scholarly journals Uncovering tissue-specific binding features from differential deep learning

2019 ◽  
Author(s):  
Mike Phuycharoen ◽  
Peyman Zarrineh ◽  
Laure Bridoux ◽  
Shilu Amin ◽  
Marta Losa ◽  
...  
Keyword(s):  
Deep Learning ◽  
Binding Sites ◽  
Specific Binding ◽  
Specific Expression ◽  
Tissue Specific ◽  
Dimensional Classification ◽  
Branchial Arches ◽  
Gradient Based ◽  
Differential Binding

ABSTRACTMotivationTranscription factors (TFs) can bind DNA in a cooperative manner, enabling a mutual increase in occupancy. Through this type of interaction, alternative binding sites can be preferentially bound in different tissues to regulate tissue-specific expression programmes. Recently, deep learning models have become state-of-the-art in various pattern analysis tasks, including applications in the field of genomics. We therefore investigate the application of convolutional neural network (CNN) models to the discovery of sequence features determining cooperative and differential TF binding across tissues.ResultsWe analyse ChIP-seq data from MEIS, TFs which are broadly expressed across mouse branchial arches, and HOXA2, which is expressed in the second and more posterior branchial arches. By developing models predictive of MEIS differential binding in all three tissues we are able to accurately predict HOXA2 co-binding sites. We evaluate transfer-like and multitask approaches to regularising the high-dimensional classification task with a larger regression dataset, allowing for creation of deeper and more accurate models. We test the performance of perturbation and gradient-based attribution methods in identifying the HOXA2 sites from differential MEIS data. Our results show that deep regularised models significantly outperform shallow CNNs as well as k-mer methods in the discovery of tissue-specific sites bound in vivo.AvailabilityFor implementation and models please visit https://doi.org/10.5281/zenodo.2635463.

scholarly journals Uncovering tissue-specific binding features from differential deep learning

2020 ◽  
Vol 48 (5) ◽  
pp. e27-e27 ◽  
Author(s):  
Mike Phuycharoen ◽  
Peyman Zarrineh ◽  
Laure Bridoux ◽  
Shilu Amin ◽  
Marta Losa ◽  
...  
Keyword(s):  
Deep Learning ◽  
Binding Sites ◽  
Specific Binding ◽  
Specific Expression ◽  
Tissue Specific ◽  
Dimensional Classification ◽  
Branchial Arches ◽  
Gradient Based ◽  
Differential Binding

Abstract Transcription factors (TFs) can bind DNA in a cooperative manner, enabling a mutual increase in occupancy. Through this type of interaction, alternative binding sites can be preferentially bound in different tissues to regulate tissue-specific expression programmes. Recently, deep learning models have become state-of-the-art in various pattern analysis tasks, including applications in the field of genomics. We therefore investigate the application of convolutional neural network (CNN) models to the discovery of sequence features determining cooperative and differential TF binding across tissues. We analyse ChIP-seq data from MEIS, TFs which are broadly expressed across mouse branchial arches, and HOXA2, which is expressed in the second and more posterior branchial arches. By developing models predictive of MEIS differential binding in all three tissues, we are able to accurately predict HOXA2 co-binding sites. We evaluate transfer-like and multitask approaches to regularizing the high-dimensional classification task with a larger regression dataset, allowing for the creation of deeper and more accurate models. We test the performance of perturbation and gradient-based attribution methods in identifying the HOXA2 sites from differential MEIS data. Our results show that deep regularized models significantly outperform shallow CNNs as well as k-mer methods in the discovery of tissue-specific sites bound in vivo.

scholarly journals FoxA Proteins Regulate H19 Endoderm Enhancer E1 and Exhibit Developmental Changes in Enhancer Binding In Vivo

2004 ◽  
Vol 24 (21) ◽  
pp. 9601-9609 ◽  
Author(s):  
Lingyun Long ◽  
Brett T. Spear
Keyword(s):  
Chromatin Immunoprecipitation ◽  
Binding Sites ◽  
Fetal Liver ◽  
Carcinoma Cells ◽  
Specific Factor ◽  
Specific Expression ◽  
Tissue Specific ◽  
Parietal Endoderm ◽  
F9 Embryonal Carcinoma Cells

