Transcription Factor Pairs Research Articles

Genome-Wide Analysis of CCT Transcript Factors to Identify Genes Contributing to Photoperiodic Flowering in Oryza rufipogon.

Photoperiod sensitivity is a dominant determinant for the phase transition in cereal crops. CCT (CONSTANS, CO-like, and TOC1) transcription factors (TFs) are involved in many physiological functions including the regulation of the photoperiodic flowering. However, the functional roles of CCT TFs have not been elucidated in the wild progenitors of crops. In this study, we identified 41 CCT TFs, including 19 CMF, 17 COL, and five PRR TFs in Oryza rufipogon, the presumed wild ancestor of Asian cultivated rice. There are thirty-eight orthologous CCT genes in Oryza sativa, of which ten pairs of duplicated CCT TFs are shared with O. rufipogon. We investigated daily expression patterns, showing that 36 OrCCT genes exhibited circadian rhythmic expression. A total of thirteen OrCCT genes were identified as putative flowering suppressors in O. rufipogon based on rhythmic and developmental expression patterns and transgenic phenotypes. We propose that OrCCT08, OrCCT24, and OrCCT26 are the strong functional alleles of rice DTH2, Ghd7, and OsPRR37, respectively. The SD treatment at 80 DAG stimulated flowering of the LD-grown O. rufipogon plants. Our results further showed that the nine OrCCT genes were significantly downregulated under the treatment. Our findings would provide valuable information for the construction of photoperiodic flowering regulatory network and functional characterization of the CCT TFs in both O. rufipogon and O. sativa.

DeepMotifSyn: a deep learning approach to synthesize heterodimeric DNA motifs.

The cooperativity of transcription factors (TFs) is a widespread phenomenon in the gene regulation system. However, the interaction patterns between TF binding motifs remain elusive. The recent high-throughput assays, CAP-SELEX, have identified over 600 composite DNA sites (i.e. heterodimeric motifs) bound by cooperative TF pairs. However, there are over 25 000 inferentially effective heterodimeric TFs in the human cells. It is not practically feasible to validate all heterodimeric motifs due to cost and labor. We introduce DeepMotifSyn, a deep learning-based tool for synthesizing heterodimeric motifs from monomeric motif pairs. Specifically, DeepMotifSyn is composed of heterodimeric motif generator and evaluator. The generator is a U-Net-based neural network that can synthesize heterodimeric motifs from aligned motif pairs. The evaluator is a machine learning-based model that can score the generated heterodimeric motif candidates based on the motif sequence features. Systematic evaluations on CAP-SELEX data illustrate that DeepMotifSyn significantly outperforms the current state-of-the-art predictors. In addition, DeepMotifSyn can synthesize multiple heterodimeric motifs with different orientation and spacing settings. Such a feature can address the shortcomings of previous models. We believe DeepMotifSyn is a more practical and reliable model than current predictors on heterodimeric motif synthesis. Contact:kc.w@cityu.edu.hk.

In Silico Prediction of Transcription Factor Collaborations Underlying Phenotypic Sexual Dimorphism in Zebrafish (Danio rerio).

The transcriptional regulation of gene expression in higher organisms is essential for different cellular and biological processes. These processes are controlled by transcription factors and their combinatorial interplay, which are crucial for complex genetic programs and transcriptional machinery. The regulation of sex-biased gene expression plays a major role in phenotypic sexual dimorphism in many species, causing dimorphic gene expression patterns between two different sexes. The role of transcription factor (TF) in gene regulatory mechanisms so far has not been studied for sex determination and sex-associated colour patterning in zebrafish with respect to phenotypic sexual dimorphism. To address this open biological issue, we applied bioinformatics approaches for identifying the predicted TF pairs based on their binding sites for sex and colour genes in zebrafish. In this study, we identified 25 (e.g., STAT6-GATA4; JUN-GATA4; SOX9-JUN) and 14 (e.g., IRF-STAT6; SOX9-JUN; STAT6-GATA4) potentially cooperating TFs based on their binding patterns in promoter regions for sex determination and colour pattern genes in zebrafish, respectively. The comparison between identified TFs for sex and colour genes revealed several predicted TF pairs (e.g., STAT6-GATA4; JUN-SOX9) are common for both phenotypes, which may play a pivotal role in phenotypic sexual dimorphism in zebrafish.

