Department of Health Information Technology, Faculty of Management and Medical Informatics, Tabriz University of Medical Sciences, Golghast St., Tabriz, 5166614711, Iran.
Pharmaceutical Analysis Research Center and Faculty of Pharmacy, Tabriz University of Medical Sciences, Tabriz, Iran.
BMC Med Inform Decis Mak. 2023 Feb 14;23(1):35. doi: 10.1186/s12911-023-02133-3.
The measurement of drug similarity has many potential applications for assessing drug therapy similarity, patient similarity, and the success of treatment modalities. To date, a family of computational methods has been employed to predict drug-drug similarity. Here, we announce a computational method for measuring drug-drug similarity based on drug indications and side effects.
The model was applied for 2997 drugs in the side effects category and 1437 drugs in the indications category. The corresponding binary vectors were built to determine the Drug-drug similarity for each drug. Various similarity measures were conducted to discover drug-drug similarity.
Among the examined similarity methods, the Jaccard similarity measure was the best in overall performance results. In total, 5,521,272 potential drug pair's similarities were studied in this research. The offered model was able to predict 3,948,378 potential similarities.
Based on these results, we propose the current method as a robust, simple, and quick approach to identifying drug similarity.
药物相似性的测量在评估药物治疗相似性、患者相似性和治疗方式的成功方面有许多潜在的应用。迄今为止，已经采用了一系列计算方法来预测药物-药物相似性。在这里，我们宣布了一种基于药物适应症和副作用的药物-药物相似性测量计算方法。
该模型应用于副作用类别中的 2997 种药物和适应症类别中的 1437 种药物。建立相应的二进制向量来确定每种药物的药物-药物相似性。进行了各种相似性度量以发现药物-药物相似性。
在所检查的相似性方法中，杰卡德相似性度量在整体性能结果中表现最好。在这项研究中，总共研究了 5521272 对潜在药物对的相似性。所提供的模型能够预测 3948378 对潜在相似性。
基于这些结果，我们提出当前的方法是一种强大、简单和快速的识别药物相似性的方法。
