Automatic infection detection based on electronic medical records

Background

Making accurate patient care decision, as early as possible, is a constant challenge, especially for physicians in the emergency department. The increasing volumes of electronic medical records (EMRs) open new horizons for automatic diagnosis. In this paper, we propose to use machine learning approaches for automatic infection detection based on EMRs. Five categories of information are utilized for prediction, including personal information, admission note, vital signs, diagnose test results and medical image diagnose.

Results

Experimental results on a newly constructed EMRs dataset from emergency department show that machine learning models can achieve a decent performance for infection detection with area under the receiver operator characteristic curve (AUC) of 0.88. Out of all the five types of information, admission note in text form makes the most contribution with the AUC of 0.87.

Conclusions

This study provides a state-of-the-art EMRs processing system to automatically make medical decisions. It extracts five types of features associated with infection and achieves a decent performance on automatic infection detection based on machine learning models.

     

OntoGene in BioCreative II

Background:

Research scientists and companies working in the domains of biomedicine and genomics are increasingly faced with the problem of efficiently locating, within the vast body of published scientific findings, the critical pieces of information that are needed to direct current and future research investment.

Results:

In this report we describe approaches taken within the scope of the second BioCreative competition in order to solve two aspects of this problem: detection of novel protein interactions reported in scientific articles, and detection of the experimental method that was used to confirm the interaction. Our approach to the former problem is based on a high-recall protein annotation step, followed by two strict disambiguation steps. The remaining proteins are then combined according to a number of lexico-syntactic filters, which deliver high-precision results while maintaining reasonable recall. The detection of the experimental methods is tackled by a pattern matching approach, which has delivered the best results in the official BioCreative evaluation.

Conclusion:

Although the results of BioCreative clearly show that no tool is sufficiently reliable for fully automated annotations, a few of the proposed approaches (including our own) already perform at a competitive level. This makes them interesting either as standalone tools for preliminary document inspection, or as modules within an environment aimed at supporting the process of curation of biomedical literature.

     

PKQuest: volatile solutes – application to enflurane, nitrous oxide, halothane, methoxyflurane and toluene pharmacokinetics

Background  

The application of physiologically based pharmacokinetic models (PBPK) to human studies has been limited by the lack of the detailed organ information that is required for this analysis. PKQuest is a new generic PBPK that is designed to avoid this problem by using a set of "standard human" default parameters that are applicable to most solutes.     

An improved sparse representation model with structural information for Multicolour Fluorescence <Emphasis Type="Italic">In-Situ</Emphasis> Hybridization (M-FISH) image classification

Background

Multicolour Fluorescence In-Situ Hybridization (M-FISH) images are employed for detecting chromosomal abnormalities such as chromosomal translocations, deletions, duplication and inversions. This technique uses mixed colours of fluorochromes to paint the whole chromosomes for rapid detection of chromosome rearrangements. The M-FISH data sets used in our research are obtained from microscopic scanning of a metaphase cell labelled with five different fluorochromes and a DAPI staining. The reliability of the technique lies in accurate classification of chromosomes (24 classes for male and 23 classes for female) from M-FISH images. However, due to imaging noise, mis-alignment between multiple channels and many other imaging problems, there is always a classification error, leading to wrong detection of chromosomal abnormalities. Therefore, how to accurately classify different types of chromosomes from M-FISH images becomes a challenging problem.

Methods

This paper presents a novel sparse representation model considering structural information for the classification of M-FISH images. In our previous work a sparse representation based classification model was proposed. This model employed only individual pixel information for the classification. With the structural information of neighbouring pixels as well as the information of themselves simultaneously, the novel approach extended the previous one to the regional case. Based on Orthogonal Matching Pursuit (OMP), we developed simultaneous OMP algorithm (SOMP) to derive an efficient solution of the improved sparse representation model by incorporating the structural information.

Results

The p-value of two models shows that the newly proposed model incorporating the structural information is significantly superior to our previous one. In addition, we evaluated the effect of several parameters, such as sparsity level, neighbourhood size, and training sample size, on the of the classification accuracy.

Conclusions

The comparison with our previously used sparse model demonstrates that the improved sparse representation model is more effective than the previous one on the classification of the chromosome abnormalities.

     

Estimating mutual information using B-spline functions – an improved similarity measure for analysing gene expression data

Background

The information theoretic concept of mutual information provides a general framework to evaluate dependencies between variables. In the context of the clustering of genes with similar patterns of expression it has been suggested as a general quantity of similarity to extend commonly used linear measures. Since mutual information is defined in terms of discrete variables, its application to continuous data requires the use of binning procedures, which can lead to significant numerical errors for datasets of small or moderate size.

