Protecting IP in bioinformatics and digital health: tips and tricks to tackle the “moving target problem”

As previously discussed, data driven approaches have revolutionised research and R&D in the life sciences sector. As a result, life science IP professionals increasingly have to support projects that involve a large computational component. These generate inventions that they may be less familiar with, and crucially these projects often move at a very different pace from the more traditional molecular biology / biochemistry wet lab driven projects. Indeed, computational projects can generate valuable insights at a relatively fast pace, can result in the development of software tools that are updated every few months, and can even be designed to change after release (i.e. through their use - for example, some machine learning / AI powered tools are designed to continuously learn, resulting in changes in the model that underlies the predictions made by the tool). Altogether, these factors mean that the protection of IP generated by these projects can feel like a “moving target”, raising questions such as what can be protected, what is worth protecting, and when is the right time to “fix” an invention in a patent application. In this post, we provide general tips and tricks to tackle this “moving target problem” using the patent system primarily, based on our experience helping clients handle this subject-matter.

What can be protected?

Bioinformatics projects will often result in the creation of various types of potentially valuable assets including data (whether raw or organised as a database), computer implemented methods / tools and new insights generated by the methods (e.g. biomarkers, drug candidate, patient stratification criteria, etc.) Broadly speaking, the latter two categories (computer implemented methods and insights that have direct technical meaning) can be protected by patents in many jurisdictions. These include Europe and the USA, contrary to widespread misconceptions that software is not (or no longer) patentable. Europe in particular is a comparatively favourable jurisdiction for both of these types of IP (as discussed in our blogs: EPO Bioinformatics Roads, Simulations in bioinformatics and A Guide to Patenting Bioinformatics Inventions at the EPO: what is technical and why does it matter?). The subtleties of how to obtain patent protection for computer implemented methods and their resulting insights in the bioinformatics field are beyond the scope of this post. However, it is well worth remembering two things in relation to what can be protected:

1. Patent protection for computer implemented bioinformatics / medical informatics inventions (i.e. what you might call software, algorithms, bioinformatics pipelines, computational methods, etc.) and for their insights (e.g. biomarkers, drug candidate, patient stratification criteria, etc.) is available.
   
   
2. These inventions often touch on aspects that are considered “excluded subject-matter” in many jurisdictions (mental activities, laws of nature, mathematical methods, etc.), each jurisdiction relying on their own approach to assess these inventions. This is by no means a road block at least in Europe and the USA, but it does mean that care has to be taken (and, crucially, the right expertise has to be recruited to the task) when drafting and prosecuting patent applications in this field. It is worth noting that the approach may differ in the fields of bioinformatics / digital health and in the fields of computer software / machine learning when applied to robotics, automotive and other engineering fields. If at all possible, make sure that you are advised by someone who specialises on bioinformatics / digital health.

When should patent protection be sought?

Broadly speaking, the question of timing depends on multiple factors that are not dissimilar to those that apply for any type of inventions. However, in a field that moves quickly, these are perhaps more crucial to ask explicitly, and to re-evaluate regularly. Useful questions to ask include:

1. What scope of protection could I hope to obtain now – i.e. based on the evidence I currently have, and the prior art that I am aware of, what scope of claim can I expect?
   
   
2. What scope of claim can I hope to obtain in e.g. 6 months, 1 year, 2 years – i.e. what are our R&D plans?
   
   
3. Are our R&D plans likely to focus, expand or divert the scope of the technology? For example, if we are planning to work on refinements of a general concept, the general concept being likely to be patentable, then patenting the general concept now, and the refinements later could be a very good way to build a strong patent portfolio. As another example, if we are likely to expand the field of application of our computational tools from e.g. cancer to e.g. fertility treatment, then each field of application could be made the object of a separate patent portfolio (although in such cases the strength of subsequent applications in terms of patentability over our own prior art should be carefully considered). These could then be managed and exploited separately, with obvious commercial advantages. Similarly, when R&D plans are likely to divert the scope to e.g. the development of entirely new tools, separate patent protection for the current and planned scope can be a sensible strategy. This is potentially true even if one line of development will be discontinued, for example if there are licensing or selling opportunities for the IP related to this line of development.
   
