A systems biology approach to transcription factor binding site prediction

PKMYT1 Is a Marker of Treatment Response and a Therapeutic Target for CDK4/6 Inhibitor-Resistance in ER+ Breast Cancer

Endocrine therapies (ET) with cyclin-dependent kinase 4/6 (CDK4/6) inhibition are the standard treatment for estrogen receptor-α-positive (ER+) breast cancer, however drug resistance is common. In this study, proteogenomic analyses of patient-derived xenografts (PDXs) from patients with 22 ER+ breast cancer demonstrated that protein kinase, membrane-associated tyrosine/threonine one (PKMYT1), a WEE1 homolog, is estradiol (E2) regulated in E2-dependent PDXs and constitutively expressed when growth is E2-independent. In clinical samples, high PKMYT1 mRNA levels associated with resistance to both ET and CDK4/6 inhibition. The PKMYT1 inhibitor lunresertib (RP-6306) with gemcitabine selectively and synergistically reduced the viability of ET and palbociclib-resistant ER+ breast cancer cells without functional p53. In vitro the combination increased DNA damage and apoptosis. In palbociclib-resistant, TP53 mutant PDX-derived organoids and PDXs, RP-6306 with low-dose gemcitabine induced greater tumor volume reduction compared to treatment with either single agent. Our study demonstrates the clinical potential of RP-6306 in combination with gemcitabine for ET and CDK4/6 inhibitor resistant TP53 mutant ER+ breast cancer.

ASPTF: A computational tool to predict abiotic stress-responsive transcription factors in plants by employing machine learning algorithms

BACKGROUND

Abiotic stresses pose serious threat to the growth and yield of crop plants. Several studies suggest that in plants, transcription factors (TFs) are important regulators of gene expression, especially when it comes to coping with abiotic stresses. Therefore, it is crucial to identify TFs associated with abiotic stress response for breeding of abiotic stress tolerant crop cultivars.

METHODS

Based on a machine learning framework, a computational model was envisaged to predict TFs associated with abiotic stress response in plants. To numerically encode TF sequences, four distinct sequence derived features were generated. The prediction was performed using ten shallow learning and four deep learning algorithms. For prediction using more pertinent and informative features, feature selection techniques were also employed.

RESULTS

Using the features chosen by the light-gradient boosting machine-variable importance measure (LGBM-VIM), the LGBM achieved the highest cross-validation performance metrics (accuracy: 86.81%, auROC: 92.98%, and auPRC: 94.03%). Further evaluation of the proposed model (LGBM prediction method + LGBM-VIM selected features) was also done using an independent test dataset, where the accuracy, auROC and auPRC were observed 81.98%, 90.65% and 91.30%, respectively.

CONCLUSIONS

To facilitate the adoption of the proposed strategy by users, the approach was implemented as a prediction server called ASPTF, accessible at https://iasri-sg.icar.gov.in/asptf/. The developed approach and the corresponding web application are anticipated to supplement experimental methods in the identification of transcription factors (TFs) responsive to abiotic stress in plants.

Pan-Cancer Analysis of lncRNA Regulation Supports Their Targeting of Cancer Genes in Each Tumor Context

Long noncoding RNAs (lncRNAs) are commonly dysregulated in tumors, but only a handful are known to play pathophysiological roles in cancer. We inferred lncRNAs that dysregulate cancer pathways, oncogenes, and tumor suppressors (cancer genes) by modeling their effects on the activity of transcription factors, RNA-binding proteins, and microRNAs in 5,185 TCGA tumors and 1,019 ENCODE assays. Our predictions included hundreds of candidate onco- and tumor-suppressor lncRNAs (cancer lncRNAs) whose somatic alterations account for the dysregulation of dozens of cancer genes and pathways in each of 14 tumor contexts. To demonstrate proof of concept, we showed that perturbations targeting OIP5-AS1 (an inferred tumor suppressor) and TUG1 and WT1-AS (inferred onco-lncRNAs) dysregulated cancer genes and altered proliferation of breast and gynecologic cancer cells. Our analysis indicates that, although most lncRNAs are dysregulated in a tumor-specific manner, some, including OIP5-AS1, TUG1, NEAT1, MEG3, and TSIX, synergistically dysregulate cancer pathways in multiple tumor contexts.