Results

In this work, we propose a method for the numerical estimation of mutual information from continuous data. We investigate the characteristic properties arising from the application of our algorithm and show that our approach outperforms commonly used algorithms: The significance, as a measure of the power of distinction from random correlation, is significantly increased. This concept is subsequently illustrated on two large-scale gene expression datasets and the results are compared to those obtained using other similarity measures.A C++ source code of our algorithm is available for non-commercial use from kloska@scienion.de upon request.

Conclusion

The utilisation of mutual information as similarity measure enables the detection of non-linear correlations in gene expression datasets. Frequently applied linear correlation measures, which are often used on an ad-hoc basis without further justification, are thereby extended.

     

Detection of distant evolutionary relationships between protein families using theory of sequence profile-profile comparison

Background  

Detection of common evolutionary origin (homology) is a primary means of inferring protein structure and function. At present, comparison of protein families represented as sequence profiles is arguably the most effective homology detection strategy. However, finding the best way to represent evolutionary information of a protein sequence family in the profile, to compare profiles and to estimate the biological significance of such comparisons, remains an active area of research.     

Mining for diagnostic information in body surface potential maps: A comparison of feature selection techniques

Background  

In body surface potential mapping, increased spatial sampling is used to allow more accurate detection of a cardiac abnormality. Although diagnostically superior to more conventional electrocardiographic techniques, the perceived complexity of the Body Surface Potential Map (BSPM) acquisition process has prohibited its acceptance in clinical practice. For this reason there is an interest in striking a compromise between the minimum number of electrocardiographic recording sites required to sample the maximum electrocardiographic information.     

Resolving deconvolution ambiguity in gene alternative splicing

Background  

For many gene structures it is impossible to resolve intensity data uniquely to establish abundances of splice variants. This was empirically noted by Wang et al. in which it was called a "degeneracy problem". The ambiguity results from an ill-posed problem where additional information is needed in order to obtain an unique answer in splice variant deconvolution.     

Nucleic acid chaperons: a theory of an RNA-assisted protein folding

Background  

Proteins are assumed to contain all the information necessary for unambiguous folding (Anfinsen's principle). However, ab initio structure prediction is often not successful because the amino acid sequence itself is not sufficient to guide between endless folding possibilities. It seems to be a logical to try to find the "missing" information in nucleic acids, in the redundant codon base.     

Requirements and design of the PROSPER protocol for implementation of information infrastructures supporting pandemic response: a Nominal Group study

Background

Advanced technical systems and analytic methods promise to provide policy makers with information to help them recognize the consequences of alternative courses of action during pandemics. Evaluations still show that response programs are insufficiently supported by information systems. This paper sets out to derive a protocol for implementation of integrated information infrastructures supporting regional and local pandemic response programs at the stage(s) when the outbreak no longer can be contained at its source.

Methods

Nominal group methods for reaching consensus on complex problems were used to transform requirements data obtained from international experts into an implementation protocol. The analysis was performed in a cyclical process in which the experts first individually provided input to working documents and then discussed them in conferences calls. Argument-based representation in design patterns was used to define the protocol at technical, system, and pandemic evidence levels.

Results

The Protocol for a Standardized information infrastructure for Pandemic and Emerging infectious disease Response (PROSPER) outlines the implementation of information infrastructure aligned with pandemic response programs. The protocol covers analyses of the community at risk, the response processes, and response impacts. For each of these, the protocol outlines the implementation of a supporting information infrastructure in hierarchical patterns ranging from technical components and system functions to pandemic evidence production.

Conclusions

The PROSPER protocol provides guidelines for implementation of an information infrastructure for pandemic response programs both in settings where sophisticated health information systems already are used and in developing communities where there is limited access to financial and technical resources. The protocol is based on a generic health service model and its functions are adjusted for community-level analyses of outbreak detection and progress, and response program effectiveness. Scientifically grounded reporting principles need to be established for interpretation of information derived from outbreak detection algorithms and predictive modeling.     

A multi-network clustering method for detecting protein complexes from multiple heterogeneous networks

Background

The accurate identification of protein complexes is important for the understanding of cellular organization. Up to now, computational methods for protein complex detection are mostly focus on mining clusters from protein-protein interaction (PPI) networks. However, PPI data collected by high-throughput experimental techniques are known to be quite noisy. It is hard to achieve reliable prediction results by simply applying computational methods on PPI data. Behind protein interactions, there are protein domains that interact with each other. Therefore, based on domain-protein associations, the joint analysis of PPIs and domain-domain interactions (DDI) has the potential to obtain better performance in protein complex detection. As traditional computational methods are designed to detect protein complexes from a single PPI network, it is necessary to design a new algorithm that could effectively utilize the information inherent in multiple heterogeneous networks.

