科研之友
微信
新浪微博
Facebook
分享链接
摘要
Most drugs are small molecules. Small molecular drugs are more likely to be absorbed in organism. A small molecular drug in the field of pharmacology is a low weight organic compound that binds to a target (for example: biopolymer) to alter the activity or function of the target. Therefore it is critical to predict drug-target interactions with computational methods in the field of pharmacology. Supervised prediction with bipartite Local Model recently has been shown to be effective for prediction of drug-target interactions. However, this pure %26quot;local%26quot; model is inapplicable to new small molecular drug or target candidates that currently have no known interactions. In this paper, we extend the existing supervised learning approach bipartite local model (BLM) by integrating a strategy for handling new small molecular drug and target candidates. Based on the assumption that similar drugs and targets have similar interaction profiles, we present a simple neighbor-based training data inferring procedure and integrate it into the frame work of BLM. This globalized BLM called bipartite learning model with neighbor-based inferring (BLOC) then has an extended functionality for prediction interactions between new small molecular drug candidates and target candidates. Consistent good performance of BLOC has been observed in the experiment for the prediction of interaction between small molecular drugs and four important categories of targets. For the Nuclear Receptors dataset, where there are more chances for the presented strategy to be applied, 20% improvement in terms of AUPR has been achieved. This demonstrates the effectiveness of BLOC and its potential in small molecular drug-target interaction prediction.
- 出版日期2012-12
- 单位南阳理工学院
