Natural Language Processing in aid of FlyBase curators

Background  

Despite increasing interest in applying Natural Language Processing (NLP) to biomedical text, whether this technology can facilitate tasks such as database curation remains unclear.     

Automatic Extraction of Nanoparticle Properties Using Natural Language Processing: NanoSifter an Application to Acquire PAMAM Dendrimer Properties

In this study, we demonstrate the use of natural language processing methods to extract, from nanomedicine literature, numeric values of biomedical property terms of poly(amidoamine) dendrimers. We have developed a method for extracting these values for properties taken from the NanoParticle Ontology, using the General Architecture for Text Engineering and a Nearly-New Information Extraction System. We also created a method for associating the identified numeric values with their corresponding dendrimer properties, called NanoSifter.We demonstrate that our system can correctly extract numeric values of dendrimer properties reported in the cancer treatment literature with high recall, precision, and f-measure. The micro-averaged recall was 0.99, precision was 0.84, and f-measure was 0.91. Similarly, the macro-averaged recall was 0.99, precision was 0.87, and f-measure was 0.92. To our knowledge, these results are the first application of text mining to extract and associate dendrimer property terms and their corresponding numeric values.     

Exact reconstruction of gene regulatory networks using compressive sensing

Background

We consider the problem of reconstructing a gene regulatory network structure from limited time series gene expression data, without any a priori knowledge of connectivity. We assume that the network is sparse, meaning the connectivity among genes is much less than full connectivity. We develop a method for network reconstruction based on compressive sensing, which takes advantage of the network’s sparseness.

Results

For the case in which all genes are accessible for measurement, and there is no measurement noise, we show that our method can be used to exactly reconstruct the network. For the more general problem, in which hidden genes exist and all measurements are contaminated by noise, we show that our method leads to reliable reconstruction. In both cases, coherence of the model is used to assess the ability to reconstruct the network and to design new experiments. We demonstrate that it is possible to use the coherence distribution to guide biological experiment design effectively. By collecting a more informative dataset, the proposed method helps reduce the cost of experiments. For each problem, a set of numerical examples is presented.

Conclusions

The method provides a guarantee on how well the inferred graph structure represents the underlying system, reveals deficiencies in the data and model, and suggests experimental directions to remedy the deficiencies.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-014-0400-4) contains supplementary material, which is available to authorized users.     

Gene regulatory network reconstruction using Bayesian networks, the Dantzig Selector, the Lasso and their meta-analysis

Modern technologies and especially next generation sequencing facilities are giving a cheaper access to genotype and genomic data measured on the same sample at once. This creates an ideal situation for multifactorial experiments designed to infer gene regulatory networks. The fifth "Dialogue for Reverse Engineering Assessments and Methods" (DREAM5) challenges are aimed at assessing methods and associated algorithms devoted to the inference of biological networks. Challenge 3 on "Systems Genetics" proposed to infer causal gene regulatory networks from different genetical genomics data sets. We investigated a wide panel of methods ranging from Bayesian networks to penalised linear regressions to analyse such data, and proposed a simple yet very powerful meta-analysis, which combines these inference methods. We present results of the Challenge as well as more in-depth analysis of predicted networks in terms of structure and reliability. The developed meta-analysis was ranked first among the 16 teams participating in Challenge 3A. It paves the way for future extensions of our inference method and more accurate gene network estimates in the context of genetical genomics.     

Identification of bacterial sRNA regulatory targets using ribosome profiling

Bacteria express large numbers of non-coding, regulatory RNAs known as ‘small RNAs’ (sRNAs). sRNAs typically regulate expression of multiple target messenger RNAs (mRNAs) through base-pairing interactions. sRNA:mRNA base-pairing often results in altered mRNA stability and/or altered translation initiation. Computational identification of sRNA targets is challenging due to the requirement for only short regions of base-pairing that can accommodate mismatches. Experimental approaches have been applied to identify sRNA targets on a genomic scale, but these focus only on those targets regulated at the level of mRNA stability. Here, we utilize ribosome profiling (Ribo-seq) to experimentally identify regulatory targets of the Escherichia coli sRNA RyhB. We not only validate a majority of known RyhB targets using the Ribo-seq approach, but also discover many novel ones. We further confirm regulation of a selection of known and novel targets using targeted reporter assays. By mutating nucleotides in the mRNA of a newly discovered target, we demonstrate direct regulation of this target by RyhB. Moreover, we show that Ribo-seq distinguishes between mRNAs regulated at the level of RNA stability and those regulated at the level of translation. Thus, Ribo-seq represents a powerful approach for genome-scale identification of sRNA targets.     