Genome-Wide Analysis of Coding and Non-coding RNA Reveals a Conserved miR164-NAC-mRNA Regulatory Pathway for Disease Defense in Populus.

MicroRNAs (miRNAs) contribute to plant defense responses by increasing the overall genetic diversity; however, their origins and functional importance in plant defense remain unclear. Here, we employed Illumina sequencing technology to assess how miRNA and messenger RNA (mRNA) populations vary in the Chinese white poplar (Populus tomentosa) during a leaf black spot fungus (Marssonina brunnea) infection. We sampled RNAs from infective leaves at conidia germinated stage [12 h post-inoculation (hpi)], infective vesicles stage (24 hpi), and intercellular infective hyphae stage (48 hpi), three essential stages associated with plant colonization and biotrophic growth in M. brunnea fungi. In total, 8,938 conserved miRNA-target gene pairs and 3,901 Populus-specific miRNA-target gene pairs were detected. The result showed that Populus-specific miRNAs (66%) were more involved in the regulation of the disease resistance genes. By contrast, conserved miRNAs (>80%) target more whole-genome duplication (WGD)-derived transcription factors (TFs). Among the 1,023 WGD-derived TF pairs, 44.9% TF pairs had only one paralog being targeted by a miRNA that could be due to either gain or loss of a miRNA binding site after the WGD. A conserved hierarchical regulatory network combining promoter analyses and hierarchical clustering approach uncovered a miR164–NAM, ATAF, and CUC (NAC) transcription factor–mRNA regulatory module that has potential in Marssonina defense responses. Furthermore, analyses of the locations of miRNA precursor sequences reveal that pseudogenes and transposon contributed a certain proportion (∼30%) of the miRNA origin. Together, these observations provide evolutionary insights into the origin and potential roles of miRNAs in plant defense and functional innovation.

Evidence of widespread, independent sequence signature for transcription factor cobinding.

Transcription factors (TFs) are the vocabulary that genomes use to regulate gene expression and phenotypes. The interactions among TFs enrich this vocabulary and orchestrate diverse biological processes. Although simple models identify open chromatin and the presence of TF motifs as the two major contributors to TF binding patterns, it remains elusive what contributes to the in vivo TF cobinding landscape. In this study, we developed a machine learning algorithm to explore the contributors of the cobinding patterns. The algorithm substantially outperforms the state-of-the-field models for TF cobinding prediction. Game theory–based feature importance analysis reveals that, for most of the TF pairs we studied, independent motif sequences contribute one or more of the two TFs under investigation to their cobinding patterns. Such independent motif sequences include, but are not limited to, transcription initiation–related proteins and known TF complexes. We found the motif sequence signatures and the TFs are rarely mutual, corroborating a hierarchical and directional organization of the regulatory network and refuting the possibility of artifacts caused by shared sequence similarity with the TFs under investigation. We modeled such regulatory language with directed graphs, which reveal shared, global factors that are related to many binding and cobinding patterns.

Master Regulators and Cofactors of Human Neuronal Cell Fate Specification Identified by CRISPR Gene Activation Screens

SUMMARYTechnologies to reprogram cell-type specification have revolutionized the fields of regenerative medicine and disease modeling. Currently, the selection of fate-determining factors for cell reprogramming applications is typically a laborious and low-throughput process. Therefore, we use high-throughput pooled CRISPR activation (CRISPRa) screens to systematically map human neuronal cell fate regulators. We utilize deactivated Cas9 (dCas9)-based gene activation to target 1,496 putative transcription factors (TFs) in the human genome. Using a reporter of neuronal commitment, we profile the neurogenic activity of these factors in human pluripotent stem cells (PSCs), leading to a curated set of pro-neuronal factors. Activation of pairs of TFs reveals neuronal cofactors, including E2F7, RUNX3, and LHX8, that improve conversion efficiency, subtype specificity, and maturation of neuronal cell types. Finally, using multiplexed gene regulation with orthogonal CRISPR systems, we demonstrate improved neuronal differentiation with concurrent activation and repression of target genes, underscoring the power of CRISPR-based gene regulation for programming complex cellular phenotypes.

