We are interested in modeling knowledge that describes cell-signaling pathways. Our ultimate goal is to provide tools and support for biology researchers who need to understand the complexities of large, tightly connected and interacting cell-signaling pathways. One example of such a support tool is Chalkboard, a system for reasoning over pathway knowledge, and for exploring downstream implications of pathway linkages.

As a domain area, we focus on the pathways that may be involved in the pathogenesis of Alzheimer's disease (AD). As an example of interacting pathways, we have found some interesting examples between these AD pathways and pathways involved in Diabetes Mellitus.

Our methods emphasize the contruction of robust ontologies for representing knowledge. Like the Open Biomedical Ontologies (OBO) efforts and the foundational model of anatomy, we follow the approach of developing reference ontologies that can support multiple applications. See also John Gennari's research interest page. This work in collaboration with Drs. Jesse Wiley and Dan Cook.

We are studying the effectiveness of drug alerting in the CPOE system at the VA Puget Sound. When clinicians enter orders, a rule-based program checks the order for any potential problems with patient allergies or interactions with current medications. Our research has shown that there is a very high rate of clinicians continuing with orders despite these warnings -- an alert override rate of 75-90 percent.

We are interested in this result as a motivating springboard to study the overall process of ordering, and how this process is changed and could be (should be) improved by the use of technology. We approach this problem broadly, using ethnographic tools and methods. More specifically, we plan to comprehensively study the ordering environment from a Cognitive Work Analysis (CWA) framework. Ultimately, our this analysis should directly impact the design of systems aimed to improve the physician ordering process. This project in collaboration with Dr. Tom Payne and Ching-Ping Lin.

Radiation treatment planning optimization has recently received much attention due to the advent of inverse planning techniques for Intensity Modulated Radiation Therapy. As useful as these algorithms are, they all have difficulty in handling the predominant problems in radiation therapy. These problems include decisions based on models formulated with incomplete data, incomplete and qualitative presriptions, and mutually contradictory constraints/objectives. Our project is aimed at using belief nets (also known as Bayes' nets) to provide better methods at guiding the optimization process and choosing the most clinically appropriate solution.

This project is led by Mark Phillips, Ph.D., Radiation Oncology.

Our current research is focused on using genome-wide laboratory and bioinformatics-based approaches to study regulation of gene expression in protozoan parasites. This work has been greatly enhanced by the recent completion of the trypanosomatid genome sequence, in which we played a leadership role. One NIH-funded sub-project seeks to characterize the molecular machinery involved in the unique polycistronic transcription of protein-coding genes in trypanosomatids. This involves both wet lab (molecular genetics, microarray, proteomics, EMSA, ChIP) and computational (sequence analysis, motif identification, microarray analysis) approaches.

A second sub-project uses DNA microarrays to identify changes in gene expression during differentiation of Leishmania between the insect and mammalian stages. As part of the Trypanosomatid program at SBRI, Dr. Myler is also actively involved in the discovery of new drugs against trypanosomes and Leishmania, as well as development of improved diagnostics for infection with these parasites.

This project aims to identify the challenges of representing and reasoning with drug mechanism knowledge and to evaluate potential informatics solutions to these challenges throughthe process of developing a knowledge-based system capable of predicting clinically relevant DDIs that occur via metabolic mechanisms. In previous work, we designed a simple, rule-based, model of metabolic inhibition and induction and applied it to a database containing assertions about 267 drugs.

This pilot system taught us that drug mechanism knowledge is often dynamic, missing, or uncertain. We are developing methods to address these properties of mechanism knowledge. In particular we are investigating the use of a Truth Maintenance System to link changes in the evidence support for assertions about drug properties to the set of drug-drug interactions and noninteractions the system predicts. This work in collaboration with Richard Boyce.

In this project, we pursue ontology-based informatics research that focuses on the representation and use of process knowledge within biomedical applications. Our primary driving use case is with bio-simulation systems. We collaborate closely with bioengineers who build simulation models of biology ranging from blood flow to pathway-level molecular processes.

A primary work product has been the Ontology of Physics for Biology, as published in the BioPortal collection of biological ontologies. See our project home page for more information, including our project's publications to date.

John Gennari and Daniel Cook co-lead this project.

Telemakus: Mining and Mapping Research Findings to Promote Knowledge Discovery

Specialization has resulted in much of the recent advancement in science. However, an unfortunate consequence of specialization has been poor communication across research domains, which hampers the knowledge discovery process. The Telemakus system offers an opportunity to re-envision the way we represent, retrieve and assimilate research findings from the published literature. The long-term goal is to develop a system that will provide tools and concept mapping algorithms to be applied to any domain which presents research findings as numeric data in support of the knowledge discovery process. Formalizing the representation methods and results of scientific research offers a new approach to conceptualize and dynamically aggregate research findings within and across scientific domains with the potential to ultimately improve and, indeed, speed up the scientific discovery process.

The first Telemakus knowledge resources provide active knowledge exploration maps and resources on: 1) the basic biology of caloric restriction in aging and 2) biomarkers in Alzheimers disease and mild cognitive impairment. Knowledge bases in communicable and infectious diseases (initial focus on influenza) are currently under development. The knowledge bases are freely available and can be found at the Telemakus web site.

