Navigating the complex landscape of drug discovery requires innovative strategies and interdisciplinary collaborations spanning biotechnology, pharmacology, medicine and engineering.

As the cost of developing new drugs surges, the pharmaceutical industry is increasingly driven to explore the potential of artificial intelligence and machine learning technologies to reshape research, expedite timelines and curtail costs.

Yet, amid this promising innovation lie potential intellectual property pitfalls due to the industry's robust reliance on proprietary drug revenue.

Use of Artificial Intelligence/Machine Learning in Drug Discovery

Escalating research and development costs, predicted to rise from approximately $1 billion to over $2 billion, are spurring interest in AI and machine learning technologies as transformative tools to streamline drug discovery, enhance patient outcomes and drive down costs.

Some key AI and machine learning applications in this sphere include predictive modeling, image analysis and pattern recognition, virtual screening and data analysis, personalized medicine, and ability to gain new insights.

Predictive Modeling

AI and machine learning can be used to build predictive models to facilitate the identification of promising drug candidates for targeted diseases, prioritizing molecules and compounds for further investigation.

Image Analysis and Pattern Recognition

AI and machine learning can be used to analyze images of cells and tissues to identify patterns and features, providing insights into disease interactions and drug efficacy.

Virtual Screening and Data Analysis

AI and machine learning can be used to analyze large and complex data sets to inform potential uses for existing drugs in treating other diseases or pinpoint diseases that may be effectively treated with novel drugs.

Personalized Medicine

Similar to traditional data analysis, AI and machine learning can analyze patient data on a much larger scale to identify specific patient populations that may benefit from a particular drug. It can also be used to reduce medical errors, detect diseases earlier, provide personalized treatment plans, improve self-care and reduce medical costs, among other benefits to improve patient outcomes.

Gaining New Insights

AI and machine learning can be used to perform various downstream biomedical tasks, generating invaluable insights for drug discovery and research. For example, Microsoft Corp.'s BioGPT tool — trained on a vast pool of biomedical research — can answer queries, classify documents, extract data and more.

Understanding the IP Toolbox for Drug Discovery Companies and How It Intersects With AI and Machine Learning

IP protection serves as a vital buffer for drug discovery companies, safeguarding investments and technology while fueling a competitive edge. Below are key IP tools for drug discovery:

Data Protection

Data serves as a pivotal asset in drug discovery, providing exclusive access to unique insights that can expedite drug discovery and development, bolstering a company's competitive edge.

Accordingly, data can serve as one of the most important, if not the most important asset, a drug discovery company has. Having proprietary data allows companies to perform nuanced analysis, identifying promising drug candidates faster and with higher precision, thereby reducing costs and accelerating time-to-market.

Data exclusivity also carries significant weight. It offers companies a defined period of time to utilize and profit from their data without interference or obligations to others. Defined data protection strategies that give the drug discovery company control over the use and access of its proprietary data ensures that this valuable asset remains secure, maintaining its value and the company's competitive advantage.

Patents

Patents are critical for drug discovery companies.

They can cover new drugs, formulations, compounds, molecules and methods of treatment. They can also protect innovations in data analysis and visualization techniques using AI and machine learning training, applications of AI and machine learning models, and processes for various aspects of drug development, design, and target identification.

Patents further provide companies an exclusive right to exclude others from making, using or selling their innovations for a limited period, typically 20 years from the patent application filing date. Patent filings related to AI and machine learning in drug discovery have been steadily growing.

This trend suggests that patents are likely to play an increasingly important role in not only the development and monetization of new drugs in the future, but also for collaborations and negotiations between drug discovery companies and large pharmaceutical companies.

Trade Secrets

Not all types of innovation are suitable for patenting. In addition to patents, trade secrets can protect valuable information generally not known to the public that provides a competitive advantage or economic benefit from being secret.

In the context of AI and machine learning and drug discovery, such information can include training data for AI and machine learning models, software code, data analysis processes, and other types of confidential information related to drug development or analysis processes. Companies relying on trade secrets can consider having robust protection protocols in place that are periodically reviewed and reinforced to maintain secrecy.

However, there is no recourse if a competitor independently develops the same technology.

Copyrights

Drug discovery companies can use copyrights to protect original works, such as scientific publications, marketing materials, training materials and software code — thus preventing unauthorized use or replication and ensuring exclusive rights to profit from these works.

Trademarks

Trademarks can play an important role in a drug discovery company's success, even though their direct impact on the science of drug discovery might not be obvious.

For instance, trademarks establish a unique identity for a company's product in a crowded marketplace and can be used to establish trust and build goodwill, assuring customers about the efficacy, safety and reliability of any new drugs developed using a company's trademarked drug discovery platform.

Licensing Agreements

Drug discovery companies may enter into IP-related agreements with other parties. These include joint development agreements, data sharing licenses, patent license agreements, technology license agreements, among others. Such agreements may allow drug discovery companies to generate revenue and collaborate with other companies or researchers while protecting their own IP.