A computational framework for identifying the transcription factors involved in enhancer-promoter loop formation

The pairwise interaction between transcription factors (TFs) plays an important role in enhancer-promoter loop formation. Although thousands of TFs in the human genome have been found, only a few TF pairs have been demonstrated to be related to loop formation. It is still a challenge to determine which TF pairs could be involved in the enhancer-promoter regulation network. This work describes a computational framework to identify TF pairs in enhancer-promoter regulation. By integrating different levels of data derived from Promoter Capture Hi-C, chromatin immunoprecipitation sequencing (ChIP-seq) of histone marks, RNA-seq, protein-protein interaction (PPI), and TF motif, we identified 361 significant TF pairs and constructed a TF interaction network. From the network, we found several hub-TFs, which may have important roles in the regulation of long-range interactions. Our studies extended TF pairs identified in other experimental and computational approaches. These findings will help the further study of long-range interactions between enhancers and promoters.

Comprehensive mapping of the human cytokine gene regulatory network.

Proper cytokine gene expression is essential in development, homeostasis and immune responses. Studies on the transcriptional control of cytokine genes have mostly focused on highly researched transcription factors (TFs) and cytokines, resulting in an incomplete portrait of cytokine gene regulation. Here, we used enhanced yeast one-hybrid (eY1H) assays to derive a comprehensive network comprising 1380 interactions between 265 TFs and 108 cytokine gene promoters. Our eY1H-derived network greatly expands the known repertoire of TF–cytokine gene interactions and the set of TFs known to regulate cytokine genes. We found an enrichment of nuclear receptors and confirmed their role in cytokine regulation in primary macrophages. Additionally, we used the eY1H-derived network as a framework to identify pairs of TFs that can be targeted with commercially-available drugs to synergistically modulate cytokine production. Finally, we integrated the eY1H data with single cell RNA-seq and phenotypic datasets to identify novel TF–cytokine regulatory axes in immune diseases and immune cell lineage development. Overall, the eY1H data provides a rich resource to study cytokine regulation in a variety of physiological and disease contexts.

Development and validation of a novel stem cell subtype for bladder cancer based on stem genomic profiling

BackgroundBladder cancer (BLCA) is the fifth most common type of cancer worldwide, with high recurrence and progression rates. Although considerable progress has been made in the treatment of BLCA through accurate typing of molecular characteristics, little is known regarding the various genetic and epigenetic changes that have evolved in stem and progenitor cells. To address this issue, we have developed a novel stem cell typing method.MethodsBased on six published genomic datasets, we used 26 stem cell gene sets to classify each dataset. Unsupervised and supervised machine learning methods were used to perform the classification.ResultsWe classified BLCA into three subtypes—high stem cell enrichment (SCE_H), medium stem cell enrichment (SCE_M), and low stem cell enrichment (SCE_L)—based on multiple cross-platform datasets. The stability and reliability of the classification were verified. Compared with the other subtypes, SCE_H had the highest degree of cancer stem cell concentration, highest level of immune cell infiltration, and highest sensitivity not only to predicted anti-PD-1 immunosuppressive therapy but also to conventional chemotherapeutic agents such as cisplatin, sunitinib, and vinblastine; however, this group had the worst prognosis. Comparison of gene set enrichment analysis results for pathway enrichment of various subtypes reveals that the SCE_H subtype activates the important pathways regulating cancer occurrence, development, and even poor prognosis, including epithelial-mesenchymal transition, hypoxia, angiogenesis, KRAS signal upregulation, interleukin 6-mediated JAK-STAT signaling pathway, and inflammatory response. Two identified pairs of transcription factors, GRHL2 and GATA6 and IRF5 and GATA3, possibly have opposite regulatory effects on SCE_H and SCE_L, respectively.ConclusionsThe identification of BLCA subtypes based on cancer stem cell gene sets revealed the complex mechanism of carcinogenesis of BLCA and provides a new direction for the diagnosis and treatment of BLCA.

