"TrAPSS - Knowledge Discovery for Disease Gene Mutation Discovery"
Tom Casavant presenting on behalf of Terry Braun,
Dept. of Biomedical Engineering
TrAPSS - a bioinformatics success story for age-related macular
degeneration.
TrAPSS is bioinformatics system that prioritizes regions of genes that
are likely to contain disabling sequence variations within the coding
regions of genes. TrAPSS relies on sequence-based annotation such as
conserved protein functional domains, transmembrane domains, and protein secondary structure predictions. Retrospective studies of the manually
optimized strategy resulted in a 2-fold increase in the mutation
identification efficiency. TrAPSS was used for a mutation screening
study that included the fibulin gene family. Sequence variations in the
fibulin 5 (FBLN5) gene were identified that represent a significant
cause of age-related macular degeneration (AMD).
PowerPoint
presentation
"Rapid Computational Identification of Therapeutic Drug Targets: Application to Protein Kinase Inhibitors"
Adrian Elcock, Dept. of Biochemistry
Therapeutic drugs that are intended to inhibit a particular protein receptor
will often also inhibit other related receptors. In some cases, this cross-reactivity may be beneficial, as a drug known to be useful for treating one
disease may also be found to be of use in treating another. In other cases,
cross-reactivity is detrimental, resulting in clinical side-effects that can
severely limit the therapeutic potential of a drug. There is therefore a
pressing need for methods capable of identifying unintended drug-receptor
interactions at an early stage of the drug development process. In this talk
I
will describe my laboratory's attempts to develop and test one such method,
that combines techniques commonly used in computational chemistry with
information obtained from recent genome sequencing efforts. By way of
example,
I will describe our attempts to predict the targets of protein kinase
inhibitors, drugs that face the difficult task of aiming to inhibit only a
handful of the ~500 closely-related human protein kinases known to exist.
The
advantages and limitations of the method will be discussed, and directions
for
future developments outlined.
PowerPoint presentation
"Genetic Approaches to Human Eye Disease: The Pitfalls and the Promise"
Michael Grassi, Dept. of Opthalmology
Numerous ophthalmologic entities have been identified and exquisitely
described. Their natural history, clinical manifestations, and
associations have been documented. What remains to be determined for
many of these conditions, though, is the basic underlying pathogenesis.
Understanding the genetic basis of disease will pave the way to future
molecularly targeted interventions. The foundation of this new era in
medicine rests on an ability to correlate genotype with phenotype. We
will discuss our multidisciplinary approach to this challenge.
PDF
of PowerPoint presentation
| Streaming video - |
http |
rtsp |
14:47
|
"Imaging-based Normative Lung Atlas: Basis for Detection, Quantitation and Tracking of Lung Pathology"
Eric Hoffman, Depts. of Radiology, Physiology and Biomedical Engineering
There is a critical need for sensitive and objective measures of regional lung status both for detection of disease and for outcomes analysis. X-ray CT provides a comprehensive solution to this need.. Significant advances are being made in both temporal and spatial resolution. Scan apertures are at sub-cardiac cycle time frames, allowing for the imaging of not only anatomy but also function.
We are developing an atlas of the human lung. This atlas allows for inter and intra subject matching and includes the lungs, lobes sub-lobar segments and airway and vascular branching structures of the lung and attached to each level of this structure will be the normal range of the CT-based measures of regional lung physiology including ventilation, perfusion.
Methods have been established for the segmentation of the lungs, lung lobes, and the airway and vascular trees. A centerline algorithm has been developed which provides airway branch points, allowing for the automatic labeling of the airway tree with a standard nomenclature and the matching of lung volumes. Quantitative airway geometry is automatically generated. Inter-subject matching then serves as a first step in establishing the normative atlas. Protocols have been established for the use of iodinated contrast agent and xenon gas to evaluate regional perfusion and ventilation respectively. Parenchymal pathology is evaluated by means of an Adaptive Multiple Feature Method (AMFM) for characterization of both anatomic and functional texture. An open source software environment, MIFAR (Medical Imaging File Archive) has been established for archiving image data along with associated non-image-based information and post image analysis data. Normative data are being integrated into the International Union of Physiological Sciences Physiome construct.