Our goal is to perform realistic molecular modeling studies relating to protein stability, function, and folding. Protein folding is one of the fundamental unsolved problems in molecular biology. A protein must assume a stable and precisely ordered conformation to perform its biological function properly. Although much is known of the structural details of the native folded conformation of proteins, very little is known about the actual folding process.

An understanding of protein folding has important implications for all biological processes, including protein degradation, protein translocation, aging, and human diseases, including cancer and amyloid diseases. The solution to the protein folding problem also has applications in the human genome project and biotechnology. Given that protein folding is of such widespread importance to human health and the fact that experimental approaches only provide limited amounts of information on the structural transitions and interactions occurring during protein folding, we are using computer simulation methods in an attempt to delineate the important forces acting during this process.

Center for Public Health Informatics

A Robert Wood Johnson Foundation funded "Common Ground" project to investigate the business processes and workflow of local health departments in the area of chronic disease prevention and control. In conjunction with Kitsap County Health District, this project will

1. identify a common set of processes for public health chronic disease activities;
2. provide a common framework for identifying possible inefficiencies in current health department practices and
3. define a set of requirements for development of a chronic disease information system.

The long-term objective of this research is to promote the use of published information by facilitating fast, easy, and effective access to the scientific literature. To achieve this objective, we are using existing knowledge bases to incorporate text-mining and categorization techniques into search and document-management interfaces.

The goal is to investigate and develop intelligent interfaces that would allow users to see a summary of the information in their entire set of search results, explore the topics that they find most interesting, share that information with others, and use the same contextualized interface to manage the documents of interest. The initial focus is on people who use the biomedical literature to perform information-intensive tasks, such as writing review articles, planning a research path, or simply keeping up with an area of interest.

Genetic Data Integration Research Group

iMed

As clinicians are forced to decrease time spent with patients and the specialization and fragmentation of care increases, patients are required to play an increasingly prominent role in their health care. Yet, few information tools exist to support patients in this active role. In particular, patients often must coordinate their health care across multiple clinicians, learn new health terminology, make treatment choices, manage their home care, track insurance benefits, etc.

At the same time, patients are trying to maintain their normal professional and personal lives, but the intense information management demands placed on them can interfere with all those activities. Little is known about this information management work of patients, but such knowledge must be a first step towards developing the tools that patients need to support their active role. The long-term objective of this research is both to understand patients’| information management work and to develop new technology that will support that work. Specifically, we propose to (1) develop a model of patients’| personal health information management work, (2) develop new technology that supports patients in that work, and (3) evaluate the effectiveness of our new technology in helping patients manage their personal health information, participate in their own health care, and maintain their daily life activities.

Structural Informatics

The goal of this research project is to develop a unified methodology for organization and retrieval of biological data from scientific experiments. This methodology will permit biomedical researchers to organize and access their critical data in ways that permit them to perform major analyses which were previously very difficult (due to the sheer volume of data), and which will provide new insights into their work.

Our project builds on existing work in experiment management, approximate queries, and content-based image retrieval. A probabilistic query framework for multimedia data is being developed that provides users with a unified way to access various types of data. Coming from multiple sources, this data is diverse and heterogeneous, including time series signals such as neuron firing patterns, 2D images such as X-rays or photographs, 3D image volumes such as CT and MRI scans, and 4-dimensional data such as fMRI image volumes over time. These different types of data will be loosely linked together to provide a unified view, and will be organized in a way that is both easy for users to understand as well as efficient for query access. The new probabilistic query framework and data organization methodologies will permit queries that are able to handle these multiple types of related data, both alphanumeric and signal/image data, in a uniform way. For example, the queries will be able to handle both single data types and multiple related data types, such as registered CT and MRI scans or neuronal firing patterns and related fMRI data. A prototype system is being built and evaluated on three different applications: a study of language sites in the human brain, an analysis of the relationship of cataract formation to genetic factors, and a study of craniofacial disorders in children.

Symbolic representations for anatomy, as exemplified by the Foundational Model of Anatomy, are combined with spatial or image-based representation in a distributed information system. Various structural information servers access the spatial and symbolic resources, and various authoring and end-user client programs implement several applications. Among these applications are a foundational model explorer (FME), a graphical query engine to the FMA called Emily, a StruQL-based query engine for the FMA called OQAFMA, a natural language front-end that composes StruQL queries, a Lisp-based FMA server, a web-based image manager, FMA-based image retrieval, interactive atlases of anatomy, and a dynamic 3-D scene generator.

Sub-projects include: BodyGen. Digital Anatomist Atlases, Digital Anatomist Jigsaw Puzzle, Dynamic Scene Generator, Image Manager WIRM Repository, Skandha4, Who Wants to Be A (Digital) Anatomist?, WIRM, and Biolucida. See web pages for more details.

Sub-projects include: Emily, Foundational Model Explorer, Foundational Model of Anatomy DB, GAPP Server, NOQAFMA, OQAFMA DB, OQAFMA Server, and WebGAPP.

An on-line information system for managing and visualizing data about the human brain. In this case a structural model is used as a framework for organizing other information.

Related subprojects: Brain Browser Web Interface, Brain Map WIRM Repository, Brain Visualizer Graphics Server, Single-Cell Recording EMS, X_Batch, XBrain and MindSeer.