell types through the inference of GRNs and MDS.

Inferring gene regulatory networks with hybrid of multi-agent genetic algorithm and random forests based on fuzzy cognitive maps

Inference of gene regulatory network from microarray data is one of the most important issues in bioinformatics. Several algorithms have been introduced for this problem, but they cannot give accurate results in case of time series gene expression data. Here, we propose an algorithm that predicts the gene regulatory network more accurately than the previous methods. A new method finds the relationship between a pair of genes by time-shifting the time series data of one gene against another when we compare the patterns of the gene pair. In addition, we increase the accuracy of the prediction method by filtering out the interactions that cannot exist in real biological network. We tested the algorithm to several simulated data which are based on realistic enzyme kinetics system, and evaluated the effectiveness of our algorithm. Results show that the present algorithm significantly improves the accuracy of the inference of gene regulatory network.

In this study, in order to deal with the noise and uncertainty in gene expression data, learning networks, especially Bayesian networks, that have the ability to use prior knowledge, were used to infer gene regulatory network. Learning networks are methods that have the structure of the network and a learning process to obtain relationships. One of the methods which have been used for measuring the relationship between genes is the correlation metrics, but the high correlated genes not necessarily mean that they have causal effect on each other. Studies on common methods in inference of gene regulatory networks are yet to pay attention to their biological importance and as such, predictions by these methods are less accurate in terms of biological significance. Hence, in the proposed method, genes with high correlation were identified in one cluster using clustering, and the existence of edge between the genes in the cluster was prevented. Finally, after the Bayesian network modeling, based on knowledge gained from clustering, the refining phase and improving regulatory interactions using biological correlation were done. In order to show the efficiency, the proposed method has been compared with several common methods in this area including GENIE3 and BMALR. The results of the evaluation indicate that the proposed method recognized regulatory relations in Bayesian modeling process well, due to using of biological knowledge which is hidden in the data collection, and is able to recognize gene regulatory networks align with important methods in this field.

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

As a tool of interpreting and analyzing genetic data, gene regulatory network (GRN) could reveal regulatory relationships between genes, proteins, and small molecules, as well as understand physiological activities and functions within biological cells, interact in pathways, and how to make changes in the organism. Traditional GRN research focuses on the analysis of the regulatory relationships through the average of cellular gene expressions. These methods are difficult to identify the cell heterogeneity of gene expression. Existing methods for inferring GRN using single-cell transcriptional data lack expression information when genes reach steady state, and the high dimensionality of single-cell data leads to high temporal and spatial complexity of the algorithm. In order to solve the problem in traditional GRN inference methods, including the lack of cellular heterogeneity information, single-cell data complexity and lack of steady-state information, we propose a method for GRN inference using single-cell transcription and gene knockout data, called SINgle-cell transcription data-KNOckout data (SIN-KNO), which focuses on combining dynamic and steady-state information of regulatory relationship contained in gene expression. Capturing cell heterogeneity information could help understand the gene expression difference in different cells. So, we could observe gene expression changes more accurately. Gene knockout data could observe the gene expression levels at steady-state of all other genes when one gene is knockout. Classifying the genes before analyzing the single-cell data could determine a large number of non-existent regulation, greatly reducing the number of regulation required for inference. In order to show the efficiency, the proposed method has been compared with several typical methods in this area including GENIE3, JUMP3, and SINCERITIES. The results of the evaluation indicate that the proposed method can analyze the diversified information contained in the two types of data, establish a more accurate gene regulation network, and improve the computational efficiency. The method provides a new thinking for dealing with large datasets and high computational complexity of single-cell data in the GRN inference.

Reconstruction of gene regulatory networks (GRNs), which explicitly represent the causality of developmental or regulatory process, is of utmost interest and has become a challenging computational problem for understanding the complex regulatory mechanisms in cellular systems. However, all existing methods of inferring GRNs from gene expression profiles have their strengths and weaknesses. In particular, many properties of GRNs, such as topology sparseness and non-linear dependence, are generally in regulation mechanism but seldom are taken into account simultaneously in one computational method. In this work, we present a novel method for inferring GRNs from gene expression data considering the non-linear dependence and topological structure of GRNs by employing path consistency algorithm (PCA) based on conditional mutual information (CMI). In this algorithm, the conditional dependence between a pair of genes is represented by the CMI between them. With the general hypothesis of Gaussian distribution underlying gene expression data, CMI between a pair of genes is computed by a concise formula involving the covariance matrices of the related gene expression profiles. The method is validated on the benchmark GRNs from the DREAM challenge and the widely used SOS DNA repair network in Escherichia coli. The cross-validation results confirmed the effectiveness of our method (PCA-CMI), which outperforms significantly other previous methods. Besides its high accuracy, our method is able to distinguish direct (or causal) interactions from indirect associations. All the source data and code are available at: http://csb.shu.edu.cn/subweb/grn.htm. lnchen@sibs.ac.cn; zpliu@sibs.ac.cn Supplementary data are available at Bioinformatics online.