PowerPoint presentation
| Streaming video - |
http |
rtsp |
25:17
|
"Spatial and Temporal Modeling of PCBs and Other Persistent and
Bioaccumulating Toxics in Lake Michigan"
Keri Hornbuckle, Dept. of Civil & Environmental Engineering
Polychlorinated biphenyls (PCBs) are persistent and bioaccumulating
toxic compounds that are easily and rapidly transported across the
globe. These compounds are found in plants, soils, and animals, in
every part of the world. These compounds can travel long distances
from their original sources because they can be transported as gases,
they are not efficiently washed out the atmosphere by rain, and they
degrade slowly. Their persistence and ubiquity is of concern because
they can bioaccumulate to toxic levels in humans and animals. Their
toxic effects include reproductive damage, learning disabilities in
children and adults, and cancer.
Although POPs have very low vapor pressures (10-3 to 10-10 atm) and
are very hydrophobic (octanol/water partition coefficients are 104 to
109), they can rapidly transform from the sorbed (condensed) phase to
the gas phase. The nature of this transformation, although a critical
aspect of any predictions for atmospheric transport, are not well
understood. Kinetic and equilibrium models have been constructed to
explain the rate and tendancy of POPs to move from water, soil, plant
surfaces to the air. These models have been coupled to provide a
dynamic mass balance on PCBs in Lake Michigan - a water body that
remains contaminated with PCBs at a level that result in fish
consumption advisories from the FDA.
PowerPoint
presentation
"We Live in Interesting Times: Electronic Health Records, Digital Libraries, and Public Policy"
Betsy Humphreys, National Library of Medicine
Presidential pronouncements on electronic health records and data standards, Congressional pressure for free electronic access to published results of government-funded research, controversy about digital registries of clinical trials - 2004 has seen a substantial increase in the public debate about once relatively obscure informatics issues. Given the range of its activities (e.g., the Unified Medical Language System, PubMed Central, ClinicalTrials.gov, MedlinePlus, GenBank, funding for informatics research, development, and training), the National Library of Medicine(NLM) and its programs play multiple roles in efforts to develop a robust National Health Information Infrastructure (NHII). This talk will review the background, current status, and likely future of key NHII developments and related NLM programs and services.
PowerPoint
presentation
"Snomed CT Reference Terminology Mandate & Post-mapping Validation Methodology for Nursing Vocabularies"
Der Fa Lu, College of Nursing
Standardized nursing
language (SNL) and reference terminology are important in the development
of the electronic health record (EHR), information to support patient care
quality and safety as well as discipline specific knowledge building. Major
standardized languages in health care have been mapped to the Systematized
Nomenclature of Medicine Clinical Terms (SNOMED CT). This study aimed to
develop and test a method for validating the fitness of cross-mapping between
the SNOMED CT and standardized nursing languages. The background of SNOMED
CT and NIC are summarized. The methodology is based on tree edit distance
between two selected terms in Nursing Intervention Classifications (NIC)
compared to the edit distance between the same two terms in SNOMED CT.
Examples of the finding are illustrated. Results of this project can be
applied to validate similar cross-mapping of other classifications, and
provide recommendations back to language developers. Discussions, implications
and future work are detailed.
PowerPoint
presentation
"Data Acquisition and Dissemination in the Large Scale Digital Cell Analysis System
"
Michael Mackey, Dept. of Biomedical Engineering
The Large Scale Digital Cell Analysis System (LSDCAS) is a live-cell imaging facility at the Holden Comprehensive Cancer Center at the University of Iowa. Using LSDCAS, biomedical researchers are conducting investigations into many areas of basic and applied biological research. The primary design goal of LSDCAS is to provide a flexible digital image platform to facilitate direct comparison between molecular studies in the laboratory and quantitative measures of the effects of such manipulations upon the dynamics of a variety of cellular and molecular endpoints in living cells. To accomplish this goal, LSDCAS is continually being refined to detect and quantify a number of experimental endpoints. These efforts will result in the first open-source live cell imaging system available to the biomedical research community. The large data production capabilities of LSDCAS (3.8 terabytes / year at peak usage) present challenges in data management, archiving, and data dissemination to the LSDCAS user community. For reliable and efficient data management, two enterprise-scale data centers have been established to support LSDCAS. As LSDCAS researchers are spread across the University campus, and are also to be found as far away as Latvia, it was necessary to provide secure internet access to the LSDCAS data. Accordingly, a multi-server system was developed using Apache Tomcat as a secure servlet container, interfaced to a PostgreSQL relational database system. To allow LSDCAS users access to digital image streams, a custom java-based client-server application was developed. For image segmentation and analysis by LSDCAS researchers, standalone analysis applications have been developed. All software runs on Linux- and Unix-based systems. This presentation will summarize our experience in the design and implementation of the LSDCAS facility, with particular emphasis on issues related to the automated management of the large volume of data produced by LSDCAS, its flexible acquisition, analysis and backup, as well as secure mechanisms to allow users to access these data.