Characterizing therapeutic signatures of transcription factors in cancer by incorporating profiles in compound treated cells.

Cancers are promoted by abnormal alterations in biological processes, such as cell cycle and apoptosis. An immediate reason for those aberrant processes is the deregulation of their involved transcription factors (TFs). Thus, the deregulated TFs in cancer have been experimented as successful therapeutic targets, such as RARA and RUNX1. This therapeutic strategy can be accelerated by characterizing new potential TF targets. Two kinds of therapeutic signatures of TFs in A375 (skin) and HT29 (colon) cancer cells were characterized by analyzing TF activities under effective and ineffective compounds to cancer. First, the therapeutic TFs (TTs) were identified as the TFs that are significantly activated or repressed under effective compared to ineffective compounds. Second, the therapeutically correlated TF pairs (TCPs) were determined as the TF pairs whose activity correlations show substantial discrepancy between the effective and ineffective compounds. It was facilitated by incorporating (i)compound-induced gene expressions (LINCS), (ii) compound-induced cell viabilities (GDSC) and (iii) TF-target interactions (TRUST2). As a result, among 627 TFs, the 35 TTs (such as MYCN and TP53) and the 214 TCPs (such as FOXO3 and POU2F2 pair) were identified. The TTs and the proteins on the paths between TCPs were compared with the known therapeutic targets, tumor suppressors, oncogenes and CRISPR-Cas9 knockout screening, which yielded significant consequences. We expect that the results provide good candidates for therapeutic TF targets in cancer. The data and Python implementations are available at https://github.com/jmjung83/TT_and_TCP. Supplementary data are available at Bioinformatics online.

Integrated Analysis of Small RNA, Transcriptome, and Degradome Sequencing Reveals the Water-Deficit and Heat Stress Response Network in Durum Wheat.

Water-deficit and heat stress negatively impact crop production. Mechanisms underlying the response of durum wheat to such stresses are not well understood. With the new durum wheat genome assembly, we conducted the first multi-omics analysis with next-generation sequencing, providing a comprehensive description of the durum wheat small RNAome (sRNAome), mRNA transcriptome, and degradome. Single and combined water-deficit and heat stress were applied to stress-tolerant and -sensitive Australian genotypes to study their response at multiple time-points during reproduction. Analysis of 120 sRNA libraries identified 523 microRNAs (miRNAs), of which 55 were novel. Differentially expressed miRNAs (DEMs) were identified that had significantly altered expression subject to stress type, genotype, and time-point. Transcriptome sequencing identified 49,436 genes, with differentially expressed genes (DEGs) linked to processes associated with hormone homeostasis, photosynthesis, and signaling. With the first durum wheat degradome report, over 100,000 transcript target sites were characterized, and new miRNA-mRNA regulatory pairs were discovered. Integrated omics analysis identified key miRNA-mRNA modules (particularly, novel pairs of miRNAs and transcription factors) with antagonistic regulatory patterns subject to different stresses. GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis revealed significant roles in plant growth and stress adaptation. Our research provides novel and fundamental knowledge, at the whole-genome level, for transcriptional and post-transcriptional stress regulation in durum wheat.