IP Considerations in Data Sharing Between Drug Discovery Companies and Large Pharmaceutical Companies

AI and machine learning models require large data sets that may be owned, developed and/or shared by multiple entities. Two example scenarios may be:

Scenario 1: A drug discovery company receives a list of compounds or molecules from a large pharmaceutical company to identify other diseases that those compounds or molecules may be used to effectively treat.

Scenario 2: A drug discovery company receives a list of genes, classifiers, biomarkers, or other indicators associated with a particular disease to identify new compounds or molecules that may be leveraged for treatment.

In both scenarios, a drug discovery company needs to be able to safeguard — via appropriate licensing agreements — its significant investment in developing, maintaining and using AI and machine learning models. They also want to maintain a competitive advantage, prevent misuse, negotiate favorable monetary/royalty benchmarks, and ensure compliance with regulatory requirements.

Some issues to consider when drafting such an agreement include data ownership, use, protection, exclusivity, and quality, regulatory compliance and IP.

Data Ownership

Clear definitions of who owns what are highly recommended in the agreement.

Ownership considerations may be complicated if third-party data is involved, if data from multiple sources is combined to create new data sets, or if raw data is pre-processed or post-processed.

For AI and machine learning applications, companies can define boundaries on who owns outputs and insights from the model, as well as any training data utilized. In general, data ownership rights can consider the current data owners, the types of data, the sources of data and how data is used.

Data Use and Protection

Like data ownership, companies can consider defining the purpose and scope of data use in the agreement to avoid misuse or unintended consequences.

Data use clauses can define how data can be used, for what purpose and by whom. Data protection is an important aspect for data sharing. Parties can have policies, protocols and protections in place to safeguard data, including data encryption, secure and authenticated storage, data back-up and replication, and limited access.

A clear understanding of how the data is kept confidential, including how it is shared, stored and protected from unauthorized access or disclosure is desirable.

Data Exclusivity

A drug discovery company may desire exclusive access to the data for a period of time, allowing them to use the data for research purposes without interference from, or obligations to, other parties.

As AI and machine learning models are updated or new models are generated, in addition to data ownership, use, and protection issues, the agreement can clearly define data exclusivity clauses and have provisions to prevent the circumvention of the data exclusivity clauses.

Data Quality

The quality and format of the data received — and the data provided — between the drug discovery company and the pharmaceutical partner can be negotiated and agreed upon in advance to ensure suitability of the data for the intended purpose.

Regulatory Compliance

Data sharing may also be subject to regulatory requirements, such as data privacy regulations, U.S. Food and Drug Administration regulations, and the Health Insurance Portability and Accountability Act. Compliance with these requirements can be carefully considered and addressed in the agreement.

Patents

The agreement can clearly identify any patents that may be affected by data sharing and their ownership.

Necessary licenses or permissions must be obtained to avoid unintended patent infringement issues. The drug discovery company should have a clear understanding of the patents they own and the rights they have to license them to a pharmaceutical partner, as well as the legal risks and exposure associated with infringement of patent rights.

Joint IP

The drug discovery company and pharmaceutical partner can negotiate patent ownership, cost sharing, litigation and legal liability associated with patenting new technology developed as a result of data sharing or joint development.

For example, who decides what to patent, who controls prosecution, who bears the cost of prosecution, and who controls litigation resulting from the patents can all be clearly specified in the agreement. This can avoid surprises and confusion where a party decides to file a patent individually without the other party's knowledge.

Three Key Issues to Consider

Finally, licensing agreements between drug discovery companies and large pharmaceutical companies involve a range of intellectual property considerations that need to be carefully considered, understood, and addressed to protect the interests of both parties.

Here are the top 3 issues that drug discovery companies can keep in mind:

1. Clearly define ownership, use, exclusivity and protection of AI and machine learning models and the data used.

2. Preserve monetary or royalty rights from any downstream updates to, or development of new, AI and machine learning models, as well as freedom to use AI and machine learning models.

3. Carefully consider implications of IP jointly developed.

As drug discovery companies embark on collaborations with large pharmaceutical companies, astute IP management has never been more crucial.

The accelerating integration of AI and machine learning technologies in drug discovery presents both unprecedented opportunities for innovation and unique challenges to intellectual property rights. As the value of data rises, these companies need to pay scrupulous attention to their IP assets to ensure their rights are adequately protected, their proprietary technologies are secured, and their competitive edge is maintained.

A harmonious balance between open collaboration and strategic protection of IP assets is vital. Companies should foster relationships underpinned by transparent and mutually beneficial IP licensing agreements, ensuring that ownership, use and exclusivity of data and AI and machine learning models are clearly defined.

Such measures not only safeguard the commercial viability of their innovations but also allow them to fully capitalize on any downstream developments. Furthermore, the implications of jointly developed IP should be carefully evaluated to ensure both parties share equitably in the benefits and responsibilities that arise from their collaborative efforts.

With AI and machine learning technologies poised to redefine the drug discovery landscape, companies that can strategically manage and protect their IP assets while fostering productive collaborations stand to gain the most.

The delicate dance between collaboration and competition in the era of AI-powered drug discovery demands a new kind of vigilance around IP rights. As such, the role of well-informed, strategically implemented IP management cannot be overstated.