March 19, 2024
Dr. Patrick Finneran's recent appearance on Software Engineering Daily marks Menten AIs significant commitment to the scientific community. Dr. Finneran, our associate director, engaged in a discussion with Sean Falconer, exploring the convergence of our platform with software engineering, artificial intelligence (AI), and biotechnology.
In the beginning, Dr. Finneran expertly outlined the complexities of preclinical research and showed how Menten AI is at the forefront of solving the varied problems in the field. He emphasized the four pillars crucial to any preclinical work: pharmacology, pharmacokinetics, toxicology, and chemistry manufacturing controls (CMC), and stated that "each of these 4 properties matter and it is a complicated loss function that, [at Menten AI], we are actually looking at, because [for example] to improve the pharmacology you don’t want to take a hit on the pharmacokinetics. It creates fun multivariable problems that we love to solve in this space!". Dr. Finneran articulated how Menten AI tackles multivariable design problems while also accelerating the discovery process: "what's going on nowadays is this super exciting realm of using computer science, and using smart experimental design [to] limit our search space… [Menten AI's platform enables testing of] only 50-100 compounds in the lab , and that's the acceleration that we are seeing today." Finally, he highlighted that Menten AI focuses on innovation by targeting hard-to-drug molecules, demonstrating our commitment to innovation and addressing unmet medical needs - after all "you don’t go after these [targets] with what everyone else is doing."
Moreover, the conversation delved into the realm of peptides, highlighting their advantages over traditional drug modalities. Dr. Finneran acknowledged the challenges in peptide discovery but underscored Menten AI's capabilities in overcoming them, for example in enhancing solubility and permeability. Sean and Patrick agreed that no platform can solve all the problems in drug discovery perfectly because we don't have enough data about humans from the start - "sometimes drugs fail in Phase II trials, and we don't know why." This missing puzzle piece is crucial for the future of drug discovery. Luckily, Menten AI stays ahead of these challenges by working closely with a large network of CROs. For example, some CROs create organ-on-a-chip systems to mimic human conditions, making testing more realistic and allow Menten AI to tackle these problems early on in the discovery process.
As the dialogue progressed, parallels were drawn between Menten AI's platform and general software engineering principles. It became evident that our approach, while innovative, requires continual refinement and collaboration with experts across disciplines. After all, utilizing Generative AI to design cyclic peptides is a niche industry and that means we have to customize our AI and machine learning (ML) models to not only better suit our strategy in drug design, but also for individual partnerships and unique projects. The host, Sean, brought up an exciting analogy and compared Menten AI to Google Brain's AlphaGo Program: "the [Alpha]Go Program was doing moves that no humans would ever think of doing and that actually allowed human go players to get better because they could see the movement and try to reverse engineer it." Patrick completely agreed with this analogy and suggests that "no machine will ever tell you why it chose to [design a specific peptide sequence]", however it's essential for scientists to comprehend the reasoning behind the AI generated output and understand how we can leverage the output data to enhance subsequent iterations. In harnessing Generative AI for cyclic peptide design, Menten AI exemplifies its dedication to cutting-edge technology.
In essence, Menten AI stands as a leader in the peptide discovery field, driven by innovation, collaboration, and an ongoing commitment to advancing drug discovery for the betterment of patients worldwide. Through our multidimensional approach and unwavering pursuit of excellence, we continue to redefine possibilities in biotechnology and beyond.
Watch the full video podcast on YouTube or below:
Software Engineering Daily (SED) is a technology podcast dedicated to uncovering the ways in which softwareis changing the world. Sean Falconer, who hosted Dr. Finneran on the podcast, is an academic, startup founder, and Googler who has published works covering a wide range of topics from information visualization to quantum computing.
Subscribe to Software Engineering Daily's YouTube Channel to gain more insights into emerging field experts and innovative industries shaping software development!