ABSTRACT Multiple enhancers govern developmental and tissue-specific expression of the H19-Igf2 locus, but factors that bind these elements have not been identified. Using chromatin immunoprecipitation, we have found two FoxA binding sites in the H19 E1 enhancer. Mutating these sites diminishes E1 activity in hepatoma cells. Additional chromatin immunoprecipitations show that FoxA binds to E1 in fetal liver, where H19 is abundantly expressed, but that binding decreases in adult liver, where H19 is no longer transcribed, even though FoxA proteins are present at both times. FoxA proteins are induced when F9 embryonal carcinoma cells differentiate into visceral endoderm (VE) and parietal endoderm (PE). We show that FoxA binds E1 in VE cells, where H19 is expressed, but not in PE cells, where H19 is silent. This correlation between FoxA binding and H19 expression indicates a role for FoxA in regulating H19, including developmental activation in the yolk sac and liver and postnatal repression in the liver. This is the first demonstration of a tissue-specific factor involved in developmental control of H19 expression. These data also indicate that the presence of FoxA proteins is not sufficient for binding but that additional mechanisms must govern the accessibility of FoxA proteins to their cognate binding sites within the H19 E1 enhancer.

scholarly journals Two-Step Regulation of Ad4BP/SF-1 Gene Transcription during Fetal Adrenal Development: Initiation by a Hox-Pbx1-Prep1 Complex and Maintenance via Autoregulation by Ad4BP/SF-1

2006 ◽  
Vol 26 (11) ◽  
pp. 4111-4121 ◽  
Author(s):  
Mohamad Zubair ◽  
Satoru Ishihara ◽  
Sanae Oka ◽  
Katsuzumi Okumura ◽  
Ken-ichirou Morohashi
Keyword(s):  
Adrenal Cortex ◽  
Binding Sites ◽  
Orphan Nuclear Receptor ◽  
Specific Expression ◽  
Steroidogenic Factor 1 ◽  
Tissue Specific ◽  
Adrenal Cells ◽  
Tissue Specific Expression ◽  
And Function ◽  
Intron 4

ABSTRACT The orphan nuclear receptor Ad4BP/SF-1 (adrenal 4 binding protein/steroidogenic factor 1) is essential for the proper development and function of reproductive and steroidogenic tissues. Although the expression of Ad4BP/SF-1 is specific for those tissues, the mechanisms underlying this tissue-specific expression remain unknown. In this study, we used transgenic mouse assays to examine the regulation of the tissue-specific expression of Ad4BP/SF-1. An investigation of the entire Ad4BP/SF-1 gene locus revealed a fetal adrenal enhancer (FAdE) in intron 4 containing highly conserved binding sites for Pbx-Prep, Pbx-Hox, and Ad4BP/SF-1. Transgenic assays revealed that the Ad4 sites, together with Ad4BP/SF-1, develop an autoregulatory loop and thereby maintain transcription, while the Pbx/Prep and Pbx/Hox sites initiate transcription prior to the establishment of the autoregulatory loop. Indeed, a limited number of Hox family members were found to be expressed in the adrenal primordia. Whether a true fetal-type adrenal cortex is present in mice remained controversial, and this argument was complicated by the postnatal development of the so-called X zone. Using transgenic mice with lacZ driven by the FAdE, we clearly identified a fetal adrenal cortex in mice, and the X zone is the fetal adrenal cells accumulated at the juxtamedullary region after birth.

