DTIP-TC2A: An analytical framework for drug-target interactions prediction methods. (August 2022)

• Record Type: Journal Article
• Authors: Keyvanpour, Mohammad Reza; Haddadi, Faraneh; Mehrmolaei, Soheila
• Is Part Of: Computational biology and chemistry. Volume 99(2022)
• Publication Date: 2022-08
• Subjects: Drug-target interactions -- Challenge -- Qualitative evaluation -- Drug discovery -- Biological network analysis -- DTIP; Chemistry -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; Biochemistry -- Data processing; Biology -- Data processing; Molecular biology -- Data processing; Periodicals; Electronic journals; 542.85
• DOI: 10.1016/j.compbiolchem.2022.107707
• Languages: English
• ISSNs: 1476-9271

Abstract

Abstract : Identifying drug-target interactions through computational methods is raised an important and key step in the process of drug discovery and drug-oriented research during the last years. In addition to the advantages of existing computational methods, there are also challenges that affect methods' efficiency and provide obstacles in the direction of developing these computational methods. However, the literature suffers from lacking a comprehensive and comparative analysis concerning drug-target interactions prediction (DTIP) focusing on the analysis of technical and challenging aspects. It seems necessary to provide a comparative perspective and a different analysis on a macro level due to the importance of the DTIP problem. In this paper, we presented the quadruple framework of analytical, named DTIP-TC2A consists of four main components for DTIP. The first component, categorizing DTIP methods based on the technical aspect ahead and investigating the strengths and weaknesses of different DTIP methods. Second, classify DTIP challenges with a major focus on a well-organized and coherent investigation of challenges and presenting a macro view of the DTIP challenges by systematic identification of them. Third, recommending some general criteria to analyze DTIP methods in form of the proposed classifications. Suggesting a suitable set of qualitative criteria along with using quantitative criteria can lead to a more proper choice of DTIP methods. Fourth, performing a two-phase qualitative analysis and comparison between each class of DTIP approaches based on the proposed functional criteria and the identified challenges ahead in order to understand the superiority of each class of DTIP methods over the other class. We believed that the DTIP-TC2A framework can offer a proper context for efficient selection of DTIP methods, improving the efficiency of a DTIP system due to the nature of computational methods, upgrading DTIP methods by removing the barriers, and presenting new directions of research for further studies through systematic identification of DTIP challenges and purposeful evaluation of challenges and methods. Graphical Abstract: ga1 Highlights: We have presented an analytical comparative framework for drug-target interactions prediction (DTIP) methods. We have proposed a technical-oriented classification for DTIP methods. We have classified challenging aspects of drug-target interactions prediction problem in biological networks. Our proposed framework can be useful to select, improve, and compare methods efficiently.