Abstract 29: Integrating computational epigenetic and statistical approaches to investigate how genome-wide transcription factor (TF)-DNA bindings affect breast cancer risk

Abstract Background: Fine-mapping and functional genomic studies of risk loci identified by genome wide association studies (GWAS) of breast cancer suggest that risk-associated regulatory variants may disrupt DNA binding affinities of transcription factors (TFs), such as FOXA1 and ESR1, thus altering gene expression and affecting breast cancer risk. To date, however, no studies have directly investigated how genome-wide TF-DNA bindings in a tissue-specific context affect breast cancer risk. Methods: We recently developed an analytic framework using computational epigenetic approaches and generalized mixed models to identify breast cancer risk associated cis-regulatory elements that are occupied by TFs. Summary statistics derived from GWAS of imputed data for approximately 11 million genetic variants were obtained from the Breast Cancer Association Consortium (BCAC). A total of 113 ChIP-seq datasets for TFs and chromatin features were annotated from the Encyclopedia of DNA Elements (ENCODE) and Roadmap projects and the Cistrome database (http://cistrome.org/) in multiple breast cancer cell lines..Chi-square values reported in the BCAC dataset were used to measure breast cancer risk associated with a genetic variant. To investigate TF-DNA bindings for breast cancer risk, we constructed generalized mixed models to evaluate the associations between the Chi-square values and binding sites of single TF or co-occupied by multiple TFs, given linkage disequilibrium (LD) blocks of variants to handle their dependence. To define approximately independent LD blocks similar to other studies, we defined LD blocks using non-overlapping segments of 100k bps. In addition, we investigated the effect of putative cis-regulatory elements characterized by TF-DNA bindings and chromatin features on breast cancer risk. Results: We identified a total of 22 TFs where their TF-DNA bindings were significantly associated with breast cancer risk at P < 1 × 10−5, with the top five TFs being FOXA1, SIN3AK, ESR1, TCF7l2, and AR (P < 1 × 10−10). We further observed stronger associations of breast cancer risk with co-occupied TF-TF DNA bindings, with the top five TF pairs (P < 1 × 10−16) being FOXA1+E2F1, FOXA1+NR2F2, FOXA1+ESR1, FOXA1+SIN3, and FOXA1+TCF12. The interactions of TF-TF DNA bindings for these five TF pairs were highly significant (P < 1 × 10−5). In addition, our results showed a significant interaction (P < 1 × 10−5) between chromatin features and the TF-DNA bindings. As compared with quiescent/low chromatin features, more TF-DNA bindings in enhancers, strong or weak transcription, or heterochromatin were associated with lower breast cancer risk. Conclusions: Our findings suggest that putative regulatory variants, which alter the TF-DNA binding affinities, particularly those located in TF-TF co-occupied sites, and TF-DNA binding colocalized with chromatin features, play an important role in breast cancer risk. Our approaches can be generalized to investigate the effect of genome-wide TF-DNA bindings on the risk of any human cancers that have comprehensive ChIP-seq and large-scale GWAS data. Citation Format: Wanqing Wen, Zhishan Chen, Quan Long, Wei Zheng, Xingyi Guo. Integrating computational epigenetic and statistical approaches to investigate how genome-wide transcription factor (TF)-DNA bindings affect breast cancer risk [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 29.

The differential expression patterns and co-expression networks of paralogs as an indicator of the TNM stages of lung adenocarcinoma and squamous cell carcinoma

Cancers constitute a severe threat to human health. Elucidating the association between the expression patterns of the paralogous genes and transcription factors (TF) and the progression of cancers by comprehensively investigating the expression patterns and co-expression networks will contribute to the in-depth understanding of the pathogenesis of cancers. Here, we identified the paralogous gene pairs and systematically analyzed the expression patterns of these paralogs and the known TFs to elucidate the associations with Tumor, Node, Metastasis (TNM) staging information across ten cancers. We found that the expression of ~60% paralogs was cancer-dependent, and more than 50% of the differentially expressed TFs pairs showed positive expression correlations. The down-regulation patterns of paralogs and TFs were closely associated with the M and N developmental stages of lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). Our results will help to understand the roles of paralogs and TFs in cancer progression and to screen prognostic biomarkers for early cancer diagnosis.

Dissection of c-AMP Response Element Architecture by Using Genomic and Episomal Massively Parallel Reporter Assays.