Tissue-specific transcription of the mouse alpha-fetoprotein gene promoter is dependent on HNF-1

1989 ◽  
Vol 9 (10) ◽  
pp. 4204-4212
Author(s):  
M H Feuerman ◽  
R Godbout ◽  
R S Ingram ◽  
S M Tilghman
Keyword(s):  
Specific Binding ◽  
Gene Promoter ◽  
Binding Activity ◽  
Alpha Fetoprotein ◽  
Transient Expression Assay ◽  
Band Shift ◽  
Specific Expression ◽  
Tissue Specific ◽  
Cis Acting ◽  
Dnase I Footprinting

Previous work identified four upstream cis-acting elements required for tissue-specific expression of the alpha-fetoprotein (AFP) gene: three distal enhancers and a promoter. To further define the role of the promoter in regulating AFP gene expression, segments of the region were tested for the ability to direct transcription of a reporter gene in transient expression assay. Experiments showed that the region within 250 base pairs of the start of transcription was sufficient to confer liver-specific transcription. DNase I footprinting and band shift assays indicated that the region between -130 and -100 was recognized by two factors, one of which was highly sequence specific and found only in hepatoma cells. Competition assays suggested that the liver-specific binding activity was HNF-1, previously identified by its binding to other liver-specific promoters. Mutation of the HNF-1 recognition site at -120 resulted in a significant reduction in transcription in transfection assays, suggesting a biological role for HNF-1 in the regulation of AFP expression.

In vitro and in vivo analysis of murine lipoprotein lipase gene promoter: tissue-specific expression

1995 ◽  
Vol 268 (2) ◽  
pp. E213-E218 ◽  
Author(s):  
J. M. Gimble ◽  
X. Hua ◽  
F. Wanker ◽  
C. Morgan ◽  
C. Robinson ◽  
...  
Keyword(s):  
Transgenic Mice ◽  
Lipoprotein Lipase ◽  
Reporter Gene ◽  
Luciferase Reporter ◽  
Luciferase Reporter Gene ◽  
Specific Expression ◽  
Tissue Specific ◽  
Brown Adipose ◽  
Octamer Motif

Lipoprotein lipase, an enzyme of central importance to lipid metabolism, is most abundant in adipose tissues, cardiac and skeletal muscle, and portions of the brain. The current work examined the murine lipoprotein lipase promoter using transient transfection, gel-retention analyses, and transgenic mice. Maximum expression of the luciferase reporter gene in transfected cells was observed with -101 bp of the promoter. Nuclear extracts from tissues expressing lipoprotein lipase contained DNA binding proteins that recognize the CCAAT box (-64 bp) and an octamer motif (-46 bp); this combination of factors was absent in nonexpressing tissues. Transgenic mice from three of five founders prepared with -1,824-bp promoter constructs expressed the luciferase reporter gene at highest levels in brown adipose tissue and brain. These findings suggest that the -1,824-bp promoter region contains sequence elements responsible for the tissue-specific transcription of lipoprotein lipase in vivo.

Endothelin ETA and ETB receptors in postnatal intestine

2001 ◽  
Vol 280 (4) ◽  
pp. G555-G562 ◽  
Author(s):  
Craig A. Nankervis ◽  
David J. Dunaway ◽  
Charles E. Miller
Keyword(s):  
Binding Sites ◽  
Specific Binding ◽  
Ischemia Reperfusion ◽  
Age Groups ◽  
Scatchard Analysis ◽  
Binding Assays ◽  
Site Model ◽  
Number Of Binding Sites ◽  
Competitive Binding Assays

We aimed to characterize endothelin (ET) receptors in the swine intestinal vasculature and to determine ischemia-reperfusion (I/R) effects on these receptors. Saturation and competitive binding assays were performed on mesenteric artery protein membranes from 1- and 40-day-old animals, both control and those subjected to 1 h of partial ischemia followed by 6 h of reperfusion in vivo. Scatchard analysis of saturation binding with 125I-labeled ET-1 in membranes from endothelium-denuded (E−) vessels revealed that the maximum number of binding sites was greater in younger animals. Competitive125I-ET-1 binding was significant for a one-site model with ET-1, ET-3, and sarafotoxin S6c (S6c) in membranes from endothelium-intact (E+) and E− vessels in both age groups. The maximum number of ET-1 binding sites was significantly greater in younger animals. In the presence of the ETAreceptor antagonist BQ-123, competitive 125I-ET-1 binding was significant for a one-site model with ET-1 and S6c in membranes from E+ vessels in both age groups. The maximum number of ET-1 binding sites was significantly greater in younger animals. After I/R, the maximum number of ET-1 binding sites was unchanged. In the presence of BQ-123, specific binding by ET-1 and S6c was eliminated in both age groups after I/R. These results suggest that both ET receptor populations are expressed to a greater degree in younger animals and I/R significantly affects the ETB receptor.

