检索结果-内蒙古大学图书馆

Higher-Order Generalization Bounds: Learning Deep Probabilistic programs via PAC-Bayes Objectives

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Warrell, Jonathan Gerstein, Mark Program in Computational Biology and Bioinformatics Department of Molecular Biophysics and Biochemistry Yale University New HavenCT06520 United States

Deep Probabilistic programming (DPP) allows powerful models based on recursive computation to be learned using efficient deep-learning optimization techniques. Additionally, DPP offers a unified perspective, where inference and learning algorithms are treated on a par with models as stochastic programs. Here, we offer a framework for representing and learning flexible PAC-Bayes bounds as stochastic programs using DPP-based methods. In particular, we show that DPP techniques may be leveraged to derive generalization bounds that draw on the compositionality of DPP representations. In turn, the bounds we introduce offer principled training objectives for higher-order probabilistic programs. We offer a definition of a higher-order generalization bound, which naturally encompasses single- and multi-task generalization perspectives (including transfer- and meta-learning) and a novel class of bound based on a learned measure of model complexity. Further, we show how modified forms of all higher-order bounds can be efficiently optimized as objectives for DPP training, using variational techniques. We test our framework using single- and multi-task generalization settings on synthetic and biological data, showing improved performance and generalization prediction using flexible DPP model representations and learned complexity measures. Copyright © 2022, The Authors. All rights reserved.

关键词： Deep learning

High torsional energy disulfides: Relationship between cross-strand disulfides and right-handed staples

学校读者我要写书评

暂无评论

Journal of bioinformatics and computational biology 2006年第1期4卷 155-168页

作者： Haworth, Naomi L. Feng, Lina L. Wouters, Merridee A. Computational Biology and Bioinformatics Program Victor Chang Cardiac Research Institute Darlinghurst NSW 2010 384 Victoria Rd. Australia School of Chemistry University of Sydney Sydney NSW 2006 Australia School of Biotechnology and Biomolecular Sciences University of New South Wales Sydney Sydney NSW 2052 Australia

Redox-active disulfides are capable of being oxidized and reduced under physiological conditions. The enzymatic role of redox-active disulfides in thiol-disulfide reductases is well-known, but redox-active disulfides are also present in non-enzymatic protein structures where they may act as switches of protein function. Here, we examine disulfides linking adjacent β-strands (cross-strand disulfides), which have been reported to be redox-active. Our previous work has established that these cross-strand disulfides have high torsional energies, a quantity likely to be related to the ease with which the disulfide is reduced. We examine the relationship between conformations of disulfides and their location in protein secondary structures. By identifying the overlap between cross-strand disulfides and various conformations, we wish to address whether the high torsional energy of a cross-strand disulfide is sufficient to confer redox activity or whether other factors, such as the presence of the cross-strand disulfide in a strained β-sheet, are required. © Imperial College Press.

关键词： Cross-strand disulfide Disulfide Redox signaling Right-handed staple

Reliable uncertainty estimate for antibiotic resistance classification with Stochastic Gradient Langevin Dynamics

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Hamid, Md-Nafiz Friedberg, Iddo Bioinformatics and Computational Biology Program Department of Veterinary Microbiology and Preventive Medicine Iowa State University AmesIA United States

Antibiotic resistance monitoring is of paramount importance in the face of this ongoing global epidemic. Deep learning models trained with traditional optimization algorithms (e.g. Adam, SGD) provide poor posterior estimates when tested against out-of-distribution (OoD) antibiotic resistant/non-resistant genes. In this paper, we introduce a deep learning model trained with Stochastic Gradient Langevin Dynamics (SGLD) to classify antibiotic resistant genes. The model provides better uncertainty estimates when tested against OoD data compared to traditional optimization methods such as Adam. Copyright © 2018, The Authors. All rights reserved.

