Looking forward means playing the long game in biopharma. In a decade’s time, it will all be about knowing every protein the AI way, something which will have transformed the industry, believes Viren Mehta, founding partner of Mehta Partners LLC. Looking forward means playing the long game in biopharma. In a decade’s time, it will all be about knowing every protein the AI way, something which will have transformed the industry, believes Viren Mehta, founding partner of Mehta Partners LLC. Looking forward means playing the long game in biopharma. In a decade’s time, it will all be about knowing every protein the AI way, something which will have transformed the industry, believes Viren Mehta, founding partner of Mehta Partners LLC.

While a great deal of ink is being spilled in crystal ball gazing about what the world holds for the New Year, the biopharma industry calls for thoughtful insights a decade out. We know that is how long it takes to achieve success in discovering, developing, and dominating with an innovative therapy. So, it is not a typo in the title above. Year-end reflection and the accelerating pace of scientific progress offer a timely opportunity to appreciate the increasingly exciting biopharma future. Information technology, especially its artificial intelligence (AI) frontier, and life science R&D are becoming one to address some of the most fundamental healthcare challenges.

It is now conceivable that the impact of life science innovations on the quality of our lives over the next generation will be greater than the impact IT has had over the past couple of generations. Admittedly I am going out on a limb with such a bold prediction, but positive results over the coming decade would make this outcome increasingly probable. We are talking about making our exquisitely choreographed dance of life healthier.

Tipping Point

Fortunately for the biopharma industry, not only is the pace of life sciences innovation at a tipping point, the resulting bounty should provide for a recalibration of management ethics amidst increasing self-confidence that even a high-risk biopharma innovation enterprise can sustain itself just as well as other similarly challenging R&D-driven industries. In the same way as a new mobile phone, edge computing or reuseable space rockets must live or die on their merits in the market place, so would truly cost-effective biopharma therapies. There should be no more need to hide behind the regulatory veil to push market forces to the hilt, ethics be damned.

Once biopharma R&D libraries include 3D structures of most of the essential proteins, predictable and increasingly curative therapeutic innovations should become practically a mechanical exercise

One clear indication of this promise is the pace of new drug approvals. In the US, after decades of regulatory struggle, where a typical year yielded only a couple of dozen new drug approvals, 2017 brought 45 new drug approvals, and 2018 is on pace to achieve close to 60 innovative biopharma products reaching the market. (Also see “Has R&D Productivity Rebounded? US FDA Officials Think So” – Pink Sheet, 13 Dec, 2018.) There is no reason why this pace should not continue, if not improve further.

Peering into the future we can now be comfortable about this growing new product flow. This week’s publication in Science of two phenomenal successes from Alphabet/Google’s DeepMind unit shows the awe-inspiring power of a fully integrated AI and biopharma R&D. The first breakthrough is the AI self-teaching game player AlphaZero starting only with the rules of the game, and the other is DeepMind’s major win at an international competition to predict the three-dimensional folding structure of proteins simply from their genetic sequence. Perhaps just a first step, but a fundamental milestone on the quickening journey to realize the promised transformation of R&D when combined with AI and all the related tools.

Biopharma’s AI Revolution

Basic science is nearing a critical mass, delineating cellular pathways and molecular mechanisms, and enabling elucidation of an optimal therapeutic agent. This precision in turn readily offers a companion diagnostic to identify and treat just the patient population most likely to benefit. Even more crucially, these tools ensure that an increasing number of therapies will not only be curative, but achieve these miracles with fewer and fewer side-effects.

Soon, biopharma R&D efforts anchored around trial and error, with companies taking pride in their libraries of millions of molecules to be tested against poorly understood assays, will be relegated to the history books.

How proteins, the workhorse molecules of life as they are sometimes described, fold their long chains of amino-acid building blocks into a compact 3D shape remains one of the most important unsolved problems in biology, because knowing a protein’s shape is key to understanding its function, and to designing a curative therapy when this protein malfunctions.

AI neural networks obviate the need for thousands of rules formulated by fallible human beings. Predicting and defining the 3D structure of every protein is now only a matter of time. Just as reusable rockets are about to transform space exploration, self-learning AI algorithms bring dynamic and unconventionally open style to rapidly decipher what had previously been done through costly and time-consuming crystallography and electron-microscopy 3D structure exercises that at best yield partial results with modest predictability. AI neural networks, in addition to defining 3D structures, can assist nurses and doctors, give confidence to patients to take charge of their health, and are already perfecting speech recognition that is simplifying countless tasks in our daily lives.

Once biopharma R&D libraries include 3D structures of most of the essential proteins, predictable and increasingly curative therapeutic innovations should become practically a mechanical exercise, just as the antisense and mRNA based drug discovery efforts of the past 25 years have now matured to offer a template for drug discovery with increasing success rates.

Back To The (Small Molecule) Future?

The real prize of this detailed understanding
of proteins could be the return to small molecule therapies, coming full circle
back to orally active drugs. The biologics with their durable and targeted
binding will remain a part of a clinician’s armamentarium, but revival of small
molecule innovations will play a valuable part in rebalancing biopharma costs
and fostering a more egalitarian healthcare offering globally.

Such a turn of events would finally reposition the industry to its rightful ranking. No more would pharma struggle on par with the used car salesman to earn society’s respect. Most biopharma companies would proudly launch quite a handful of such effective, if not curative drugs each year, enabling them to earn an honest living by pricing their bountiful flow such that every world citizen who may benefit would have access.

And isn’t access the real obligation that our industry must fulfill to continue to deserve its patent-protected period of profitable market position?

This column originally appeared on Scrip Biopharma Intelligence, December 17th, 2018