Oct. 1, 2019
Extension of the informatics framework for the structurization of free text diagnostic reports
In this seminar presentation, we will briefly discuss our recent project to extend our informatics pipeline for the structurization of free-text diagnostic reports with an information theory-based approach. We will focus on the steps of this approach that are for quantifying and measuring information in free text diagnostic reports. This is a work in progress.
Sep. 24, 2019
Analysis of Curated Toll-Like-Receptor Co-expression Network Reveals associations among RNAseq Expression and Immune system modulation in patients with Colorectal Cancer
This project is concerned with improving immune system capabilities in the context of immunotherapy for colorectal cancer treatments. Immunotherapies called checkpoint inhibitors are monoclonal antibody treatments that have been proven to be effective in small percentages of sample populations, especially upon a combination of treatments (e.g. PD-1 with CTLA-4). However, the effectiveness of the treatment response often accompanied by contraindications of autoimmunity. Allocation of our efforts towards computational modeling of communication networks specific to TLR signaling affords a more intimate study of innate immunity, and the corresponding responses generated, as they pertain to the Tumor Microenvironment. Targeting the mechanisms…
Sep. 24, 2019
Addressing Challenges of Deep Learning for Post-Disaster Damage Classification
The effectiveness of natural disaster response functions has increasingly relied on information extracted from remotely-sensed imagery – particularly for the time-critical and often large-scale task of assessing post-disaster damage to the built environment. Despite a recent push to operationalize deep learning (DL) technologies to assist this task, several overlooked and unaddressed challenges limit DL’s applicability. First, very little DL research in the remote sensing domain utilizes the geographic variation in objects of interest or ancillary spatial data layers as features which could help the algorithms learn. Second, there is a scarcity of labelled training data depicting the effects of common…
Aug. 28, 2019
Using Geospatial Context: Facilitating Geospatial Analysis in Research
This presentation will briefly review the Geospatial Context Big Table status and its current evolution. Then the presentation turns to two projects that have used or are using geospatial context in the analysis and visualization of their results. The first project is a zoonotic disease study originally undertaken as a class project and lab rotation. The second is a study of Thyroid Cancer and possible exposome radiation influences and population stability metrics. In each case focus will be on developing focus and assembling data and possible surrogates, compiling into units of analysis, and the modeling of those relationships.
Aug. 19, 2019
Applying Blockchain Technology to Enhance Clinical Trial Recruitment
Patient recruitment for clinical trials is known to be a challenging aspect of clinical research. There are multiple competing concerns from the sponsor, patient and principal investigator’s perspectives resulting in most clinical trials not meeting recruitment requirements on time. Conducting under-enrolled clinical trials affects the power of conclusive results or causes premature trial termination. Blockchain is a distributed ledger technology originally applied in the financial sector. Its features as a peer-to-peer system with publicly audited transactions, data security, and patient privacy are a good fit for the needs of clinical trials recruitment. The “Smart Contract” is a programmable self-executing protocol…
May 6, 2019
Discovery of Homogeneous Subgroups from Heterogeneous Populations for Precision Health – A Deep Exploratory Mining and X2AI Approach
Today, six of the top ten highest-grossing drugs in the US are effective in less than 10% of patients and even the most effective drugs from that list have positive responses in only 25% of patients. This “imprecision medicine” practice not only harms certain populations of patients, it also burdens the healthcare system financially. By finding meaningful and homogeneous subgroups prior to conducting costly clinical trials, researchers can further study focused populations and identify potential risk factors through slicing and dicing from complex phenotypic/genotypic information sources. Advancements in machine learning algorithms have shown promising results in many biomedical applications and…
May 1, 2019
Using Predictive Analytics to Improve Surveillance of Heat-Related Illnesses During Military Training
Heat-related illnesses are important occupational risks in military personnel, especially for soldiers who do not have experience with hot climate regions. Our study focuses on predictive analysis of heat-related illnesses to improve prevention program. The Royal Thai Army (RTA) Medical Department collected data from conscripts during a 10-week military training program in 2013. To build predictive analytic models, we applied various machine learning and deep learning methods, including generalized linear model (GLM), k-nearest neighbors (kNN), random forests (RF), eXtreme gradient boosting (XGB), deep neural networks (DNN), and convolutional neural networks (CNN). We compared accuracy between different models and we found…
May 1, 2019
Uncovering complex relationships in biomedical data through hierarchical logic pattern contrast mining
Understanding the relationships which exist between features in medical data ranging from electronic health records to genetic variations in sequenced genomes is key to understanding how these features impact the medical condition of an individual. Existing pattern mining methods are unable to discover relationships more complex than co-occurrence which limits their usefulness in searching for patterns associated with medical conditions which may include inhibitory and many-to-one relationships between features. Our algorithm can extract complex nested logical relationships which can provide additional information about how individual features interact to affect medical outcomes. We demonstrate the effectiveness of our algorithm on two…
April 9, 2019
Linking EMR and Exposome Data for Risk Prediction and Interventions: A Translational Approach
Precision medicine (PM) is a medical model that proposes the customization of healthcare, with medical decisions, treatments, practices, or products being tailored to the individual patient. An individual’s “Social Determinants of Health” (SDOH) have been demonstrated as a key factor in obtaining successful clinical outcomes for individual patients which necessitate individualized interventions. Two major problems exist in addressing SDOH within a clinical setting. First, interventions that have shown to be successful in addressing challenges presented by various Social Determinants of Health often scarce and span outside of those services that are available and/or reimbursed within a healthcare setting. Because of this, it is critical…