How a computer cheats the patent system of “Big Pharma»
The algorithm allowing to find ways of synthesis of pharmacological preparations bypassing patent restrictions is developed. The same program will help to formulate patent applications so as to make them invulnerable
In modern pharmacology, intellectual property is the most important asset, and its protection is a separate industry where a lot of know-how is accumulated. Starting work on the Chematika project, researchers from Poland, led by Bartosz Grzybowski, who is now working in South Korea, did not expect that they developed a computer algorithm will be able to revolutionize this area of knowledge. Nevertheless, they seem to have succeeded: the program easily finds strategies for the synthesis of complex chemical compounds that bypass the restrictions imposed by the relevant patents. “It turned out that there is still a lot of loopholes – says Grzybowski. — We were able to find new retrosynthetic ways around patents.”
Pharmacological patents protect intellectual property by preventing competitors from using key technological solutions. These solutions are designed to optimize the output of the target compound, ensure maximum purity and reduce production costs. A patent usually formulated so as to avoid detours, which, although less efficient, still could attract competitors that save them from having to acquire a license.
However, Polish researchers tried to find such ways. They have “frozen” key elements of the chemical compound, prompting the program to offer, albeit less effective, but still possible from the point of view of the chemical bypass. The program was tested on three commercial drugs. Linezolid is an antibiotic from the oxazolidinone group — the so-called last resort antibiotic used against resistant infections. Sitagliptin is prescribed for diabetes, and panobinostat — anticancer therapy.
In each of the three cases, when a computer program was allowed to choose the path of synthesis at its own discretion, it offered exactly the technology that is patented by the manufacturer. But then the authors of the algorithm fixed some atoms and chemical bonds, declaring them “untouchable”. In this case, the program found an alternative way to assemble the required molecule from other components. It was found that in all three cases considered, even when the algorithm was forbidden to touch all parts of the molecule protected by a patent, it was able to offer feasible chemical solutions based on alternative source materials — sometimes unexpected, but quite acceptable from an economic and technological point of view.
Chematika’s algorithm can have a major impact on patent practice: a machine search for alternatives will make it possible to formulate a patent application in such a way as to completely eliminate the possibility of circumvention, thus making the patent invulnerable to competitors.
Bartosz Grzybowski nevertheless notes that this invulnerability is not absolute: the development of chemistry can lead to the fact that the ways of synthesis, which the program considers unacceptable today, will be proposed. Thus, the described synthesisdoes not limit healthy technological competition, driving progress in the field of pharmacology. Moreover, Polish scientists believe that their program, closing the obvious workarounds, can push chemists to more actively search for new opportunities and approaches to organic synthesis. Another participant in the study, mathematician Peter Dittwald, draws attention to how the cooperation of chemists and programmers brings the creation of “chemical artificial intelligence”, which can be used not only in academic research, but also in industry.