BACKGROUND: Machine learning and deep learning are powerful tools for analyzing electronic health records (EHRs) in healthcare research. Although family health history has been recognized as a major predictor for a wide spectrum of diseases, research has so far adopted a limited view of family relations, essentially treating patients as independent samples in the analysis. METHODS: To address this gap, we present ALIGATEHR, which models inferred family relations in a graph attention network augmented with an attention-based medical ontology representation, thus accounting for the complex influence of genetics, shared environmental exposures, and disease dependencies. RESULTS: Taking disease risk prediction as a use case, we demonstrate that explicitly modeling family relations significantly improves predictions across the disease spectrum. We then show how ALIGATEHR's attention mechanism, which links patients' disease risk to their relatives' clinical profiles, successfully captures genetic aspects of diseases using longitudinal EHR diagnosis data. Finally, we use ALIGATEHR to successfully distinguish the 2 main inflammatory bowel disease subtypes with highly shared risk factors and symptoms (Crohn's disease and ulcerative colitis). CONCLUSION: Overall, our results highlight that family relations should not be overlooked in EHR research and illustrate ALIGATEHR's great potential for enhancing patient representation learning for predictive and interpretable modeling of EHRs.

As the scale of available on-line data grows ever larger, individuals and businesses must cope with increasing complexity in decision-making processes which utilize large volumes of unstructured, semi-structured and/or structured data to satisfy multiple, interrelated information needs which contribute to an overall decision. Traditional decision support systems (DSSs) have been developed to address this need, but such systems are typically expensive to build, and are purpose-built for a particular decision-making scenario, making them difficult to extend or adapt to new decision scenarios. In this paper, we propose a novel decision representation which allows decision makers to formulate and organize natural language questions or assertions into an analytic hierarchy, which can be evaluated as part of an ad hoc decision process or as a documented, repeatable analytic process. We then introduce a new decision support framework, QUADS, which takes advantage of automatic question answering (QA) technologies to automatically understand and process a decision representation, producing a final decision by gathering and weighting answers to individual questions using a Bayesian learning and inference process. An open source framework implementation is presented and applied to two real world applications: target validation, a fundamental decision-making task for the pharmaceutical industry, and product recommendation from review texts, an everyday decision-making situation faced by on-line consumers. In both applications, we implemented and compared a number of decision synthesis algorithms, and present experimental results which demonstrate the performance of the QUADS approach versus other baseline approaches.

Next-generation sequencing technologies offer extraordinary high-throughput capacity and broad applications in biological research, drug development and molecular diagnostics. The use of targeted exome capture provides an efficient strategy sequence the complete coding regions (the exome) of the genome at much deeper read-depths without the high costs of the whole genome sequencing approaches. With the increasingly wide acceptance of high-throughput sequencing technologies in the bio-pharmaceutical industry, it becomes critical to have fast, robust, and high quality in silico methods for secondary and tertiary analyses of large amounts of exome sequencing data in an industrial environment.

With increasing numbers of genomic datasets becoming publicly available, new analysis and filtering tools are needed to serve user communities. Here we describe the development of the Roche In Silico Target Portal (ISTP) for Oncology research, a system that integrates meta-analysis results of gene expression studies with gene-disease, gene-mutation as well as gene-compound (competitor information) relationships into a single entry application. This system uses a three-tier architecture that includes an Oracle relational database, a set of data processing pipelines, and a Java-based web interface. The data processing pipelines were developed in house using Accelrys' Pipeline Pilot. Currently, the ISTP database contains data for seven major tumor types with tens of thousands of records on genes, compounds, mutations, and gene-disease relationship, accessible to Roche internal users via a simple user interface. The ISTP application highlights the potential of such a data integration system for drug discovery and development.

Identifying drug target candidates is an important task for early development throughout the drug discovery process. This process is supported by the development of new high-throughput technologies that enable better understanding of disease mechanism. With the push for personalized medicine, more experimental data are produced to identify how the genetics differ among individuals with respect to disease mechanism and drug response. It becomes critical to facilitate effective analysis of the large amount of biological data. In this paper, we describe our solution in employing text mining as a technique for finding scientific information for target and biomarker discovery from the biomedical literature. Additionally, we discuss how the extracted knowledge can be an effective resource for the analysis of biological data such as next-generation sequencing data.