关键词： Stochastic systems

Gene Co-expression Network analysis of Lung Squamous Cell Carcinoma data

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Hamid, Md-Nafiz Bioinformatics and Computational Biology program Department of Veterinary Microbiology and Preventive Medicine Iowa State University AmesIA United States

We performed a gene co-expression analysis on Lung Squamous Cell Carcinoma data to find modules (groups) of genes that may highly impact the growth of these type of tumors. Additionally, we used cancer survival data to relate modules to prognostic significance in terms of survival time. Analysis on RNA-seq data revealed modules which are significant in gene enrichment analysis. Specifically, two genes - RFC4 and ECT2 - have been found to be significant in terms of survival time. We also performed a second gene co-expression analysis on a second dataset of microarray data, and many significant genes found in this analysis could also be found in the RNA-seq data implying that these genes might indeed play a crucial role in Lung Squamous Cancer. All the R code for the analysis can be found at: https://***/nafizh/Gene_ coexpression_analysis_lung_cancer Copyright © 2018, The Authors. All rights reserved.

关键词： Genes

Rank Projection Trees for Multilevel Neural Network Interpretation

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Warrell, Jonathan Mohsen, Hussein Gerstein, Mark Program in Computational Biology and Bioinformatics Yale University Molecular Biophysics and Biochemistry Yale University Department of Computer Science Yale University

A variety of methods have been proposed for interpreting nodes in deep neural networks, which typically involve scoring nodes at lower layers with respect to their effects on the output of higher-layer nodes (where lower and higher layers are closer to the input and output layers, respectively). However, we may be interested in picking out a prioritized collection of subsets of the inputs across a range of scales according to their importance for an output node, and not simply a prioritized ranking across the inputs as singletons. Such a situation may arise in biological applications, for instance, where we are interested in epistatic effects between groups of genes in determining a trait of interest. Here, we outline a flexible framework which may be used to generate multiscale network interpretations, using any previously defined scoring function. We demonstrate the ability of our method to pick out biologically important genes and gene sets in the domains of cancer and psychiatric genomics. Copyright © 2018, The Authors. All rights reserved.

关键词： Genes

An approach to comparing tiling array and high throughput sequencing technologies for genomic transcript mapping

学校读者我要写书评

暂无评论

BMC Research Notes 2009年第1期2卷 1-7页

作者： Sasidharan, Rajkumar Agarwal, Ashish Rozowsky, Joel Gerstein, Mark Molecular Biophysics and Biochemistry Department Yale University New Haven CT 06520 United States Department of Plant Biology Carnegie Institution for Science Stanford CA 94305 United States Interdepartmental Program in Computational Biology and Bioinformatics Yale University New Haven CT 06520 United States Department of Computer Science Yale University New Haven CT 06520 United States

Background. There are two main technologies for transcriptome profiling, namely, tiling microarrays and high-throughput sequencing. Recently there has been a tremendous amount of excitement about the latter because of the advent of next-generation sequencing technologies and its promises. Consequently, the question of the moment is how these two technologies compare. Here we attempt to develop an approach to do a fair comparison of transcripts identified from tiling microarray and MPSS sequencing data. Findings. This comparison is a challenging task because the sequencing data is discrete while the tiling array data is continuous. We use the published rice and Arabidopsis datasets which provide currently best matched sets of arrays and sequencing experiments using a slightly earlier generation of sequencing, the MPSS tag sequencing technology. After scoring the arrays consistently in both the organisms, a first pass comparison reveals a surprisingly small overlap in transcripts of 22% and 66% respectively, in rice and Arabidopsis. However, when we do the analysis in detail, we find that this is an underestimate. In particular, when we map the probe intensities onto the sequencing tags and then look at their intensity distribution, we see that they are very similar to exons. Furthermore, restricting our comparison to only protein-coding gene loci revealed a very good overlap between the two technologies. Conclusion. Our approach to compare genome tiling microarray and MPSS sequencing data suggests that there is actually a reasonable overlap in transcripts identified by the two technologies. This overlap is distorted by the scoring and thresholding in the tiling array scoring procedure. © 2009 Sasidharan et al;licensee BioMed Central Ltd.

