Person-millenia are spent each year seeking useful molecules for medicine, food, agriculture and other uses. For biomolecules, the near infinite universe of possibilities is staggering and humbling. As an example, antibodies, which make up the majority of the top-grossing medicines today, are comprised of 1,100 amino acids chosen from the twenty used by living things. The binding part (variable region) that allows the antibody to bind and recognize pathogens, is about 110 amino acids, giving rise to 10143 possible combinations. There are only about 1080 atoms in the universe, illustrating the intractability of exploring the entire space of possibility. This is just one example…
Presently, machine learning (ML), artificial intelligence (AI), quantum computing, and “big data” are often put forth as the solutions to all problems, particularly by pontificating TED presenters and in Sand Hill pitches dripping with hyperbole. Expecting these methods to provide intelligent prediction of molecular structure and function within our lifetimes is utter bullsh*t. For example, a neural network trained on daily weather patterns in Palo Alto cannot develop an internal model for global weather. In a similar way, finite and reasonable molecular training sets will not magically cause a generalizable model of molecular quantum mechanics to arise within a neural network, no matter how many layers it is endowed with.
With that provocative preface, we turn to the notion of letting matter compute itself. Massive combinatorial libraries can now be intelligently and efficiently mined with appropriate molecular readouts (AKA “the question vector”) at ever-increasing throughputs presently surpassing 1012 unique molecules in a few hours. Once “matter-in-the-loop” exploration is embraced, AI, ML and other methods can be brought to bear usefully in closed-loop methods to follow veins of opportunity in molecular space. Several examples of mining massive molecular spaces will be presented, including drug discovery, digital pathology, and AI-guided continuous-flow chemical synthesis – all real, all working today.
About the Speaker:
Greg Kovacs received a BASc degree in electrical engineering from the University of British Columbia, an MS degree in bioengineering from the University of California, Berkeley, and a PhD in electrical engineering and an MD degree from Stanford University. Greg serves as Chief Technology Officer of SRI International in Menlo Park, California. He is a Professor Emeritus of Electrical Engineering at Stanford University with a courtesy appointment in the Department of Medicine, having joined the faculty in 1991. He has published more than 180 peer-reviewed papers, has written one textbook and has 64 issued patents.
From 2008 – 2010 he was Director of the Microelectronics Technology Office at DARPA. He also has extensive industry experience including co-founding several companies, including Cepheid in Sunnyvale, CA, and Stanford cardiovascular medicine spin-off PhysioWave. In 2003, he served as the Investigation Scientist for the debris team of the Columbia Accident Investigation Board, having worked for the first four months after the accident at the Kennedy Space Center, Florida.
Greg received an NSF Young Investigator Award, held the Noyce Family Chair, and was a Terman and then University Fellow at Stanford. In 2010 he was awarded the Secretary of Defense Medal for Exceptional Public Service. Kovacs is a Fellow of the IEEE, the American Institute for Medical and Biological Engineering and of the IEEE and the Explorer’s Club.
Images are licensed under Creative Commons License.