In eukaryotes, transcription factors (TFs) orchestrate gene expression by binding to TF-binding sites (TFBSs) and localizing transcriptional co-regulators and RNA polymerase II to cis-regulatory elements. However, we lack a basic understanding of the relationship between TFBS composition and their quantitative transcriptional responses. Here, we measured expression driven by 17,406 synthetic cis-regulatory elements with varied compositions of a model TFBS, the c-AMP response element (CRE) by using massively parallel reporter assays (MPRAs). We find CRE number, affinity, and promoter proximity largely determines expression. In addition, we observe expression modulation based on the spacing between CREs and CRE distance to the promoter, where expression follows a helical periodicity. Finally, we compare library expression between an episomal MPRA and a genomically integrated MPRA, where a single cis-regulatory element is assayed per cell at a defined locus. These assays largely recapitulate each other, although weaker, non-canonical CREs exhibit greater activity in a genomic context.

MicroRNA and transcription factor co-regulatory networks and subtype classification of seminoma and non-seminoma in testicular germ cell tumors

Recent studies have revealed that feed-forward loops (FFLs) as regulatory motifs have synergistic roles in cellular systems and their disruption may cause diseases including cancer. FFLs may include two regulators such as transcription factors (TFs) and microRNAs (miRNAs). In this study, we extensively investigated TF and miRNA regulation pairs, their FFLs, and TF-miRNA mediated regulatory networks in two major types of testicular germ cell tumors (TGCT): seminoma (SE) and non-seminoma (NSE). Specifically, we identified differentially expressed mRNA genes and miRNAs in 103 tumors using the transcriptomic data from The Cancer Genome Atlas. Next, we determined significantly correlated TF-gene/miRNA and miRNA-gene/TF pairs with regulation direction. Subsequently, we determined 288 and 664 dysregulated TF-miRNA-gene FFLs in SE and NSE, respectively. By constructing dysregulated FFL networks, we found that many hub nodes (12 out of 30 for SE and 8 out of 32 for NSE) in the top ranked FFLs could predict subtype-classification (Random Forest classifier, average accuracy ≥90%). These hub molecules were validated by an independent dataset. Our network analysis pinpointed several SE-specific dysregulated miRNAs (miR-200c-3p, miR-25-3p, and miR-302a-3p) and genes (EPHA2, JUN, KLF4, PLXDC2, RND3, SPI1, and TIMP3) and NSE-specific dysregulated miRNAs (miR-367-3p, miR-519d-3p, and miR-96-5p) and genes (NR2F1 and NR2F2). This study is the first systematic investigation of TF and miRNA regulation and their co-regulation in two major TGCT subtypes.

Core transcriptional regulatory circuits in prion diseases

Complex diseases involve dynamic perturbations of pathophysiological processes during disease progression. Transcriptional programs underlying such perturbations are unknown in many diseases. Here, we present core transcriptional regulatory circuits underlying early and late perturbations in prion disease. We first identified cellular processes perturbed early and late using time-course gene expression data from three prion-infected mouse strains. We then built a transcriptional regulatory network (TRN) describing regulation of early and late processes. We found over-represented feed-forward loops (FFLs) comprising transcription factor (TF) pairs and target genes in the TRN. Using gene expression data of brain cell types, we further selected active FFLs where TF pairs and target genes were expressed in the same cell type and showed correlated temporal expression changes in the brain. We finally determined core transcriptional regulatory circuits by combining these active FFLs. These circuits provide insights into transcriptional programs for early and late pathophysiological processes in prion disease.

Identification and Characterization of a Novel Basic Helix-Loop-Helix Transcription Factor of Phospholipid Synthesis Regulation in Aspergillus niger.

