scholarly journals HebbPlot: An intelligent tool for learning and visualizing chromatin mark signatures

2017 ◽  
Author(s):  
Hani Z Girgis ◽  
Alfredo Velasco ◽  
Zachary E Reyes
Keyword(s):  
Neural Network ◽  
Gene Regulation ◽  
Case Studies ◽  
Histone Modifications ◽  
Software Tool ◽  
Histone Code ◽  
Human Diseases ◽  
Computational Tools ◽  
Chromatin Mark

AbstractHistone modifications play important roles in gene regulation, heredity, imprinting, and many human diseases. The histone code is complex, consisting of about 100 marks. Biologists need computational tools for characterizing general signatures representing the distributions of tens of chromatin marks around thousands of regions. To this end, we developed a software tool called HebbPlot, which utilizes a Hebbian neural network to learn such signatures. HebbPlot presents a signature as a digitized image, which can be easily interpreted. We validated HebbPlot in six case studies. HebbPlot is applicable to a wide array of studies, facilitating the deciphering of the histone code.

Binary Precision Neural Network Manycore Accelerator

2021 ◽  
Vol 17 (2) ◽  
pp. 1-27
Author(s):  
Morteza Hosseini ◽  
Tinoosh Mohsenin
Keyword(s):  
Neural Network ◽  
Low Power ◽  
Image Classification ◽  
Case Studies ◽  
Average Power ◽  
Total Power ◽  
Fabrication Technology ◽  
Population Count ◽  
Cluster Architecture ◽  
Domain Specific

This article presents a low-power, programmable, domain-specific manycore accelerator, Binarized neural Network Manycore Accelerator (BiNMAC), which adopts and efficiently executes binary precision weight/activation neural network models. Such networks have compact models in which weights are constrained to only 1 bit and can be packed several in one memory entry that minimizes memory footprint to its finest. Packing weights also facilitates executing single instruction, multiple data with simple circuitry that allows maximizing performance and efficiency. The proposed BiNMAC has light-weight cores that support domain-specific instructions, and a router-based memory access architecture that helps with efficient implementation of layers in binary precision weight/activation neural networks of proper size. With only 3.73% and 1.98% area and average power overhead, respectively, novel instructions such as Combined Population-Count-XNOR , Patch-Select , and Bit-based Accumulation are added to the instruction set architecture of the BiNMAC, each of which replaces execution cycles of frequently used functions with 1 clock cycle that otherwise would have taken 54, 4, and 3 clock cycles, respectively. Additionally, customized logic is added to every core to transpose 16×16-bit blocks of memory on a bit-level basis, that expedites reshaping intermediate data to be well-aligned for bitwise operations. A 64-cluster architecture of the BiNMAC is fully placed and routed in 65-nm TSMC CMOS technology, where a single cluster occupies an area of 0.53 mm 2 with an average power of 232 mW at 1-GHz clock frequency and 1.1 V. The 64-cluster architecture takes 36.5 mm 2 area and, if fully exploited, consumes a total power of 16.4 W and can perform 1,360 Giga Operations Per Second (GOPS) while providing full programmability. To demonstrate its scalability, four binarized case studies including ResNet-20 and LeNet-5 for high-performance image classification, as well as a ConvNet and a multilayer perceptron for low-power physiological applications were implemented on BiNMAC. The implementation results indicate that the population-count instruction alone can expedite the performance by approximately 5×. When other new instructions are added to a RISC machine with existing population-count instruction, the performance is increased by 58% on average. To compare the performance of the BiNMAC with other commercial-off-the-shelf platforms, the case studies with their double-precision floating-point models are also implemented on the NVIDIA Jetson TX2 SoC (CPU+GPU). The results indicate that, within a margin of ∼2.1%--9.5% accuracy loss, BiNMAC on average outperforms the TX2 GPU by approximately 1.9× (or 7.5× with fabrication technology scaled) in energy consumption for image classification applications. On low power settings and within a margin of ∼3.7%--5.5% accuracy loss compared to ARM Cortex-A57 CPU implementation, BiNMAC is roughly ∼9.7×--17.2× (or 38.8×--68.8× with fabrication technology scaled) more energy efficient for physiological applications while meeting the application deadline.

scholarly journals FIREcaller: Detecting Frequently Interacting Regions from Hi-C Data

2019 ◽  
Author(s):  
Cheynna Crowley ◽  
Yuchen Yang ◽  
Yunjiang Qiu ◽  
Benxia Hu ◽  
Armen Abnousi ◽  
...  
Keyword(s):  
Gene Regulation ◽  
Spatial Organization ◽  
R Package ◽  
Specific Gene ◽  
List Type ◽  
Cell Type ◽  
R Software ◽  
Computational Tools ◽  
Cell Type Specific ◽  
User Friendly

