Identifying the most effective chemical compounds to target specific proteins is critical in drug discovery. Traditional methods often depend on single screening techniques, which can miss important compounds or rank them inaccurately. To overcome this challenge, Said Moshawih, a Master of Data Science scholar at INTI International University, under Professor Dr. Goh Khang Wen, Pro Vice-Chancellor at INTI International University, has developed an innovative consensus scoring method. This approach integrates multiple screening techniques, leveraging machine learning to provide a more accurate and comprehensive evaluation of potential drug candidates.
In his paper “Consensus Holistic Virtual Screening for Drug Discovery: A Novel Machine Learning Model Approach,” published in a highly ranked journal, Said evaluates the efficacy of this consensus holistic virtual screening method across diverse datasets. The study demonstrates that combining various screening methods through a consensus approach leverages the most favourable aspects of multiple screening metrics, leading to more accurate and reliable results.
Said explains, “In this research, we analysed various protein targets, including G protein-coupled receptors (GPCRs), kinases, nuclear proteins, proteases, DNA repair enzymes, and suppressor proteins, focusing on eight targets across these categories to highlight the versatility of the consensus scoring method.”
The study found that consensus scoring improves the accuracy of ranking chemical compounds by their effectiveness. Certain compounds were ranked higher and found to be more active than results obtained using a single method. Key measures, such as the success rate and accuracy of identifying active compounds, were similar to or better than traditional methods, especially for a protein called CDK2.
Overall, the combined method, powered by machine learning, is useful across various proteins, making it a versatile tool in drug research. It ranks compounds more accurately, helping researchers concentrate on the most promising candidates. While some proteins may require tailored approaches, the consensus method can be optimised for high performance. This comprehensive analysis underscores the importance of considering quantitative metrics and prioritising active compounds, which can vary significantly across different methods.
Said stated, “The analysis emphasises the effectiveness of consensus scoring as a crucial virtual screening technique, often yielding superior performance in early detection of actives and prioritising compounds with the highest biological activities.”
He concluded, “These findings significantly advance our understanding of screening techniques’ performance in diverse protein target contexts, ultimately enhancing the effectiveness of virtual screening approaches. These discoveries suggest that consensus scoring is a powerful tool in drug discovery, improving the chances of finding effective new drugs.”
Said Moshawih, a Master of Data Science scholar at INTI International University, has developed a novel consensus scoring approach that integrates multiple traditional screening methods through machine learning. This approach significantly enhances the identification of potential drug candidates in virtual screening applications.