Rui Loureiro of Hovione shares some results of development using the PRIME tool *
Sustainable API development guided by historical and AI evaluation
Rui Loureiro of Hovione shares some results of development using the PRIME tool *
Synthetic molecules need to be developed faster and at a lower cost than biopharmaceutical drugs . This increases the pressure on chemists and engineers to develop faster and more sustainable processes , which requires that they follow Green Chemistry principles .
One key societal pressure is to expedite new drug development . Traditionally , API chemical process development takes , on average , six to eight years of the total five to 15 that is takes to bring a drug from early discovery to the commercial stage . However , the industry is being pushed to cut this in half . 1
Chemical processes vary in complexity and involve many unit operations and variables that need to be considered . Once a process is developed at laboratory scale , it must be scaled up and transferred to production , introducing further complexity and variables .
To accelerate the pace , there is a growing emphasis on leveraging technologies to enhance experimentation and data management . This includes using automated systems for experimentation , electronic notebooks to enable digitalisation , and Artificial Intelligence ( AI ) and machine learning ( ML ) algorithms to expedite process development . Furthermore , digital twins are emerging as a valuable tool to mitigate risk and expedite scale-up processes . 2
The rush towards rapid deployment must be balanced with the development and optimisation of
Figure 1 – Process metrics at different phases
robust and sustainable processes . This requirement is a challenge for CDMOs , yet also presents an opportunity to capitalise on the historical knowledge gathered from the different drugs that they manufacture over the years . Using this historical data and leveraging AI , ML and other tools , they can identify and implement previous learnings , thereby accelerating process development .
Use of historical data
Over the years , there has been a remarkable evolution in the accessibility and dissemination of pharmaceutical data . A wide range of tools is now available to streamline data management and analysis in the industry .
Commercial databases offer vast repositories of chemical and biological information , facilitating literature searches and compound screening . Electronic laboratory
notebooks ( ELNs ) provide digital platforms for documenting experimental procedures , results and observations , enhancing reproducibility , transfer and traceability . 3
Additionally , computational software suites enable molecular modelling , virtual screening and predictive analytics , empowering researchers to expedite drug discovery and optimisation processes . By integrating historical data with advanced computational tools , researchers can uncover hidden insights and identify promising drug candidates .
AI combined with ML holds immense potential for accelerating drug development from route scouting to process development . AI algorithms can analyse vast chemical databases , predict reaction outcomes and propose optimal synthetic routes with high efficiency and accuracy . The routes proposed need to be
22 SPECIALITY CHEMICALS MAGAZINE ESTABLISHED 1981