Results

In this paper, we introduce a novel multi-network clustering algorithm to detect protein complexes from multiple heterogeneous networks. Unlike existing protein complex identification algorithms that focus on the analysis of a single PPI network, our model can jointly exploit the information inherent in PPI and DDI data to achieve more reliable prediction results. Extensive experiment results on real-world data sets demonstrate that our method can predict protein complexes more accurately than other state-of-the-art protein complex identification algorithms.

Conclusions

In this work, we demonstrate that the joint analysis of PPI network and DDI network can help to improve the accuracy of protein complex detection.

     

HAb18G/CD147 cell-cell contacts confer resistance of a HEK293 subpopulation to anoikis in an E-cadherin-dependent manner

Background  

Acquisition of resistance to "anoikis" facilitates the survival of cells under independent matrix-deficient conditions, such as cells in tumor progression and the production of suspension culture cells for biomedical engineering. There is evidence suggesting that CD147, an adhesion molecule associated with survival of cells in tumor metastasis and cell-cell contacts, plays an important role in resistance to anoikis. However, information regarding the functions of CD147 in mediating cell-cell contacts and anoikis-resistance remains limited and even self-contradictory.     

Usability evaluation of a clinical decision support tool for osteoporosis disease management

Background

There is an extensive body of literature on health system quality reporting that has yet to be characterized. Scoping is a novel methodology for systematically assessing the breadth of a body of literature in a particular research area. Our objectives were to showcase the scoping review methodology in the review of health system quality reporting, and to report on the extent of the literature in this area.

Methods

A scoping review was performed based on the York methodology outlined by Arksey and O'Malley from the University of York, United Kingdom. We searched 14 peer reviewed and grey literature databases limiting the search to English language and non-English language articles with English abstracts published between 1980 and June 2006 with an update to November 2008. We also searched specific websites, reference lists, and key journals for relevant material and solicited input from key stakeholders. Inclusion/exclusion criteria were applied to select relevant material and qualitative information was charted from the selected literature.

Results

A total of 10,102 articles were identified from searching the literature databases, 821 were deemed relevant to our scoping review. An additional 401 were identified from updates, website searching, references lists, key journals, and stakeholder suggestions for a total of 1,222 included articles. These were categorized and catalogued according to the inclusion criteria, and further subcategories were identified through the charting process. Topic areas represented by this review included the effectiveness of health system report cards (n = 194 articles), methodological issues in their development (n = 815 articles), stakeholder views on report cards (n = 144 articles), and ethical considerations around their development (n = 69 articles).

Conclusions

The scoping review methodology has permitted us to characterize and catalogue the extensive body of literature pertaining to health system report cards. The resulting literature repository that our review has created can be of use to researchers and health system stakeholders interested in the topic of health system quality measurement and reporting.     

Quantitative model for inferring dynamic regulation of the tumour suppressor gene p53

Background  

The availability of various "omics" datasets creates a prospect of performing the study of genome-wide genetic regulatory networks. However, one of the major challenges of using mathematical models to infer genetic regulation from microarray datasets is the lack of information for protein concentrations and activities. Most of the previous researches were based on an assumption that the mRNA levels of a gene are consistent with its protein activities, though it is not always the case. Therefore, a more sophisticated modelling framework together with the corresponding inference methods is needed to accurately estimate genetic regulation from "omics" datasets.     

Identifying differential correlation in gene/pathway combinations

Background  

An important emerging trend in the analysis of microarray data is to incorporate known pathway information a priori. Expression level "summaries" for pathways, obtained from the expression data for the genes constituting the pathway, permit the inclusion of pathway information, reduce the high dimensionality of microarray data, and have the power to elucidate gene-interaction dependencies which are not already accounted for through known pathway identification.     

Identification of pathogens in culture-negative infective endocarditis cases by metagenomic analysis

Background

Pathogens identification is critical for the proper diagnosis and precise treatment of infective endocarditis (IE). Although blood and valve cultures are the gold standard for IE pathogens detection, many cases are culture-negative, especially in patients who had received long-term antibiotic treatment, and precise diagnosis has therefore become a major challenge in the clinic. Metagenomic sequencing can provide both information on the pathogenic strain and the antibiotic susceptibility profile of patient samples without culturing, offering a powerful method to deal with culture-negative cases.

Methods

To assess the feasibility of a metagenomic approach to detect the causative pathogens in resected valves from IE patients, we employed both next-generation sequencing and Oxford Nanopore Technologies MinION nanopore sequencing for pathogens and antimicrobial resistance detection in seven culture-negative IE patients. Using our in-house developed bioinformatics pipeline, we analyzed the sequencing results generated from both platforms for the direct identification of pathogens from the resected valves of seven clinically culture-negative IE patients according to the modified Duke criteria.