•

Hundreds of lncRNAs target cancer genes and pathways in each tumor context

•

lncRNA copy numbers are predictive of target cancer gene dysregulation

•

Most lncRNAs are predicted to be transcriptional or post-transcriptional specialists

•

lncRNAs are predicted to synergistically regulate proliferation pathways in cancer

Genomic analysis of hepatoblastoma identifies distinct molecular and prognostic subgroups

Despite being the most common liver cancer in children, hepatoblastoma (HB) is a rare neoplasm. Consequently, few pretreatment tumors have been molecularly profiled, and there are no validated prognostic or therapeutic biomarkers for HB patients. We report on the first large-scale effort to profile pretreatment HBs at diagnosis. Our analysis of 88 clinically annotated HBs revealed three risk-stratifying molecular subtypes that are characterized by differential activation of hepatic progenitor cell markers and metabolic pathways: high-risk tumors were characterized by up-regulated nuclear factor, erythroid 2-like 2 activity; high lin-28 homolog B, high mobility group AT-hook 2, spalt-like transcription factor 4, and alpha-fetoprotein expression; and high coordinated expression of oncofetal proteins and stem-cell markers, while low-risk tumors had low lin-28 homolog B and lethal-7 expression and high hepatic nuclear factor 1 alpha activity.

CONCLUSION

Analysis of immunohistochemical assays using antibodies targeting these genes in a prospective study of 35 HBs suggested that these candidate biomarkers have the potential to improve risk stratification and guide treatment decisions for HB patients at diagnosis; our results pave the way for clinical collaborative studies to validate candidate biomarkers and test their potential to improve outcome for HB patients. (Hepatology 2017;65:104-121).

Computational identification of potential transcriptional regulators of TGF-ß1 in human atherosclerotic arteries

TGF-ß is protective in atherosclerosis but deleterious in metastatic cancers. Our aim was to determine whether TGF-ß transcriptional regulation is tissue-specific in early atherosclerosis. The computational methods included 5 steps: (i) from microarray data of human atherosclerotic carotid tissue, to identify the 10 best co-expressed genes with TGFB1 (TGFB1 gene cluster), (ii) to choose the 11 proximal promoters, (iii) to predict the TFBS shared by the promoters, (iv) to identify the common TFs co-expressed with the TGFB1 gene cluster, and (v) to compare the common TFs in the early lesions to those identified in advanced atherosclerotic lesions and in various cancers. Our results show that EGR1, SP1 and KLF6 could be responsible for TGFB1 basal expression, KLF6 appearing specific to atherosclerotic lesions. Among the TFs co-expressed with the gene cluster, transcriptional activators (SLC2A4RG, MAZ) and repressors (ZBTB7A, PATZ1, ZNF263) could be involved in the fine-tuning of TGFB1 expression in atherosclerosis.

Complex structural and regulatory evolution of the pro-opiomelanocortin gene family

The melanocortin system is a neuroendocrine machinery that has been associated with phenotypic diversification in a number of vertebrate lineages. Central to the highly pleiotropic melanocortin system is the pro-opiomelanocortin (pomc) gene family, a family of pre-prohormones that each give rise to melanocyte stimulating hormone (MSH), adrenocorticotropic releasing hormone (ACTH), β-lipotropin hormone, and β-endorphin. Here we examine the structure, tissue expression profile, and pattern of cis transcriptional regulation of the three pomc paralogs (α1, α2, and β) in the model cichlid fish Astatotilapia burtoni and other cichlids, teleosts, and mammals. We found that the hormone-encoding regions of pomc α1, pomc α2 and pomc β are highly conserved, with a few notable exceptions. Surprisingly, the pomc β gene of cichlids and pomacentrids (damselfish) encodes a novel melanocortin peptide, ε-MSH, as a result of a tandem duplication of the segment encoding ACTH. All three genes are expressed in the brain and peripheral tissues, but pomc α1 and α2 show a more spatially restricted expression profile than pomc β. In addition, the promoters of each pomc gene have diverged in nucleotide sequence, which may have facilitated the diverse tissue-specific expression profiles of these paralogs across species. Increased understanding of the mechanisms regulating pomc gene expression will be invaluable to the study of pomc in the context of phenotypic evolution.

