|
|
Eugene Agichtein. MathCS.
Phone: 404-727-7962
Email: eugene@mathcs.emory.edu
Institution: Emory
Location: On Campus (Emory main campus)
Availability: Spring,Summer,Fall
Lab Positions: 0
Project Description: Project 1: Extracting structured information from textual data in biomedical databases
Text data is ubiquitous and the information and knowledge locked therein are of crucial importance for a wide range of tasks. For example, key nuggets of information are buried in text annotations in biological databases, in laboratory results, and in millions of published biomedical literature abstracts. Information extraction and text mining techniques can be used to automatically structure, transform, and organize this information to fully exploit the textual data for life sciences applications. The focus will be on extracting structured information from biomedical literature (e.g., Medline), as well as textual annotations in biological databases. The methodology will be to apply state-of-the-art information extraction and machine learning techniques for developing proof-of-concept prototypes for biomedical information extraction and text mining. For example, within a 10-week period a student might implement a published information extraction/machine learning algorithm that has been used in other domains (e.g., Web information extraction) and evaluate the performance on biomedical text. This effort would be in the context of the long-term research goal of develop adaptive, robust, and scalable text information extraction and knowledge discovery techniques for the life sciences domain. The experimentation will be performed within the framework of Emory Text Mining framework (EmText) that I am developing.
Student Requirements: Seniors only, with substantial computer science and programming coursework.
Suggested Reading (References):
(1) Hong Yu and Eugene Agichtein, Extracting Synonymous Gene and Protein Terms from Biological Literature, Bioinformatics, 2003, Vol 19, pages i340-i349
(2) Eleazar Eskin and Eugene Agichtein, Combining Text Mining and Sequence Analysis to Discover Protein Functional Regions, 2004, Proceedings of Pacific Symposium on Biocomputing (PSB)
(3) Eugene Agichtein, Confidence Estimation Methods for Partially Supervised Relation Extraction, 2006, SIAM Conference on Data Mining (SDM)
(4) Eugene Agichtein and Venkatesh Ganti, Mining Reference Tables for Automatic Text Segmentation, 2004, Proceedings of the ACM International Conference on Knowledge Discovery and Data Mining (KDD)
(5) Eugene Agichtein and Silviu Cucerzan, Predicting Accuracy of Extracting Information from Unstructured Text Collections, 2005, Proceedings of the ACM Conference on Information and Knowledge Management (CIKM)
Techniques used in this lab: Statistical Natural Language Processing, focusing on applying machine learning techniques for structured Information Extraction from biomedical text data.
|
Keith Berland. Physics.
Phone: 404 712 9061
Email: kberland@physics.emory.edu
Institution: Emory
Location: On Campus (Emory main campus)
Availability: Spring,Summer,Fall
Lab Positions: 2
Project Description: A wide variety of undergraduate research opportunities are available in the biophotonics lab, ranging from the development of novel optical instrumentation to the application of high-sensitivity fluorescence measurements to investigate protein dynamics and interactions in living cells. Many of our research projects are highly interdisciplinary, and appropriate for students interested in physics, biophysics, biochemistry, and even cell biology. Students interested in instrumentation can participate in designing and building new optical devices, or in writing software for instrument control and data analysis. Interested students should contact the PI about specific current opportunities.
Student Requirements: Preferable to have a strong math background and or computer programming skills, but not specifically required. Molecular biology skills are also useful.
Accepts 2nd year students? Y
Suggested Reading (References):
(1) Observation Volumes and Gamma Factors in Two-Photon Fluctuation Spectroscopy. Biophysical Journal. Vol. 89, 2077-2090
(2) Characterizing Observation Volumes and the Role of Photophysical Dynamics in One-Photon Fluorescence Fluctuation Spectroscopy. Journal of Biomedical Optics Vol 10(4). 044015, 1-9
(3) Saturation Modified Point Spread Functions in Two-Photon Microscopy. Microscopy Research and Technique. Vol. 64, 135-141.
(4) High Sensitivity Detection of Specific DNA Molecules Using Dual-Color Two-Photon Fluorescence Correlation Spectroscopy. Journal of Biotechnology. Vol. 108, 127-136.
(5) Fluorescence Correlation Spectroscopy: A New Tool for Quantification of Molecular Interactions , Protein-Protein Interactions: Methods and Protocols (ed. H. Fu), Humana Press. Pp. 383-397.
Techniques used in this lab: Some reserach tools you may learn about and use in the lab include: Fluorescence Microscopy, Fluctuation Spectroscopy, Laser Physics, Two-photon microscopy, Nuclear Localization Signal Biophysics, Protein Conjugation, Amyloid Peptide Self-assembly, Biophysics of the Intracellular Environment
|
Li Xiong. Math & Computer Science.
Phone: 404-778-0758
Email: lxiong@mathcs.emory.edu
Institution: Emory
Location: On Campus (Emory main campus)
Availability: Spring,Summer,Fall
Lab Positions: 2
Project Description: Evaluation of De-identification Methods - Current information technology enables many organizations to collect, store, and use various types of information about individuals in large repositories. Such a wealth of information provides enormous opportunities for data analysis but also raises serious concerns about the privacy of individuals. For example, individually identifiable health information is protected under the Health Insurance Portability and Accountability Act (HIPAA). Privacy preserving data publishing deals with release of anonymized views of personal information for analysis while preserving privacy of individuals. The project will evaluate a number de-identification methods for medical text document in terms of their accuracy, efficiency and implications on data utility.