The phosphate regulon and bacterial virulence: a regulatory network connecting phosphate homeostasis and pathogenesis

Bacterial pathogens regulate virulence factor gene expression coordinately in response to environmental stimuli, including nutrient starvation. The phosphate (Pho) regulon plays a key role in phosphate homeostasis. It is controlled by the PhoR/PhoB two-component regulatory system. PhoR is an integral membrane signaling histidine kinase that, through an interaction with the ABC-type phosphate-specific transport (Pst) system and a protein called PhoU, somehow senses environmental inorganic phosphate (P(i)) levels. Under conditions of P(i) limitation (or in the absence of a Pst component or PhoU), PhoR activates its partner response regulator PhoB by phosphorylation, which, in turn, up- or down-regulates target genes. Single-cell profiling of PhoB activation has shown recently that Pho regulon gene expression exhibits a stochastic, "all-or-none" behavior. Recent studies have also shown that the Pho regulon plays a role in the virulence of several bacteria. Here, we present a comprehensive overview of the role of the Pho regulon in bacterial virulence. The Pho regulon is clearly not a simple regulatory circuit for controlling phosphate homeostasis; it is part of a complex network important for both bacterial virulence and stress response.     

Hub-centered gene network reconstruction using automatic relevance determination

Network inference deals with the reconstruction of biological networks from experimental data. A variety of different reverse engineering techniques are available; they differ in the underlying assumptions and mathematical models used. One common problem for all approaches stems from the complexity of the task, due to the combinatorial explosion of different network topologies for increasing network size. To handle this problem, constraints are frequently used, for example on the node degree, number of edges, or constraints on regulation functions between network components. We propose to exploit topological considerations in the inference of gene regulatory networks. Such systems are often controlled by a small number of hub genes, while most other genes have only limited influence on the network's dynamic. We model gene regulation using a Bayesian network with discrete, Boolean nodes. A hierarchical prior is employed to identify hub genes. The first layer of the prior is used to regularize weights on edges emanating from one specific node. A second prior on hyperparameters controls the magnitude of the former regularization for different nodes. The net effect is that central nodes tend to form in reconstructed networks. Network reconstruction is then performed by maximization of or sampling from the posterior distribution. We evaluate our approach on simulated and real experimental data, indicating that we can reconstruct main regulatory interactions from the data. We furthermore compare our approach to other state-of-the art methods, showing superior performance in identifying hubs. Using a large publicly available dataset of over 800 cell cycle regulated genes, we are able to identify several main hub genes. Our method may thus provide a valuable tool to identify interesting candidate genes for further study. Furthermore, the approach presented may stimulate further developments in regularization methods for network reconstruction from data.     

A Natural Language Processing Tool for Large-Scale Data Extraction from Echocardiography Reports

Large volumes of data are continuously generated from clinical notes and diagnostic studies catalogued in electronic health records (EHRs). Echocardiography is one of the most commonly ordered diagnostic tests in cardiology. This study sought to explore the feasibility and reliability of using natural language processing (NLP) for large-scale and targeted extraction of multiple data elements from echocardiography reports. An NLP tool, EchoInfer, was developed to automatically extract data pertaining to cardiovascular structure and function from heterogeneously formatted echocardiographic data sources. EchoInfer was applied to echocardiography reports (2004 to 2013) available from 3 different on-going clinical research projects. EchoInfer analyzed 15,116 echocardiography reports from 1684 patients, and extracted 59 quantitative and 21 qualitative data elements per report. EchoInfer achieved a precision of 94.06%, a recall of 92.21%, and an F1-score of 93.12% across all 80 data elements in 50 reports. Physician review of 400 reports demonstrated that EchoInfer achieved a recall of 92–99.9% and a precision of >97% in four data elements, including three quantitative and one qualitative data element. Failure of EchoInfer to correctly identify or reject reported parameters was primarily related to non-standardized reporting of echocardiography data. EchoInfer provides a powerful and reliable NLP-based approach for the large-scale, targeted extraction of information from heterogeneous data sources. The use of EchoInfer may have implications for the clinical management and research analysis of patients undergoing echocardiographic evaluation.     

Automatic diagnosis of imbalanced ophthalmic images using a cost-sensitive deep convolutional neural network

Background

Ocular images play an essential role in ophthalmological diagnoses. Having an imbalanced dataset is an inevitable issue in automated ocular diseases diagnosis; the scarcity of positive samples always tends to result in the misdiagnosis of severe patients during the classification task. Exploring an effective computer-aided diagnostic method to deal with imbalanced ophthalmological dataset is crucial.

Methods

In this paper, we develop an effective cost-sensitive deep residual convolutional neural network (CS-ResCNN) classifier to diagnose ophthalmic diseases using retro-illumination images. First, the regions of interest (crystalline lens) are automatically identified via twice-applied Canny detection and Hough transformation. Then, the localized zones are fed into the CS-ResCNN to extract high-level features for subsequent use in automatic diagnosis. Second, the impacts of cost factors on the CS-ResCNN are further analyzed using a grid-search procedure to verify that our proposed system is robust and efficient.

Results

Qualitative analyses and quantitative experimental results demonstrate that our proposed method outperforms other conventional approaches and offers exceptional mean accuracy (92.24%), specificity (93.19%), sensitivity (89.66%) and AUC (97.11%) results. Moreover, the sensitivity of the CS-ResCNN is enhanced by over 13.6% compared to the native CNN method.

Conclusion

Our study provides a practical strategy for addressing imbalanced ophthalmological datasets and has the potential to be applied to other medical images. The developed and deployed CS-ResCNN could serve as computer-aided diagnosis software for ophthalmologists in clinical application.