Reverse engineering to suggest biologically relevant redox activities of phenolic materials

Phenolics are among the most abundant redox-active organics in nature, but the intractability of phenolic materials (e.g., melanin) has precluded study of their biological activities and functions. Previous studies demonstrated that a model abiotic catecholic matrix can rapidly exchange electrons with biological oxidants and reductants without the need for enzymes. Here, a novel electrochemically based reverse engineering approach was employed to probe redox interactions between this model matrix and a population of bacteria. Specifically, this method employs redox-active natural products (e.g., pyocyanin) to shuttle electrons between the bacteria and the abiotic matrix, and imposed oscillating potential inputs to engage redox-cycling mechanisms that switch the matrix's redox state. The oscillating output currents were observed to be amplified, gated, and partially rectified, while the overall magnitude and direction of electron flow across the matrix depended on the biological and environmental context. These response characteristics support hypotheses that natural phenolic materials may be integral to extracellular electron transport for processes that include anaerobic respiration, redox signaling, and redox-effector action.

STIFDB2: an updated version of plant stress-responsive transcription factor database with additional stress signals, stress-responsive transcription factor binding sites and stress-responsive genes in Arabidopsis and rice

Understanding the principles of abiotic and biotic stress responses, tolerance and adaptation remains important in plant physiology research to develop better varieties of crop plants. Better understanding of plant stress response mechanisms and application of knowledge derived from integrated experimental and bioinformatics approaches are gaining importance. Earlier, we showed that compiling a database of stress-responsive transcription factors and their corresponding target binding sites in the form of Hidden Markov models at promoter, untranslated and upstream regions of stress-up-regulated genes from expression analysis can help in elucidating various aspects of the stress response in Arabidopsis. In addition to the extensive content in the first version, STIFDB2 is now updated with 15 stress signals, 31 transcription factors and 5,984 stress-responsive genes from three species (Arabidopsis thaliana, Oryza sativa subsp. japonica and Oryza sativa subsp. indica). We have employed an integrated biocuration and genomic data mining approach to characterize the data set of transcription factors and consensus binding sites from literature mining and stress-responsive genes from the Gene Expression Omnibus. STIFDB2 currently has 38,798 associations of stress signals, stress-responsive genes and transcription factor binding sites predicted using the Stress-responsive Transcription Factor (STIF) algorithm, along with various functional annotation data. As a unique plant stress regulatory genomics data platform, STIFDB2 can be utilized for targeted as well as high-throughput experimental and computational studies to unravel principles of the stress regulome in dicots and gramineae. STIFDB2 is available from the URL: http://caps.ncbs.res.in/stifdb2.

A validated regulatory network for Th17 cell specification

Th17 cells have critical roles in mucosal defense and are major contributors to inflammatory disease. Their differentiation requires the nuclear hormone receptor RORγt working with multiple other essential transcription factors (TFs). We have used an iterative systems approach, combining genome-wide TF occupancy, expression profiling of TF mutants, and expression time series to delineate the Th17 global transcriptional regulatory network. We find that cooperatively bound BATF and IRF4 contribute to initial chromatin accessibility and, with STAT3, initiate a transcriptional program that is then globally tuned by the lineage-specifying TF RORγt, which plays a focal deterministic role at key loci. Integration of multiple data sets allowed inference of an accurate predictive model that we computationally and experimentally validated, identifying multiple new Th17 regulators, including Fosl2, a key determinant of cellular plasticity. This interconnected network can be used to investigate new therapeutic approaches to manipulate Th17 functions in the setting of inflammatory disease.