AbstractHi-C experiments have been widely adopted to study chromatin spatial organization, which plays an essential role in genome function. We have recently identified frequently interacting regions (FIREs) and found that they are closely associated with cell-type-specific gene regulation. However, computational tools for detecting FIREs from Hi-C data are still lacking. In this work, we present FIREcaller, a stand-alone, user-friendly R package for detecting FIREs from Hi-C data. FIREcaller takes raw Hi-C contact matrices as input, performs within-sample and cross-sample normalization, and outputs continuous FIRE scores, dichotomous FIREs, and super-FIREs. Applying FIREcaller to Hi-C data from various human tissues, we demonstrate that FIREs and super-FIREs identified, in a tissue-specific manner, are closely related to gene regulation, are enriched for enhancer-promoter (E-P) interactions, tend to overlap with regions exhibiting epigenomic signatures of cis-regulatory roles, and aid the interpretation or GWAS variants. The FIREcaller package is implemented in R and freely available at https://yunliweb.its.unc.edu/FIREcaller.Highlights– Frequently Interacting Regions (FIREs) can be used to identify tissue and cell-type-specific cis-regulatory regions.– An R software, FIREcaller, has been developed to identify FIREs and clustered FIREs into super-FIREs.

scholarly journals Deciphering gene regulation from gene expression dynamics using deep neural network

2018 ◽  
Author(s):  
Jingxiang Shen ◽  
Mariela D. Petkova ◽  
Yuhai Tu ◽  
Feng Liu ◽  
Chao Tang
Keyword(s):  
Neural Network ◽  
Gene Expression ◽  
Machine Learning ◽  
Gene Regulation ◽  
Gene Network ◽  
Deep Neural Network ◽  
Biological Functions ◽  
Gene Regulation Network ◽  
Regulation Network ◽  
Gene Expression Dynamics

AbstractComplex biological functions are carried out by the interaction of genes and proteins. Uncovering the gene regulation network behind a function is one of the central themes in biology. Typically, it involves extensive experiments of genetics, biochemistry and molecular biology. In this paper, we show that much of the inference task can be accomplished by a deep neural network (DNN), a form of machine learning or artificial intelligence. Specifically, the DNN learns from the dynamics of the gene expression. The learnt DNN behaves like an accurate simulator of the system, on which one can perform in-silico experiments to reveal the underlying gene network. We demonstrate the method with two examples: biochemical adaptation and the gap-gene patterning in fruit fly embryogenesis. In the first example, the DNN can successfully find the two basic network motifs for adaptation – the negative feedback and the incoherent feed-forward. In the second and much more complex example, the DNN can accurately predict behaviors of essentially all the mutants. Furthermore, the regulation network it uncovers is strikingly similar to the one inferred from experiments. In doing so, we develop methods for deciphering the gene regulation network hidden in the DNN “black box”. Our interpretable DNN approach should have broad applications in genotype-phenotype mapping.SignificanceComplex biological functions are carried out by gene regulation networks. The mapping between gene network and function is a central theme in biology. The task usually involves extensive experiments with perturbations to the system (e.g. gene deletion). Here, we demonstrate that machine learning, or deep neural network (DNN), can help reveal the underlying gene regulation for a given function or phenotype with minimal perturbation data. Specifically, after training with wild-type gene expression dynamics data and a few mutant snapshots, the DNN learns to behave like an accurate simulator for the genetic system, which can be used to predict other mutants’ behaviors. Furthermore, our DNN approach is biochemically interpretable, which helps uncover possible gene regulatory mechanisms underlying the observed phenotypic behaviors.

scholarly journals Phosphorylation of repressive histone code readers by casein kinase 2 plays diverse roles in heterochromatin regulation

2019 ◽  
Vol 166 (1) ◽  
pp. 3-6 ◽  
Author(s):  
Yota Murakami
Keyword(s):  
Histone Modifications ◽  
Constitutive Heterochromatin ◽  
Histone H3 ◽  
Casein Kinase ◽  
Casein Kinase 2 ◽  
Histone Code ◽  
Effector Proteins ◽  
Regulatory Proteins ◽  
Silent Chromatin ◽  
Facultative Heterochromatin