关键词： Tiling Array Transcript Mapping Tile Path Tiling Microarray Intergenic Transcript

Characterizing the tumor microenvironmental changes in a network perspective

学校读者我要写书评

暂无评论

Characterizing the tumor microenvironmental changes in a net...

第四届全国生物信息学与系统生物学学术大会

作者：李文婷王瑞孙之荣 Institute of Bioinformatics School of Life SciencesTsinghua University Computational Biology and Bioinformatics Program Institute for Genome Science and PolicyDuke University Medical CenterDurhamNC27705

Alternations in tumor microenvironment are critical in driving tumor development. However,it proves difficult to characterize their molecular components and further relate to important pathological *** using a network approach,we characterized systematic expression changes at multi-gene module levels between colorectal tumors with and without recurrence after surgery and confirmed their association with two critical

关键词： Characterizing the tumor microenvironmental changes in a network perspective

Distribution-based Sketching of Single-Cell Samples

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Baskaran, Vishal Athreya Ranek, Jolene Shan, Siyuan Stanley, Natalie Oliva, Junier B. Department of Computer Science UNC-Chapel Hill United States Computational Medicine Program Curriculum in Bioinformatics and Computational Biology UNC-Chapel Hill United States Computational Medicine Program Department of Computer Science UNC-Chapel Hill United States

Modern high-throughput single-cell immune profiling technologies, such as flow and mass cytometry and single-cell RNA sequencing can readily measure the expression of a large number of protein or gene features across the millions of cells in a multi-patient cohort. While bioinformatics approaches can be used to link immune cell heterogeneity to external variables of interest, such as, clinical outcome or experimental label, they often struggle to accommodate such a large number of profiled cells. To ease this computational burden, a limited number of cells are typically sketched or subsampled from each patient. However, existing sketching approaches fail to adequately subsample rare cells from rare cell-populations, or fail to preserve the true frequencies of particular immune cell-types. Here, we propose a novel sketching approach based on Kernel Herding that selects a limited subsample of all cells while preserving the underlying frequencies of immune cell-types. We tested our approach on three flow and mass cytometry datasets and on one single-cell RNA sequencing dataset and demonstrate that the sketched cells (1) more accurately represent the overall cellular landscape and (2) facilitate increased performance in downstream analysis tasks, such as classifying patients according to their clinical outcome. An implementation of sketching with Kernel Herding is publicly available at https://***/vishalathreya/Set-Summarization. © 2022, CC0.

关键词： bioinformatics

A pathway analysis applied to Genetic Analysis Workshop 16 genome-wide rheumatoid arthritis data

学校读者我要写书评

暂无评论

BMC proceedings 2009年第7期3 Suppl 7卷 S91页

作者： David H Ballard Chatchawit Aporntewan Ji Young Lee Joon Sang Lee Zheyang Wu Hongyu Zhao Program in Computational Biology and Bioinformatics Yale University 300 George Street Suite 503 New Haven Connecticut 06511 USA. david.ballard@yale.edu.

The identification of several hundred genomic regions affecting disease risk has proven the ability of genome-wide association studies have proven their ability to identify genetic contributors to disease. Currently, single-nucleotide polymorphism (SNP) association analysis is the most widely used method of genome-wide association data, but recent research shows that multi-marker tests of association may provide greater power, especially when more than one mutation is present within a gene and the mutations are in low linkage disequilibrium with each other. Here we use a multi-marker association test based on regression to SNPs located within known genes to obtain a gene-level score of association. We then perform pathway analysis using this score as a measure of gene importance. We use two tests of pathway enrichment - a binomial test and a random set method. By utilizing publicly available gene and pathway information, we identify B cell, cytokine and inflammation response, and antigen presentation pathways as being associated with rheumatoid arthritis. These results confirm known biological mechanisms for auto-immunity disorders, of which rheumatoid arthritis is one.

关键词： Rheumatoid Arthritis Major Histocompatibility Complex Major Histocompatibility Complex Class Binomial Test Gene Score