     

Analyzing stationary states of gene regulatory network using petri nets

We introduce and formally define the notion of a stationary state for Petri nets. We also propose a fully automatic method for finding such states. The procedure makes use of the Presburger arithmetic to describe all the stationary states. Finally we apply this novel approach to find stationary states of a gene regulatory network describing the flower morphogenesis of A. thaliana. This shows that the proposed method can be successfully applied in the study of biological systems.     

Dissecting microregulation of a master regulatory network

Background

All jawed-vertebrates have four T cell receptor (TCR) chains: alpha (TRA), beta (TRB), gamma (TRG) and delta (TRD). Marsupials appear unique by having an additional TCR: mu (TRM). The evolutionary origin of TRM and its relationship to other TCR remain obscure, and is confounded by previous results that support TRM being a hybrid between a TCR and immunoglobulin locus. The availability of the first marsupial genome sequence allows investigation of these evolutionary relationships.

Results

The organization of the conventional TCR loci, encoding the TRA, TRB, TRG and TRD chains, in the opossumMonodelphis domesticaare highly conserved with and of similar complexity to that of eutherians (placental mammals). There is a high degree of conserved synteny in the genomic regions encoding the conventional TCR across mammals and birds. In contrast the chromosomal region containing TRM is not well conserved across mammals. None of the conventional TCR loci contain variable region gene segments with homology to those found in TRM; rather TRM variable genes are most similar to that of immunoglobulin heavy chain genes.

Conclusion

Complete genomic analyses of the opossum TCR loci continue to support an origin of TRM as a hybrid between a TCR and immunoglobulin locus. None of the conventional TCR loci contain evidence that such a recombination event occurred, rather they demonstrate a high degree of stability across distantly related mammals. TRM, therefore, appears to be derived from receptor genes no longer extant in placental mammals. These analyses provide the first genomic scale structural detail of marsupial TCR genes, a lineage of mammals used as models of early development and human disease.     

Integrated network reconstruction,visualization and analysis using YANAsquare

Background  

Modeling of metabolic networks includes tasks such as network assembly, network overview, calculation of metabolic fluxes and testing the robustness of the network.     

Mapping and manipulating the Mycobacterium tuberculosis transcriptome using a transcription factor overexpression-derived regulatory network

Background

Mycobacterium tuberculosis is characterized by a low mutation rate and a lack of genetic recombination. Yet, the rise of extensively resistant strains paints a picture of a microbe with an impressive adaptive potential. Here we describe the first documented case of extensively drug-resistant tuberculosis evolved from a susceptible ancestor within a single patient.

Results

Genome sequences of nine serial M. tuberculosis isolates from the same patient uncovered a dramatic turnover of competing lineages driven by the emergence, and subsequent fixation or loss of single nucleotide polymorphisms. For most drugs, resistance arose through independent emergence of mutations in more than one clone, of which only one ultimately prevailed as the clone carrying it expanded, displacing the other clones in the process. The vast majority of mutations identified over 3.5 years were either involved in drug resistance or hitchhiking in the genetic background of these. Additionally, RNA-sequencing of isolates grown in the absence of drug challenge revealed that the efflux-associated iniBAC operon was up-regulated over time, whereas down-regulated genes include those involved in mycolic acid synthesis.

Conclusions

We observed both rapid acquisitions of resistance to antimicrobial compounds mediated by individual mutations as well as a gradual increase in fitness in the presence of antibiotics, likely driven by stable gene expression reprogramming. The rapid turnover of resistance mutations and hitchhiking neutral mutations has major implications for inferring tuberculosis transmission events in situations where drug resistance evolves within transmission chains.     

Efficient reverse-engineering of a developmental gene regulatory network

Understanding the complex regulatory networks underlying development and evolution of multi-cellular organisms is a major problem in biology. Computational models can be used as tools to extract the regulatory structure and dynamics of such networks from gene expression data. This approach is called reverse engineering. It has been successfully applied to many gene networks in various biological systems. However, to reconstitute the structure and non-linear dynamics of a developmental gene network in its spatial context remains a considerable challenge. Here, we address this challenge using a case study: the gap gene network involved in segment determination during early development of Drosophila melanogaster. A major problem for reverse-engineering pattern-forming networks is the significant amount of time and effort required to acquire and quantify spatial gene expression data. We have developed a simplified data processing pipeline that considerably increases the throughput of the method, but results in data of reduced accuracy compared to those previously used for gap gene network inference. We demonstrate that we can infer the correct network structure using our reduced data set, and investigate minimal data requirements for successful reverse engineering. Our results show that timing and position of expression domain boundaries are the crucial features for determining regulatory network structure from data, while it is less important to precisely measure expression levels. Based on this, we define minimal data requirements for gap gene network inference. Our results demonstrate the feasibility of reverse-engineering with much reduced experimental effort. This enables more widespread use of the method in different developmental contexts and organisms. Such systematic application of data-driven models to real-world networks has enormous potential. Only the quantitative investigation of a large number of developmental gene regulatory networks will allow us to discover whether there are rules or regularities governing development and evolution of complex multi-cellular organisms.