Computational prediction of the polyQ and CAG repeat spinocerebellar ataxia network based on sequence identity to untranslated regions

Computational prediction of biological networks would be a tremendous asset to systems biology and personalized medicine. In this paper, we use a moving window bioinformatic screen to identify transcripts with partial identity to the 5' and 3'UTRs of the polyQ spinocerebellar ataxia (SCA) genes ATXN1, ATXN2, ATXN3, ATXN7, TBP and CACNA1A and the CAG repeat expansion gene PPP2R2B. We find that the bioinformatic screen enriches for transcripts that encode proteins that interact and that have functions relevant to polyQ SCA. Transcription control and RNA binding are the primary functional groups represented in the proteins from the combined screens. The insulin growth factor pathway, the WNT pathway, long term potentiation, melanogenesis and ATM mediated DNA repair pathways were identified as important pathways. UGUUU repeats were identified as an abundant motif in the SCA network and PAXIP1, CELF2, CREBBP, EBF1, PLEKHG4, SRSF4, C5orf42, NFIA, STK24, and YWHAG were identified as statistically significant proteins in the polyQ and PPP2R2B network.

Regulatory link mapping between organisms

Background

Identification of gene regulatory networks is useful in understanding gene regulation in any organism. Some regulatory network information has already been determined experimentally for model organisms, but much less has been identified for non-model organisms, and the limited amount of gene expression data available for non-model organisms makes inference of regulatory networks difficult.

Results

This paper proposes a method to determine the regulatory links that can be mapped from a model to a non-model organism. Mapping a regulatory network involves mapping the transcription factors and target genes from one genome to another. In the proposed method, Basic Local Alignment Search Tool (BLAST) and InterProScan are used to map the transcription factors, whereas BLAST along with transcription factor binding site motifs and the GALF-P tool are used to map the target genes. Experiments are performed to map the regulatory network data of S. cerevisiae to A. thaliana and analyze the results. Since limited information is available about gene regulatory network links, gene expression data is used to analyze results. A set of rules are defined on the gene expression experiments to identify the predicted regulatory links that are well supported.

Conclusions

Combining transcription factors mapped using BLAST and subfamily classification, together with target genes mapped using BLAST and binding site motifs, produced the best regulatory link predictions. More than two-thirds of these predicted regulatory links that were analyzed using gene expression data have been verified as correctly mapped regulatory links in the target genome.

Genome-wide histone acetylation data improve prediction of mammalian transcription factor binding sites

Motivation: Histone acetylation (HAc) is associated with open chromatin, and HAc has been shown to facilitate transcription factor (TF) binding in mammalian cells. In the innate immune system context, epigenetic studies strongly implicate HAc in the transcriptional response of activated macrophages. We hypothesized that using data from large-scale sequencing of a HAc chromatin immunoprecipitation assay (ChIP-Seq) would improve the performance of computational prediction of binding locations of TFs mediating the response to a signaling event, namely, macrophage activation.

Results: We tested this hypothesis using a multi-evidence approach for predicting binding sites. As a training/test dataset, we used ChIP-Seq-derived TF binding site locations for five TFs in activated murine macrophages. Our model combined TF binding site motif scanning with evidence from sequence-based sources and from HAc ChIP-Seq data, using a weighted sum of thresholded scores. We find that using HAc data significantly improves the performance of motif-based TF binding site prediction. Furthermore, we find that within regions of high HAc, local minima of the HAc ChIP-Seq signal are particularly strongly correlated with TF binding locations. Our model, using motif scanning and HAc local minima, improves the sensitivity for TF binding site prediction by ∼50% over a model based on motif scanning alone, at a false positive rate cutoff of 0.01.

Availability: The data and software source code for model training and validation are freely available online at http://magnet.systemsbiology.net/hac.

Contact:aderem@systemsbiology.org; ishmulevich@systemsbiology.org

Supplementary information:Supplementary data are available at Bioinformatics online.