MOTIVATION: Identifying drug-drug interactions (DDIs) is a critical process in drug administration and drug development. Clinical support tools often provide comprehensive lists of DDIs, but they usually lack the supporting scientific evidences and different tools can return inconsistent results. In this article, we propose a novel approach that integrates text mining and automated reasoning to derive DDIs. Through the extraction of various facts of drug metabolism, not only the DDIs that are explicitly mentioned in text can be extracted but also the potential interactions that can be inferred by reasoning. RESULTS: Our approach was able to find several potential DDIs that are not present in DrugBank. We manually evaluated these interactions based on their supporting evidences, and our analysis revealed that 81.3% of these interactions are determined to be correct. This suggests that our approach can uncover potential DDIs with scientific evidences explaining the mechanism of the interactions.

Remote sensing technologies have long since been relegated to the implementation of airborne or spaceborne configurations. We recognize that distributed sensors in all configurations can contribute to improved sensor performance and expanded measurement configurations. These new sensor systems require designs that can be scaled to hundreds to potentially thousands of individual sensors or sensor components. This paper describes a novel sensor network design strategy and some suggested operational scenarios. Mega-Mesh sensor networks require special design to support effective communication protocols. There will be a requirement for reliable sensor-to-sensor communication, ad-hoc self organizing network, multiple individual communication network structures, wireless base stations, remotely charging techniques, etc. All of these approaches dictate unique network architectures that eventually interface to conventional communication infrastructure and/or legacy networks.

Storage area networks (SAN) are important components of today's corporate networks. Two key features of SAN are the ability to provide distributed access to stored files of any size in a timely manner, and to use network bandwidth efficiently. To achieve these goals, the paper proposes a novel bandwidth reservation scheme for the LightRing - a recently proposed all-optical WDM ring with distributed multi-token access control. For each file transfer, the proposed scheme reserves a bandwidth that: (1) is proportional to the file size; and (2) takes into account the ring latency. The file is then transmitted using one or more optical bursts. As shown in the paper, the proposed reservation scheme achieves both of the goals, i.e., a good utilization of the ring bandwidth and fast file transfer time under a variety of file sizes.

A number of bandwidth reservation schemes have been proposed for multi-channel rings, generally classified as centralized and distributed. Besides simulation results, the existing analytical models address only the blocking probability and achievable throughput of such schemes. However, in determining the properties and advantages of each reservation scheme (especially the LightRing, a scheme previously proposed by the authors) the above performance figures are not sufficient, as network latency plays a determinant role when fast bandwidth provisioning is of the essence. This paper presents an analytical framework that allows one to derive the response time and throughput of such schemes under the assumption that a single request for bandwidth can be stored and processed by the source at a time. Numerical results obtained using the proposed model reveal a number of interesting properties of the existing bandwidth reservation schemes.

One of the approaches currently pursued to provide an optical network architecture for supporting the next generation Internet consists of transmitting the Internet protocol (IP) packets directly over the wavelength division multiplexing (WDM) layer. The challenge of this approach is to make the optical bandwidth directly accessible to IP with the statistical multiplexing necessary to support the bursty nature of IP traffic in a cost effective way. This paper proposes a WDM ring architecture with a multitoken interarrival time (MTIT) access protocol designed to achieve a bandwidth efficient multiplexing technique in local and metropolitan network applications. The MTIT protocol achieves efficient multiplexing of all-optically transmitted packets by means of multiple control tokens whose rotation speed along the ring is regulated via the "token interarrival time," as opposed to the "token rotation time" adopted in existing standards. The proposed WDM architecture and protocol provide a unique medium access control that allows one to maintain the channel access delay at a node within a limited range of values, possibly reaching average values that are below the roundtrip propagation time of the ring. In the paper it is shown that the MTIT protocol efficiency grows with the number of wavelengths-in agreement with today's WDM technology trends-and it is not a function of the packet size-a mandatory feature to support IP traffic.

One of the concepts currently pursued to provide the ideal network architecture for supporting the next generation Internet consists of transmitting the Internet protocol (IP) packets directly over the wavelength division multiplexing (WDM) layer of extant optical networks. The challenge of this concept is to make the optical bandwidth directly accessible to IP with the statistical multiplexing necessary to support the bursty nature of IP traffic in a cost effective way. This paper proposes the multi-token inter-arrival time (MTIT) access protocol designed to achieve a bandwidth efficient multiplexing technique in WDM ring networks. The network throughput is maximized without violating the desired access delay requirement by controlling the "token inter-arrival time" as opposed to controlling the "token rotation time" as proposed in existing standards, e.g., FDDI. It is shown that the MTIT protocol efficiency grows with the number of wavelengths-in full agreement with the WDM technology trend-and does not depend on the packet length-a feature ideal for IP implementation. In addition, the protocol provides a unique token control technique that allows to maintain the channel access delay at a node bounded even within the round trip propagation time of a ring.