Atrial natriuretic factor binding sites in the jejunum

1989 ◽  
Vol 256 (2) ◽  
pp. G436-G441 ◽  
Author(s):  
C. Bianchi ◽  
G. Thibault ◽  
A. De Lean ◽  
J. Genest ◽  
M. Cantin
Keyword(s):  
Binding Sites ◽  
Atrial Natriuretic Factor ◽  
Specific Binding ◽  
Rat Tissues ◽  
Rat Jejunum ◽  
Specific Binding Sites ◽  
Natriuretic Factor ◽  
Atrial Natriuretic

We have studied the localization and the characterization of atrial natriuretic factor (ANF) binding sites by radioautographic techniques. Quantitative in vitro radioautography with a computerized microdensitometer demonstrated the presence of high-affinity, low-capacity 125I-ANF-(99-126) binding sites (Kd, 48 pM; Bmax, 63 fmol/mg protein) mainly in the villi of 20-microns slide-mounted transverse sections of the rat jejunum. Competition curves showed 50% inhibitory concentrations of 55 and 1,560 pM for ANF-(99-126) and ANF-(103-123), respectively. In vivo electron microscope radioautography showed that 80% of the silver grains were localized on the lamina propria fibroblast-like cells, 18% on mature enterocytes, and 2% on capillaries. Bradykinin and adrenocorticotropin did not compete with ANF binding. These results demonstrate that ANF binding sites in the rat jejunum possess the pharmacological characteristics of functional ANF receptors encountered in other rat tissues, and ultrastructural radioautographs show their cellular distribution. Taken together, these results demonstrate the presence and the localization of specific binding sites for ANF in the jejunal villi of the rat small intestine.

scholarly journals Binding of disparate transcriptional activators to nucleosomal DNA is inherently cooperative.

1995 ◽  
Vol 15 (3) ◽  
pp. 1405-1421 ◽  
Author(s):  
C C Adams ◽  
J L Workman
Keyword(s):  
Transcription Factors ◽  
Dna Binding ◽  
Binding Sites ◽  
Specific Binding ◽  
General Mechanism ◽  
Nf Kappa B ◽  
Nucleosomal Dna ◽  
Nucleosome Binding

To investigate mechanisms by which multiple transcription factors access complex promoters and enhancers within cellular chromatin, we have analyzed the binding of disparate factors to nucleosome cores. We used a purified in vitro system to analyze binding of four activator proteins, two GAL4 derivatives, USF, and NF-kappa B (KBF1), to reconstituted nucleosome cores containing different combinations of binding sites. Here we show that binding of any two or all three of these factors to nucleosomal DNA is inherently cooperative. Thus, the binuclear Zn clusters of GAL4, the helix-loop-helix/basic domains of USF, and the rel domain of NF-kappa B all participated in cooperative nucleosome binding, illustrating that this effect is not restricted to a particular DNA-binding domain. Simultaneous binding by two factors increased the affinity of individual factors for nucleosomal DNA by up to 2 orders of magnitude. Importantly, cooperative binding resulted in efficient nucleosome binding by factors (USF and NF-kappa B) which independently possess little nucleosome-binding ability. The participation of GAL4 derivatives in cooperative nucleosome binding required only DNA-binding and dimerization domains, indicating that disruption of histone-DNA contacts by factor binding was responsible for the increased affinity of additional factors. Cooperative nucleosome binding required sequence-specific binding of all transcription factors, appeared to have spatial constraints, and was independent of the orientation of the binding sites on the nucleosome. These results indicate that cooperative nucleosome binding is a general mechanism that may play a significant role in loading complex enhancer and promoter elements with multiple diverse factors in chromatin and contribute to the generation of threshold responses and transcriptional synergy by multiple activator sites in vivo.