   
4. Do commercial / research imperatives have an impact on our timeline? In most cases the answer to this question will be ‘yes’. The follow up question will then be: how long do we have until something crucial has to be disclosed, and how much more information will we have if we wait e.g. 1, 6, 12 months before filing a patent application?

What should be protected?

As discussed above, projects in this field can often generate two types of patent-eligible aspects: computational methods, and insights. Both of these should in general be discussed, and could be made the object of separate protection strategies.

Insights (such as biomarkers, drug candidates etc.) are “traditional” prime candidates for patent protection. They are likely to be familiar to life science IP professionals and patent offices alike, generate IP that is comparatively easy to enforce, and are in general not usefully protected by trade secrets since their commercial exploitation often inherently discloses the invention. 

When it comes to software tools, it is often worth having the discussion of whether and to what extent some things should be kept as trade secrets and others should be patent protected. This discussion would typically involve the following questions: 

1. What patent claim scope could we expect to obtain (in view of the subject-matter, the prior art, the evidence at hand, etc.)? Would this claim scope:
   
   • cover planned products / services at least in the short term; it is worth noting that even with software tools that evolve by design, it is often possible to obtain a scope of protection that is broad enough to cover the product as it evolves (for example, because the protection relates to how the software analyses data to produce an insight, and does not include limitations that would tie it to a fixed training data set);
   • be commercially useful for another reason (e.g. attract investment, represent a possible road block for competitors to develop their own products/services, provides credibility to the product/service, have publicity value, provides a solid proof of background IP before entering discussions with a partner, etc.);
   • be easy to work around (e.g. because it is limited to a particular set of parameters that are not in fact crucial to obtain a commercially useful product); and
   • be enforceable (e.g. would we be able to identify infringers with reasonable effort).
2. What could we expect to keep secret? In many academic projects the only scientifically ethical answer to this is “nothing of relevance”. Similar considerations are likely to apply to software tools that are regulated, such as e.g. medical devices. In such cases, a careful look at patentable aspects makes even more sense since the patent system is designed to provide a competitive advantage to applicants in exchange for an enabling disclosure (in the patent, not prior to it!)
   
   
3. What is the short term and long term commercial strategy? In particular, what will be commercialised in the near future, and in the long term? How long am I expecting to use the technology and what do I expect my competitors to do in that time? Who will commercialise what, and who will own the IP for what? For example, for a company that has developed a new tool for computational drug design, protecting the tool as core company IP and selling/licensing the rights to drug candidates to partners may be a sensible strategy to balance the interests of all parties.
   
   
4. What is the short term and long term R&D strategy? In particular, where will the development work focus, what solutions will be developed for internal and/or external use? 

All of these points essentially aim to answer two questions: what patent scope is realistic, and what would be the value of this (i.e. how would it support short and/or long term commercial plans)?

In the software world, competitors and competing products can appear very quickly, and products can become obsolete just as quickly. This can lead to a feeling that the patent system is just too slow to provide any competitive advantage, and too expensive to provide value for money in this context. This does not have to be true, if a patenting strategy is developed and regularly updated with commercial and R&D plans in mind, with the ultimate goal of optimising value in view of these goals. This means applying for patent protection whenever the scope of protection that can reasonably be expected supports short and/or long term commercial plans, and being ruthless about letting go of rights when they no longer provide value. As an example, if a patent (or even a pending application) provides a competitive advantage for 2 years, after which the product has changed enough that the patent/application is no longer useful, then the right has done its job and can be dropped to focus resources on the next wave of innovation. Aging portfolios are expensive to maintain, so looking at the glass half full it’s no bad thing if it makes commercial sense to only protect something for a few years then move on! Finally, it is worth bearing in mind that some patent systems are quicker than others, that acceleration strategies exist in many systems if quick grant is of importance, and that a patent application can have value even before if grants.

In conclusion, it is more important here than ever to have frequent conversations about IP. These should discuss the current state of technology development but also the bigger picture, to understand both the potential for protection and what, for a particular venture, has value now and will have value in the future.