Please use this identifier to cite or link to this item:
|Title:||Prediction of putative adverse drug reaction-related proteins from primary sequence by support vector machines|
|Citation:||Zhi, L.J., Lian, Y.H., Chan, J.Z., Zhi, W.C., Yu, Z.C. (2005). Prediction of putative adverse drug reaction-related proteins from primary sequence by support vector machines. International Journal of Pharmaceutical Medicine 19 (5-6) : 317-322. ScholarBank@NUS Repository. https://doi.org/10.2165/00124363-200519050-00009|
|Abstract:||Introduction: Adverse drug reactions (ADRs) are responsible for the failure of a significant portion of investigative drugs trials and the major reason for the withdrawal of drugs from clinical research. A number of ADRs are caused by the (undesired) interaction of drugs with key proteins involved in normal biological processes. Identification of these ADR-related proteins facilitates the design of drugs with fewer adverse effects by rationally avoiding unwanted interaction with these proteins. Method: This work explores the use of a statistical learning method, support vector machines (SVMs), for the identification of potential ADR-related proteins. A SVM classification system was trained and tested by using 759 ADR-related proteins of different species and 2280 non-ADR-related proteins. Results: 93.9% of the ADR-related proteins and 98.2% of non-ADR-related proteins were correctly classified. Discussion: The SVM is potentially useful for facilitating the identification of ADR-related proteins. The development of methods to identify ADR indications of ADR-related proteins are progressing well, an example of which is the web-based ADR-related protein prediction tool SVMDART, which can be accessed at http://jing.cz3.nus.edu.sg/cgi-bin/dart.cgi. © 2005 Adis Data Information BV. All rights reserved.|
|Source Title:||International Journal of Pharmaceutical Medicine|
|Appears in Collections:||Staff Publications|
Show full item record
Files in This Item:
There are no files associated with this item.
checked on Nov 2, 2018
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.