Objective: Breast cancer is a common tumor and has a high mortality rate. Gene regulatory networks(GRNs) can genetically facilitate targeted therapies for this disease. Impact Statement: This study proposes a new method to infer GRNs. This new method combining genetic modules and convolutional neural networks is presented to infer GRNs from the RNA sequencing data of breast cancer. Introduction: GRNs play an essential role in many disease treatments. Previous studies showed that GRNs will accelerate tumor therapy. However, most of the existing network inference methods are based on large-scale gene collections, which ignore the characteristics of different tumors. Methods: In this work, weighted gene coexpression network analysis was deployed to screen key genes and gene modules. The gene regulatory associations in gene modules were then transformed into 2-dimensional histogram types. A convolutional neural network was chosen as the main framework to fit the gene regulatory types and infer the GRN. Results: The method integrates genetic data analysis and deep learning perspectives to screen key genes and predict GRNs among key genes. The key genes screened were validated by multiple methods, and the inferred gene regulatory associations were widely validated in real datasets. Conclusion: The method can be used as an auxiliary tool with the potential to predict key genes and the GRNs of key genes. It has the potential to facilitate the therapeutic process and targeted therapy for breast cancer.

Gene regulatory networks (GRN) have been studied by computational scientists and biologists over 20 years to gain a fine map of gene functions. With large-scale genomic and epigenetic data generated under diverse cells, tissues, and diseases, the integrative analysis of multi-omics data plays a key role in identifying casual genes in human disease development. Bayesian inference (or integration) has been successfully applied to inferring GRNs. Learning a posterior distribution than making a single-value prediction of model parameter makes Bayesian inference a more robust approach to identify GRN from noisy biomedical observations. Moreover, given multi-omics data as input and a large number of model parameters to estimate, the automatic preference of Bayesian inference for simple models that sufficiently explain data without unnecessary complexity ensures fast convergence to reliable results. In this chapter, we introduced GRN modeling using hierarchical Bayesian network and then used Gibbs sampling to identify network variables. We applied this model to breast cancer data and identified genes relevant to breast cancer recurrence. In the end, we discussed the potential of Bayesian inference as well as Bayesian deep learning for large-scale and complex GRN inference.

Biological tissues are composed of distinct microenvironments that spatially orchestrate gene expression and cell identity. However, the regulatory principles governing domain-specific cellular functions remain poorly understood due to the lack of effective methods for mapping gene regulatory networks (GRNs) in situ . To address this gap, we introduce STARNet, a representation learning approach that leverages heterogeneous hypergraph modeling of spatial transcriptomic and epigenomic data to resolve tissue-domain-specific regulatory interactions. By integrating graph neural networks with contrastive learning in a self-supervised framework, STARNet learns unified embeddings that preserve both multi-modal molecular features and anatomical spatial context, enabling accurate and domain-resolved GRN reconstruction within complex tissues. Benchmarking on both simulated and real datasets demonstrates that STARNet achieves state-of-the-art performance. We further demonstrate its broad applicability across diverse biological contexts, including neural development, genetic disease risk, and drug-induced developmental toxicity. In the mouse brain, it delineates region-specific regulatory networks and reconstructs spatiotemporal programs underlying neural stem cell differentiation. In human genetics, it provides a mechanistic link between genotypes and phenotypes by showing how genome-wide association study (GWAS) variants for complex diseases perturb hippocampus-specific GRNs. In developmental toxicology, STARNet reveals that drug-induced disruptions of GRNs in defined embryonic regions underlie tissue-specific vulnerability. Collectively, STARNet offers a powerful and versatile framework for resolving the spatial regulatory logic of complex tissues, providing multi-angle insights into tissue patterning, development, and disease mechanisms.