The development of a commercializable software prototype and technology that can be built into artificial intelligence workflows to create novel and patentable scaffolds and vast virtual compound libraries with no human intervention
The extremely costly and time-consuming drug discovery workflow employs high throughput (a lot of experiments or data generation in very short time) technologies in almost every part of the R&D phase. The only area that remains a very manual process is the design of the new molecules or molecule libraries for synthesis to tackle biological problems or to improve the composition of compound libraries used for screening. To date the idea generation capability of the human brain has not been replaceable by computational tools.
In the last 5 years the use of artificial intelligence has made a huge progress in all industrial segments and it is an emerging technology in the pharmaceutical industry as well. The multidimensional lead optimization can only partially and incompletely be supported by manual or partially automated data mining in huge databases which leads to loss of the information content. To address these issues research has turned to AI in the pharmaceutical industry in order to enable the computer using all available valuable data to judge the quality of the lead molecule series and to decide the next best candidates for synthesis in the drug development program. Synthetic creativity and the knowledge of design of a novel structure is missing from the capabilities of an AI therefore AI-powered drug discovery needs large libraries of novel and patentable molecule structures from which it can select after evaluating and ranking the next molecules recommended for synthesis.
Synthetic feasibility is critical because rapid synthesis of the selected molecules is obviously critical to reduce cycle time. The novel synthetic know-how database and the associated design intelligence recently developed by ChemPass enable us to achieve the development of an automated design engine for truly synthesizable novel scaffolds and libraries to support the decision making of AI-powered drug discovery. Thus, the goal of the proposed project is the development of a commercializable software prototype that can be built into artificial intelligence workflows to create novel and patentable scaffolds and vast virtual compound libraries with no human intervention.