Abstract Heterochromatin is a condensed and transcriptionally silent chromatin structure and that plays important roles in epigenetic regulation of the genome. Two types of heterochromatin exist: constitutive heterochromatin is primarily associated with trimethylation of histone H3 at lysine 9 (H3K9me3), and facultative heterochromatin with trimethylation of H3 at lysine 27 (H3K27me3). The methylated histones are bound by the chromodomain of histone code ‘reader’ proteins: HP1 family proteins for H3K9me3 and Polycomb family proteins for H3K27me3. Each repressive reader associates with various ‘effector’ proteins that provide the functional basis of heterochromatin. Heterochromatin regulation is primarily achieved by controlling histone modifications. However, recent studies have revealed that the repressive readers are phosphorylated, like other regulatory proteins, suggesting that phosphorylation also participates in heterochromatin regulation. Detailed studies have shown that phosphorylation of readers affects the binding specificities of chromodomains for methylated histone H3, as well as the binding of effector proteins. Thus, phosphorylation adds another layer to heterochromatin regulation. Interestingly, casein kinase 2, a strong and predominant kinase within the cell, is responsible for phosphorylation of repressive readers. In this commentary, I summarize the regulation of repressive readers by casein kinase 2-dependent phosphorylation and discuss the functional meaning of this modification.

scholarly journals Ash2 acts as an ecdysone receptor coactivator by stabilizing the histone methyltransferase Trr

2013 ◽  
Vol 24 (3) ◽  
pp. 361-372 ◽  
Author(s):  
Albert Carbonell ◽  
Alexander Mazo ◽  
Florenci Serras ◽  
Montserrat Corominas
Keyword(s):  
Gene Regulation ◽  
Histone Modifications ◽  
Transcriptional Activation ◽  
Histone H3 ◽  
Histone Methyltransferase ◽  
Ecdysone Receptor ◽  
Related Protein ◽  
Molting Hormone ◽  
Set Domain ◽  
Present Evidence

The molting hormone ecdysone triggers chromatin changes via histone modifications that are important for gene regulation. On hormone activation, the ecdysone receptor (EcR) binds to the SET domain–containing histone H3 methyltransferase trithorax-related protein (Trr). Methylation of histone H3 at lysine 4 (H3K4me), which is associated with transcriptional activation, requires several cofactors, including Ash2. We find that ash2 mutants have severe defects in pupariation and metamorphosis due to a lack of activation of ecdysone-responsive genes. This transcriptional defect is caused by the absence of the H3K4me3 marks set by Trr in these genes. We present evidence that Ash2 interacts with Trr and is required for its stabilization. Thus we propose that Ash2 functions together with Trr as an ecdysone receptor coactivator.

A Chebyshev polynomial feedforward neural network trained by differential evolution and its application in environmental case studies

2020 ◽  
Vol 126 ◽  
pp. 104663
Author(s):  
Ioannis A. Troumbis ◽  
George E. Tsekouras ◽  
John Tsimikas ◽  
Christos Kalloniatis ◽  
Dias Haralambopoulos
Keyword(s):  
Neural Network ◽  
Differential Evolution ◽  
Case Studies ◽  
Chebyshev Polynomial ◽  
Feedforward Neural Network

Assessing the factor of safety using an artificial neural network: case studies on landslides in Giresun, Turkey

2015 ◽  
Vol 73 (12) ◽  
pp. 8639-8646 ◽  
Author(s):  
Kenan Gelisli ◽  
Türkan Kaya ◽  
Ali Erden Babacan
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Case Studies ◽  
Factor Of Safety ◽  
Artificial Neural

A multi-layer cellular automata approach for algorithmic generation of virtual case studies: VIBe

2010 ◽  
Vol 61 (1) ◽  
pp. 37-45 ◽  
Author(s):  
R. Sitzenfrei ◽  
S. Fach ◽  
H. Kinzel ◽  
W. Rauch
Keyword(s):  
Boundary Conditions ◽  
Case Studies ◽  
Real World ◽  
New Technologies ◽  
Single Case ◽  
Software Tool ◽  
Simulation Software ◽  
Data Availability ◽  
Study Results

Analyses of case studies are used to evaluate new or existing technologies, measures or strategies with regard to their impact on the overall process. However, data availability is limited and hence, new technologies, measures or strategies can only be tested on a limited number of case studies. Owing to the specific boundary conditions and system properties of each single case study, results can hardly be generalized or transferred to other boundary conditions. virtual infrastructure benchmarking (VIBe) is a software tool which algorithmically generates virtual case studies (VCSs) for urban water systems. System descriptions needed for evaluation are extracted from VIBe whose parameters are based on real world case studies and literature. As a result VIBe writes Input files for water simulation software as EPANET and EPA SWMM. With such input files numerous simulations can be performed and the results can be benchmarked and analysed stochastically at a city scale. In this work the approach of VIBe is applied with parameters according to a section of the Inn valley and therewith 1,000 VCSs are generated and evaluated. A comparison of the VCSs with data of real world case studies shows that the real world case studies fit within the parameter ranges of the VCSs. Consequently, VIBe tackles the problem of limited availability of case study data.

Sign in / Sign up

Export Citation Format

Share Document