Student Requirements: Experience with computer science and programming (e.g. CS170) and database systems and information management (e.g. CS447).
Accepts 1st year students? Y
Accepts 2nd year students? Y
Suggested Reading (References):
(1) L. Xiong, K. Boronda, C. Flowers, M. Graiser. De-identification of Medical Text. Technical Report TR-2007-012-A. Emory University Department of Mathematics and Computer Science. 2007
Techniques used in this lab: Computer programming, data collection and analysis
Additional Comments: Please feel free to contact me if you are interested or have any questions.
|
EUGENE DEMCHUK. Division of Toxicology.
Phone: 770-488-3327
Email: edemchuk@cdc.gov
Institution: CDC/ATSDR
Location: Off-campus (personal vehicle required, carpool possible but not guaranteed)
Availability: Spring,Summer,Fall
Lab Positions: 2
Project Description: Autism Spectrum Disorder (ASD) is an increasingly common developmental disability in industrial nations. ASD is thought to result from gene-environment interactions. Despite research progress in identifying candidate genes associated with ASD, no clear etiology or causative marker has been found. If and when a genetic predisposition is identified, the next research question will be: What environmental trigger is responsible for the development or manifestation of the clinical phenotype? To address this question we develop a rapid-screening computational toxicology methodology which can be applied to large numbers of environmental pollutants (ligands) and a known or suspected biological target for autism. Starting with a model database of hypothesized chemical triggers and a set of critical-pathway genes, we screen the chemicals against known genetic variants using state-of-the-art molecular docking techniques. Top scored gene/chemical combinations potentially offer an educated choice for further in-depth analysis of gene-environment interactions using laboratory and/or epidemiological methods.
Additional Project Information: The Agency for Toxic Substances and Disease Registry (ATSDR) Computational Toxicology and Method Development Laboratory implements the full range of methods in support of ATSDR mission to protect human populations from exposure to environmental contaminants. These include benchmark dose, chemical-specific adjustment factor, physiologically-based pharmacokinetic, quantitative structure-activity relationship (QSAR), genetic-susceptibility- and meta-analysis modeling, and modeling the toxicity of chemical mixtures. Computational toxicology methods are used as an integrated systematic approach in the development of ATSDR Minimal Risk Levels to be used as health guidance values to protect populations exposed to toxic chemicals at hazardous waste sites. These methods are also used in the development of ATSDR Toxicological Profiles, to support environmental health consultations and prioritization of environmental chemical hazards, when experimental information is insufficient, and to improve study design, when filling the priority data needs as mandated by the Congress. Also, the Laboratory is engaged in the development of response strategies to new emerging chemical threats. We develop methods for assessing toxicological effects of potentially hazardous chemicals from their chemical structure alone. A need for analysis of this type is especially imminent during the times of emergencies, whether it is an accidental chemical release, major natural disaster, or terrorist threat in all situations when time is a critical element of public health response.
Student Requirements: chemistry, toxicology and/or physiology, statistics, biochemistry, basic understanding of principles in physics, basic math
Accepts 1st year students? Y
Accepts 2nd year students? Y
Suggested Reading (References):
(1) Demchuk, E.; Ruiz, P.; Wilson, J.D.; Scinicariello, F.; Pohl, H.R.; Fay, M.; Mumtaz, M.; Hansen, H.; De Rosa, C.T. Computational toxicology methods in public health practice . Toxicol. Mech. Method. 2008, 18, 119135.
(2) Snyder, J.A.; Demchuk, E.; McCanlies E.C.; Schuler, C.R.; Kreiss, K.; Frye, B.; Ensey, J.; Stanton, M.; Weston, A. Impact of negatively charged patches on the surface of MHC class II antigen-presenting proteins on risk of chronic beryllium disease . J. R. Soc. Int. 2008, 5, 749758.
(3) Demchuk, E.; Albin, B.C.; Fay, M.; Garrett, R.M.; Hansen, H. Structure-activity analysis of chemical health guidance values . Toxicologist (Suppl. to Toxicol. Sci.) 2006, 90, 186.
(4) Demchuk, E.; Yucesoy, B.; Johnson, V.J.; Weston, A.; Germolec, D.; De Rosa, C.T.; Luster, M.I. A statistical model to assess genetic susceptibility as a risk factor in multifactorial diseases: Lessons from occupational asthma . Environ. Health Persp. 2007, 115, 231234.
(5) Hnizdo, V.; Darian, E.; Fedorowicz, A.; Demchuk, E.; Li, S.; Singh, H. Nearest-neighbor nonparametric method for estimating the configurational entropy of complex molecules . J. Comp. Chem. 2007, 28, 655668.
Techniques used in this lab: Students may learn various computational toxicology techniques, including benchmark dose modeling, chemical-specific adjustment factor modeling, physiologically-based pharmacokinetic/pharmacodynamic modeling, (quantitative) structure-activity relationship -- (Q)SAR modeling, genetic-susceptibility- and meta-analysis modeling, modeling the toxicity of chemical mixtures and chemical-chemical interactions, molecular docking, protein homology structure modeling, and other.
Additional Comments: A brief description of the ATSDR Computational Toxicology lab can be found at http://www.atsdr.cdc.gov/dtem/programs/comptox/index.html
|
|