PowerPoint
presentation
"21st Century Biomedical Informatics: Confronted by Insurmountable Opportunity"
Daniel Masys, UC San Diego School of Medicine
Medical Informatics
applies the principles of information science and the
technologies of computing and telecommunications to patient care,
education
and biomedical research. This talk will review recent developments in
the
field, and the challenges confronted by clinicians, researchers, and
patients in the skillful use of information technologies for improving
health.
PowerPoint
presentation
"Terminology to Capture Patient Care: Normalization Challenges"
Sue Moorhead, College of Nursing
Over the past several decades nursing has identified the need for standardized languages to describe the care provided to individuals, families, and communities and has developed nursing languages to capture the essence of this care in 3 areas that are essential to nursing: patient problems, nursing interventions or treatments, and patient outcomes. This work has been in concert with the steps of the nursing process and has developed the content or knowledge components for this approach. This ongoing work provides organized classifications of standardized terms for use in nursing practice as well as terms for the electronic health record yet the languages of nursing and the electronic health record have developed along side each other. Today many different groups are working together to make an electronic record that will meet the needs of a diverse health care system. This paper will briefly describe 3 nursing languages used by nurses and will focus on the issues of normalization from the view point of language developers.
PowerPoint
presentation
| Streaming video - |
http |
rtsp |
33:50
|
"Patient Care Documentation - Structured Vocabulary - Natural Language Dichotomy"
Aleta Porcella, College of Nursing
Today's rapidly changing health care environment creates pressure for the computerization of the patient record. Two requirements for inclusion of nursing activities into the computerized patient record (CPR) are a standardized nursing language of sufficient granularity and a database that allows for one time collection of data for multiple uses. Documentation systems raise issues of data completeness. Using a descriptive methodology, nursing documentation in one CPR was examined for prevalence and content of free text documentation in an otherwise structured nursing information system (NIS). Results demonstrate house wide use of free text (narrative note) fields. Variability in use unrelated to patient acuity suggests idiosyncratic individual or unit documentation practices. Findings support the use of quality management activities to improve documentation practices and point to areas of database enhancement and information system development.
PowerPoint
presentation
"Exploring Spatial Patterns of Disease with
Geographic Information Systems"
Gerard Rushton, Dept. of Geography
I will review recent developments in mapping the geographic patterns of disease. I will show how the purpose of making disease maps is changing from the atlas format to interactive, web-based, mapping services that supply "linked maps" from which users can explore relationships between disease patterns and other socio-economic or environmental patterns. Using colorectal and breast cancer incidence records from the Iowa Cancer Registry, I will illustrate some of the problems of mapping and interpreting disease patterns. Finally, I will show how the process of geo-coding health and environmental data can become the basis of scientific studies that explore relationships between environmental exposures and health effects. Problems with maintaining the privacy and confidentiality of health records will also be discussed.
PowerPoint
presentation
"Informatics for Efficient EST-based Gene Discovery"
Todd Scheetz, Center for Bioinformatics
Several large-scale efforts in EST-based gene discovery in Rat, Human, Mouse, and other species have been conducted at the University of Iowa, beginning in 1997. To date, these efforts have led to the sequencing of more than 1,000,000 ESTs accounting for more than 100,000 novel ESTs, including 40,000 previously undescribed genes. This effort has been augmented by a set of custom informatics tools to gather, analyze, store, and retrieve the sequence data generated. The high rate of gene discovery associated with this work was primarily due to novel cDNA library normalization and subtraction methodologies (Bonaldo et al. 1996). A critical aspect of this work is to periodically perform subtractive hybridizations of cDNA libraries against a set of previously sequenced clones from that library. The informatics necessary for this effort consists first of examination of individual sequences to verify the presence of a 3' end, and to confirm the clone identity. Second, it is necessary to cluster the sequences to determine the current novelty of the library and to allow feedback to the cDNA library subtraction process to remove redundant clones and improve overall discovery efficiency. Finally, informatics is required for the submission of the EST sequences and associated annotation to both public and local databases.