The synthesis of phospholipids relies on a sort of genes, whose promoter regions contain inositol-sensitive upstream activation sequence (UASINO) and are regulated by the basic helix-loop-helix (bHLH)-type ino2/ino4 transcription factor (TF) pair. Ten putative bHLH TFs have been found through whole genome sequencing of Aspergillus niger, but none of these TFs have been characterized. In this study, we identified and characterized the bHLH-type TF ino2(An02g04350) in A. niger. Electrophoretic mobility shift assay (EMSA) and yeast two-hybrid assay demonstrated that ino2 functions as a homodimer in UASINO genes (e.g., ino1 and cho1) and binds to opi1(An1g02370) in vitro. Real-time quantitative PCR of ino1 and quantification of total phospholipid indicated that the ino2 disruptant downregulated the transcription of ino1 and the amount of total cellular phosphatidylinositol. In addition, phenotype analyses showed that a loss of ino2 led to resistance to cell wall interference and DNA damage. Comparative transcriptome analyses showed that more than 1000 genes and GO terms associated with UASINO, endoplasmic reticulum–associated protein degradation, phosphatidylinositol synthesis, chitin synthesis, and fatty acid synthesis were differentially expressed in Δino2 compared to the wild type (WT). Taken together, these observations indicate that the bHLH TF ino2 functions as a homodimer that regulates the synthesis of phosphatidylinositol, fatty acid, and chitin and influences the homeostasis of the endoplasmic reticulum membrane.

Inductive inference of gene regulatory network using supervised and semi-supervised graph neural networks

Discovering gene regulatory relationships and reconstructing gene regulatory networks (GRN) based on gene expression data is a classical, long-standing computational challenge in bioinformatics. Computationally inferring a possible regulatory relationship between two genes can be formulated as a link prediction problem between two nodes in a graph. Graph neural network (GNN) provides an opportunity to construct GRN by integrating topological neighbor propagation through the whole gene network. We propose an end-to-end gene regulatory graph neural network (GRGNN) approach to reconstruct GRNs from scratch utilizing the gene expression data, in both a supervised and a semi-supervised framework. To get better inductive generalization capability, GRN inference is formulated as a graph classification problem, to distinguish whether a subgraph centered at two nodes contains the link between the two nodes. A linked pair between a transcription factor (TF) and a target gene, and their neighbors are labeled as a positive subgraph, while an unlinked TF and target gene pair and their neighbors are labeled as a negative subgraph. A GNN model is constructed with node features from both explicit gene expression and graph embedding. We demonstrate a noisy starting graph structure built from partial information, such as Pearson’s correlation coefficient and mutual information can help guide the GRN inference through an appropriate ensemble technique. Furthermore, a semi-supervised scheme is implemented to increase the quality of the classifier. When compared with established methods, GRGNN achieved state-of-the-art performance on the DREAM5 GRN inference benchmarks. GRGNN is publicly available at https://github.com/juexinwang/GRGNN.

A synthetic transcription factor pair mimic for precise recruitment of an epigenetic modifier to the targeted DNA locus.

We developed an epigenetically active, cooperative DNA binding transcription factor platform assisted by cucurbit[7]uril (CB7) host-guest modules. This new type of molecule termed ePIP-HoGu not only mimics the operation of transcription factors as a pair but also recruits the epigenetic modifier to a particular DNA locus.

Systematic identification of metabolites controlling gene expression in E. coli

Metabolism controls gene expression through allosteric interactions between metabolites and transcription factors. These interactions are usually measured with in vitro assays, but there are no methods to identify them at a genome-scale in vivo. Here we show that dynamic transcriptome and metabolome data identify metabolites that control transcription factors in E. coli. By switching an E. coli culture between starvation and growth, we induce strong metabolite concentration changes and gene expression changes. Using Network Component Analysis we calculate the activities of 209 transcriptional regulators and correlate them with metabolites. This approach captures, for instance, the in vivo kinetics of CRP regulation by cyclic-AMP. By testing correlations between all pairs of transcription factors and metabolites, we predict putative effectors of 71 transcription factors, and validate five interactions in vitro. These results show that combining transcriptomics and metabolomics generates hypotheses about metabolism-transcription interactions that drive transitions between physiological states.