A comparison study of artificial intelligense-driven no-code appications for drug discovery and development

Abstract

The aim. The aim of this study was to evaluate the functionality and effectiveness of selected AI-driven no-code applications in drug discovery. This research assessed ease of use, interface design, user experience, speed, resource utilisation, accuracy, and scalability to determine their suitability for various drug development tasks. Materials and methods. The study used an evaluation methodology to test six AI-driven no-code applications: Insilico Medicine’s Pharma.AI, Atomwise, Schrödinger’s LiveDesign, Exscientia, BenevolentAI, and Cyclica. Quantitative data were collected from performance metrics, and qualitative data were obtained through expert interviews. Data analysis was conducted using descriptive statistics, repeated measures ANOVA, and post hoc Tukey’s Honestly Significant Difference (HSD) tests. Results. The analysis revealed that Insilico Medicine’s Pharma.AI and Atomwise consistently outperformed other applications regarding usability and predictive accuracy. Schrödinger’s LiveDesign demonstrated high accuracy but required significant computational resources. BenevolentAI and Exscientia showed limitations in usability and accuracy, particularly in toxicity prediction. Cyclica was noted for its ease of use but was less effective in scalability and resource utilisation.

Conclusions. The findings provide valuable insights for researchers and pharmaceutical companies, guiding the

integration and application of AI-driven solutions to accelerate the drug discovery process and improve the suc- cess rate of developing new therapeutic drugs. Future research should focus on broadening the evaluation to

include more diverse scenarios and real-world applications to further validate and enhance these tools.