PowerPoint
presentation
"Dynamic Image Analysis System (DIAS) Technologies"
David Soll, Dept. of Biological Sciences
To answer a variety of biomedical research questions, one must be able to reconstruct live migrating cells in 2D and 3D at short time intervals, and quantitate their behavior. These questions include how a cell responds behaviorally to extracellular signals, how mutated genes affect the regulation of the actin cytoskeleton, how cells organize during embryogenesis, how metastasizing cancer cell behavior differs from normal cell behavior and how white blood cell diseases in children affect cell motility. To this end, we have developed computer-assisted 2D and 3D dynamic image analysis systems (2D-DIAS, 3D-DIAS) that provide not only 2D and 3D representations of crawling cells, including reconstructed pseudopods and nuclei, but also reconstructions of every cell and nucleus at short time intervals during early embryogenesis. In addition, 3D reconstructions have been generated of human white blood cells crawling towards a wound, a model for studying the effects of relevant drugs. Because every cell, pseudopod and nucleus reconstructed in 2D is isolated in a separate trace slot, 40 parameters of motility and dynamic morphology are computed for each component of a cell at time intervals as short as a thirtieth of a second. For 3D cell reconstructions, over 100 parameters are computed for each component at one second intervals and for 3D embryo reconstructions, over 100 parameters are computed for every individual cell and nucleus at five second intervals. DIAS-4.0, a new JAVA-based system that performs both 2D and 3D reconstructions, and possesses new reconstruction and motion analysis capabilities, will be described. In particular, new programs in progress will be demonstrated for reconstructing and motion-analyzing in 4D intracellular fiber networks, filopodia, GFP-tagged subcellular complexes, cells in complex tissues and embryos, and echocardiograms.
PDF
of PowerPoint presentation
"Virtual Liver Surgery Planning System using Augmented Reality"
Milan Sonka, College of Engineering
A system for virtual planning of liver tumor resections will be discussed that consists of the environments for diaphragm, liver, liver tumor, and liver vasculature segmentation and collaborative surgery planning visualization. The system is under development and many individual modules will be discussed and their functionality presented.
PDF
version of PowerPoint presentation
"Text Mining: Challenges and Opportunities"
Padmini Srinivasan, School of Library and Information Science
Serendipity has often shaped the pathways of science. Classic examples include Flemming's discovery of penicillin upon observing a culture of Staphylococcus dissolving when the plate was accidentally contaminated with a blue-green mold. But what if there were systems designed to make such discoveries more likely and in effect make them less serendipitous? Clearly such systems would have to separate the potentially meaningful connections from a vast and mostly noisy background of random associations. The area of text mining, also known as literature-based knowledge discovery, is concerned with this goal. More recently this area applied to biology has also been termed `conceptual biology'. In this talk I will present an overview of our text mining approach which may be used to explore individual topics or groups of topics. We have conducted several experiments testing our overall approach that have yielded reasonable results. This talk will briefly outline completed work and then present an overview of current research directions being undertaken as well as describe particular challenges in this area.
PowerPoint
presentation
"CDOCS: A Tool for Collaborative Consensus for New Phenotype Vocabularies"
Andrew Williams, Spelman College, Atlanta, GA
Often clinicians observe variations of disease in the clinic but do not
have a standardized vocabulary for them. These new disease subtypes
provide phenotype information that may aid in the disease gene discovery
process. The Collaborative Distributed Ontology Consensus Systems
(CDOCS) is being developed to aid clinicians that are distributed
geographically and temporally to collaboratively reach a consensus on
new disease subtypes. This talk will describe the CDOCS system and
approach for using artificial intelligence to aid clinicians in
translational cancer and eye disease research. We will examine the
issues involved and describe the technical solutions being developed.
PowerPoint
presentation
