This research topic can only be taken by Thesis route and Project route student. This research topic include the following data mining mechanism design and system development. The project outcome needs to be examined by well-designed experiment for quanititative and qualitative data analysis (for Thesis route student) and performance evaluation include speed and accuracy comparisons (for both Thesis and Project students).
Many people may have chance to see more than one doctor and receive prescriptions from different doctors. Due to privacy issue and accessiblility, a doctor may never know what kind of medications his or her patient is currently having. As we all know that some interactions existed between drugs, that is so-called drug-drug interactions. This research aims to design a mechanism which can extract and identify the potential interactions that two drugs may have and to provide users (including doctors and patients) alarm when the names of two drugs entered by the users.
Preparation Readings (and also for Speed Comparison purpose):
- Wishart, D. S., Knox, C., Guo, A. C., Cheng, D., Shrivastava, S., Tzur, D., Gautam, B., Hassanali, M. (2008). DrugBank: a knowledgebase for drugs, drug actions and drug targets. Nucleic Acids Research, 36(suppl 1), D901-D906. Retrieved on May 10, 2014, from http://nar.oxfordjournals.org/content/36/suppl_1/D901.full
- Magro, L., Conforti, A., Del Zotti, F., Leone, R., Iorio, M. L., Meneghelli, I., Massignani, D., Visona, E., Moretti, U. (2008). Identification of severe potential drug-drug interactions using an Italian general-practitioner database. European Journal of Clinical Pharmacology, 64(3), 303-309.
- Duda, S., Aliferis, C., Miller, R., Statnikov, A., and Johnson K. (2005). Extracting Drug-Drug Interaction Articles from MEDLINE to Improve the Content of Drug Databases. AIMA Annual Symposium Proceedings, 2005, 216-220. Retrieved on May 9, 2014, from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1560879/
- Bjorne, J., Airola, A., Pahikkala, T., and Salakoski, T. (2011). Drug-Drug Interaction Extraction from Biomedical Texts with SVM and RLS Classfiers. In the Proceedings of the 1st Challenge task on Drug-Drug Interaction Extraction, (DDIExtraction 2011), Huelva, Spain, September, 2011, 35-42. Retrieved on May 9, 2014, from http://ceur-ws.org/Vol-761/paper4.pdf
- Pathak, J., Kiefer, R. C., and Chute, C. G. (2013). Using Linked Data for Mining Drug-Drug Interactions in Electronic Health Records. Studies in Health Technology and Informatics, 192, 682-686. Retrieved on May 9, 2014, from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3909652/
- Kim, S., Jin, D., Lee, H. (2013). Predicting Drug-Target Interactions Using Drug-Drug Interactions. PLOS ONE, November 21, 2013. doi:10.1371/journal.pone.0080129. Retrieved on May 9, 2014, from http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0080129
- Rastegar-Mojarad, M. (2013). Extraction and Classification of Drug-Drug Interaction from Biomedical Text Using a Two-Stage Classifier. Theses and Dissertation. Paper 304. Retrieved on May 9, 2014, from http://dc.uwm.edu/cgi/viewcontent.cgi?article=1309&context=etd
- Cheng, F., and Zhao, Z. (2014). Machine learning-based prediction of drug–drug interactions by integrating drug phenotypic, therapeutic, chemical, and genomic properties. Journal of the American Medical Informatics Association. doi:10.1136/amiajnl-2013-002512. Abstract Retrieved on May 9, 2014, from http://jamia.bmj.com/content/early/2014/03/21/amiajnl-2013-002512.abstract
- Law, V., Knox, C., Djoumbou, Y., Jewison, T., Guo, A. C., Liu, Y., Maciejewski, A., Arndt, D., Wilson, M., Neveu, V., Tang, A., Gabriel, G., Ly, C., Adamjee, S., Dame, Z. T., Han, B., Zhou, Y., Wishart, D. S. (2014). DrugBank 4.0: shedding new light on drug metabolism. Nucleic Acids Research, 42(D1), D1091-1097. doi: 10.1093/nar/gkt1068. Retrieved on May 9, 2014, from http://nar.oxfordjournals.org/content/42/D1/D1091.full.pdf+html
For Accuracy Comparison purpose (you can NEVER use it directly): existing open access Drug-Drug Interaction datasets
- Boyce, R. D. (2011). Drug Interaction Knowledge Base 1.2. Retrieved on May 9, 2014, from http://dbmi-icode-01.dbmi.pitt.edu/dikb-evidence/front-page.html
- Boyce, R. D. (n.d.). D2R Server for the Drug Interaction Knowledge Base. Retrieved on May 9, 2014, from http://dbmi-icode-01.dbmi.pitt.edu/dikb/
- Boyce, R. D. (n.d.). All DDIPredictions. Retrieved on May 9, 2014, from http://dbmi-icode-01.dbmi.pitt.edu/dikb/all/DDIPredictions
- DrugBank. (2014). DrugBank - Open Data Drug & Drug Target Database. Retrieved on May 9, 2014, from http://www.drugbank.ca/
- The FreeMedForms Project. (2014). What is FreeDiams. Retrieved on May 10, 2014, from http://www.freemedforms.com/en/manuals/freediams/presentation
- Flockhart D. A. (2009). Drug Interactions: Cytochrome P450 Drug Interaction Table. Retrieved on May 10, 2014, from http://medicine.iupui.edu/clinpharm/ddis/
Tasks of this research also include:
- Conducting in-depth and comprehensive survey on relevant academic literature and research.
- Designing and implementing a mechanism (in standalone Java application running in a Debian virtual box) which can extract drug-drug interactions from text or relationships among drug metabolism. All relevant functions have to be developed in modularized way with classes and memeber functions to give flexibility and reusability in future development and enhancement for working with other systems.
- Having proper user interface for the standalone Java application to get the inputs (e.g., text, xml dataset, url for web crawler searching) and start extracting drug-drug interactions.
- Recording all the time spent that each stage in the proposed mechanism has, so the performance evaluation can be done later.
- Comparing the results that your proposed mechanism generated with existing datasets (i.e., the existing databases which have drug-drug interactions stored can be used as benchmark sets). Sometimes your proposed mechanism may perform better in one benchmark set while existing approaches performed not good enough. Also, sometimes due to the source your proposed mechanism uses as inputs may not be covered by particular benchmark set, and your proposed mechanism may not perform well when you compare the results with the particular benchmark set. So it is important to test your results with vary benchmark sets.
- Implementing existing approaches -- the existing approaches may or may not use similar method like your proposed mechanism, as we all know that if your proposed mechanism is not better than the existing ones in terms of performance, accuracy, or simpler degree, then the contribution of your research may not be important enough for researchers in the same field -- and recording all the time spent and results that each apprach has by using the same input sets. Also, the results are evaluated with the same benchmark sets.
- Develop web user interfaces for users to (1) enter two drug names; (2) enter two drug identification number; (3) enter a drug name and a drug identification number; and (4) enter two prescriptions with mixed drug names and drug identification numbers, to check if there is any drug-drug interaction. All relevant functions have to be developed in modularized JSPs and inputs/outputs should be packed into XML/JSON format to give flexibility and reusability in future development and enhancement for working with other systems.
- Having a project DVD which includes sources, executables, documents, database, demo video clips, and exported Debian image in .ova format.
For taking this topic, the student is better to have the following skills or knowledge:
- Familiar with XML, JSON, HTML, CSS, Java Script, and AJAX
- Having experience in developing application in Java standalone application (acts as startup daemons running in Debian OS), and JSP.
- Information collection, re-organization, summarization, and reports.
- Using screen capture software and video editing software/tool to creating instructional video clips
|