Results

Our results showed both metagenomics methods can be applied for the causative pathogen detection in all IE samples. Moreover, we were able to simultaneously characterize respective antimicrobial resistance features.

Conclusion

Metagenomic methods for IE detection can provide clinicians with valuable information to diagnose and treat IE patients after valve replacement surgery. However, more efforts should be made to optimize protocols for sample processing, sequencing and bioinformatics analysis.

     

Information about ADRs explored by pharmacovigilance approaches: a qualitative review of studies on antibiotics, SSRIs and NSAIDs

Background  

Despite surveillance efforts, unexpected and serious adverse drug reactions (ADRs) repeatedly occur after marketing. The aim of this article is to analyse ADRs reported by available ADR signal detection approaches and to explore which information about new and unexpected ADRs these approaches have detected.     

To See or Not to See: Do Front of Pack Nutrition Labels Affect Attention to Overall Nutrition Information?

Background

Front of pack (FOP) nutrition labels are concise labels located on the front of food packages that provide truncated nutrition information. These labels are rapidly gaining prominence worldwide, presumably because they attract attention and their simplified formats enable rapid comparisons of nutritional value.

Methods

Eye tracking was conducted as US consumers interacted with actual packages with and without FOP labels to (1) assess if the presence of an FOP label increases attention to nutrition information when viewers are not specifically tasked with nutrition-related goals; and (2) study the effect of FOP presence on consumer use of more comprehensive, traditional nutrition information presented in the Nutritional Facts Panel (NFP), a mandatory label for most packaged foods in the US.

Results

Our results indicate that colored FOP labels enhanced the probability that any nutrition information was attended, and resulted in faster detection and longer viewing of nutrition information. However, for cereal packages, these benefits were at the expense of attention to the more comprehensive NFP. Our results are consistent with a potential short cut effect of FOP labels, such that if an FOP was present, participants spent less time attending the more comprehensive NFP. For crackers, FOP labels increased time spent attending to nutrition information, but we found no evidence that their presence reduced the time spent on the nutrition information in the NFP.

Conclusions

The finding that FOP labels increased attention to overall nutrition information by people who did not have an explicit nutritional goal suggests that these labels may have an advantage in conveying nutrition information to a wide segment of the population. However, for some food types this benefit may come with a short-cut effect; that is, decreased attention to more comprehensive nutrition information. These results have implications for policy and warrant further research into the mechanisms by which FOP labels impact use of nutrition information by consumers for different foods.     

GHOST: global hepatitis outbreak and surveillance technology

Background

Hepatitis C is a major public health problem in the United States and worldwide. Outbreaks of hepatitis C virus (HCV) infections associated with unsafe injection practices, drug diversion, and other exposures to blood are difficult to detect and investigate. Effective HCV outbreak investigation requires comprehensive surveillance and robust case investigation. We previously developed and validated a methodology for the rapid and cost-effective identification of HCV transmission clusters. Global Hepatitis Outbreak and Surveillance Technology (GHOST) is a cloud-based system enabling users, regardless of computational expertise, to analyze and visualize transmission clusters in an independent, accurate and reproducible way.

Results

We present and explore performance of several GHOST implemented algorithms using next-generation sequencing data experimentally obtained from hypervariable region 1 of genetically related and unrelated HCV strains. GHOST processes data from an entire MiSeq run in approximately 3 h. A panel of seven specimens was used for preparation of six repeats of MiSeq libraries. Testing sequence data from these libraries by GHOST showed a consistent transmission linkage detection, testifying to high reproducibility of the system. Lack of linkage among genetically unrelated HCV strains and constant detection of genetic linkage between HCV strains from known transmission pairs and from follow-up specimens at different levels of MiSeq-read sampling indicate high specificity and sensitivity of GHOST in accurate detection of HCV transmission.

Conclusions

GHOST enables automatic extraction of timely and relevant public health information suitable for guiding effective intervention measures. It is designed as a virtual diagnostic system intended for use in molecular surveillance and outbreak investigations rather than in research. The system produces accurate and reproducible information on HCV transmission clusters for all users, irrespective of their level of bioinformatics expertise. Improvement in molecular detection capacity will contribute to increasing the rate of transmission detection, thus providing opportunity for rapid, accurate and effective response to outbreaks of hepatitis C. Although GHOST was originally developed for hepatitis C surveillance, its modular structure is readily applicable to other infectious diseases. Worldwide availability of GHOST for the detection of HCV transmissions will foster deeper involvement of public health researchers and practitioners in hepatitis C outbreak investigation.