Evidence for the existence of F2-isoprostane receptors on rat vascular smooth muscle cells

1993 ◽  
Vol 264 (6) ◽  
pp. C1619-C1624 ◽  
Author(s):  
M. Fukunaga ◽  
N. Makita ◽  
L. J. Roberts ◽  
J. D. Morrow ◽  
K. Takahashi ◽  
...  
Keyword(s):  
Smooth Muscle ◽  
Vascular Smooth Muscle ◽  
Smooth Muscle Cells ◽  
Binding Sites ◽  
Tissue Injury ◽  
Specific Binding ◽  
Muscle Cells ◽  
Free Oxygen Radicals ◽  
Txa2 Receptor

The isoprostanes are nonenzymatically generated prostanoids synthesized in vivo in humans and rats through reactions catalyzed by free oxygen radicals. 8-Epi-prostaglandin F2 alpha (8-epi-PGF2 alpha), an F2-isoprostane, is a potent smooth muscle constrictor. A thromboxane A2 (TxA2) receptor antagonist, SQ 29548, blocks renal vasoconstriction during 8-epi-PGF2 alpha administration in rats. With the use of cultured rat aortic smooth muscle cells, we found specific binding sites for [3H]SQ 29548 and for [125I]BOP, a TxA2 agonist. Both ligands were displaced from these binding sites by 8-epi-PGF2 alpha, although with significantly lesser potency than nonlabeled SQ 29548, I-BOP, or U-46619, a TxA2 agonist. In contrast, 8-epi-PGF2 alpha stimulated inositol 1,4,5-trisphosphate production and DNA synthesis in these cells with significantly greater potency than any TxA2 agonist, effects only partially inhibited by SQ 29548. In human TxA2 receptor cDNA-transfected cells, competition by 8-epi-PGF2 alpha for specific [3H]SQ 29548 binding was negligible. Thus 8-epi-PGF2 alpha probably exerts its biological actions in vascular smooth muscle through activation of receptor sites related to but distinct from TxA2 receptors. The existence of such binding sites suggests novel avenues for investigation into the biology of TxA2 and of free radical-mediated tissue injury.

Human glucagon gene promoter sequences regulating tissue-specific versus nutrient-regulated gene expression

2002 ◽  
Vol 282 (1) ◽  
pp. R173-R183 ◽  
Author(s):  
Min Nian ◽  
Jun Gu ◽  
David M. Irwin ◽  
Daniel J. Drucker
Keyword(s):  
Gene Expression ◽  
Gene Promoter ◽  
Regulatory Sequences ◽  
Dependent Manner ◽  
Specific Expression ◽  
Tissue Specific ◽  
High Fiber ◽  
Fiber Diet ◽  
Regulated Gene Expression

The glucagon-like peptides (GLPs) are synthesized and secreted in a nutrient-dependent manner in rodents; however, the factors regulating human GLP-1 and GLP-2 biosynthesis remain unclear. To understand how nutrients regulate human proglucagon gene expression, we studied the expression of a human proglucagon promoter-growth hormone (GH) transgene in 1.6 human glucagon-GH transgenic mice. Fasting-refeeding significantly decreased and increased the levels of circulating mouse insulin and transgene-derived hGH ( P < 0.05 fasting vs. refeeding) and decreased and upregulated, respectively, the levels of endogenous mouse proglucagon RNA in the ileum but not in the jejunum or colon. High-fiber feeding significantly increased the levels of glucose-stimulated circulating hGH and upregulated levels of mouse intestinal proglucagon gene expression in the jejunum, ileum, and colon ( P < 0.05, 0 vs. 30% fiber diet). In contrast, neither fasting-refeeding nor a high-fiber diet upregulated the expression of the human proglucagon promoter-hGH transgene. These findings demonstrate that human proglucagon gene regulatory sequences specifying tissue-specific expression in gut endocrine cells are not sufficient for recognition of energy-derived signals regulating murine glucagon gene expression in enteroendocrine cells in vivo.

Sign in / Sign up

Export Citation Format

Share Document