We’re looking at the potency, the selectivity, the off-targeting effects, the pharmacokinetics, the pharmacodynamics, and so on,” Ellis said. “And what we learned as we ratcheted up the parameters is that we were starting to see a significant amount of attrition in the laboratory when we would make molecules.”
Seeing that attrition, the team shifted to a predict-first approach.
“We anchored our understanding of the molecule that we wanted to make in a computational prediction,” Ellis said. “And as we’ve done that, we’ve seen a significant increase in the number of successful molecules that we’ve been able to make and that have advanced to the stage where we’re starting to look at them more deeply in the lab.”
Ultimately, the goal is to have AI guide the entire process from start to finish, from the concept of a molecule to its creation and validation.
“We want to be able to increase the probability of success by having an understanding of what we’re getting into before we actually go into the lab,” Ellis said. “The more we can do that, the fewer molecules we have to make to get to a successful clinical candidate.”
As BMS continues to refine its AI capabilities, Ellis sees a future where the company is able to predict not just the safety and efficacy of molecules, but also other critical parameters like the likelihood of a molecule becoming a medicine that can reach patients.
“We’re trying to get to the point where we can predict all of those things in parallel and not just the safety and efficacy,” Ellis said. “It’s a very exciting time to be in the industry and to be a part of this transformation.”
With AI and machine learning leading the way, the future of drug discovery at BMS and beyond is looking brighter than ever.
In the world of pharmaceutical research and development, the process of creating new molecules with specific properties can be a daunting task. Scientists often find themselves faced with the challenge of balancing multiple parameters that need to be met in order to create an effective compound.
According to a recent interview with a representative from a leading pharmaceutical company, the process of synthesizing molecules can be compared to pulling on a thread – if one parameter is adjusted, it can cause another to break. This complexity can lead to frustration and may result in months-long plateaus in progress.
To overcome this challenge, the company decided to bring in more computational scientists and utilize advanced modeling and prediction tools. By applying these tools to the most important parameters in the composite, they were able to make significant advancements in a matter of weeks.
This shift towards incorporating computational tools into their research process was described as a “cultural evolution” within the company. It not only influenced how molecules were selected for further study in the lab, but also impacted the overall structure of their research teams. The company also embraced the concept of “hybrid intelligence,” which combines human expertise with computational capabilities to enhance the quality of ideas and accelerate decision-making.
The representative expressed pride in the company’s progress and emphasized that this cultural shift towards computational tools is becoming increasingly ingrained in their research practices. They believe that this approach will continue to play a crucial role in their future endeavors.
In conclusion, the integration of computational tools into pharmaceutical research has proven to be a game-changer for this company. By marrying human intelligence with advanced technology, they have been able to overcome obstacles and achieve their goals more efficiently. This commitment to innovation and adaptation is a testament to their dedication to pushing the boundaries of scientific discovery.
