04 November 2010

Technical Board of Appeal rejects Computer-implemented Invention in the field of Bioinformatics (T 784/06)

I just came across the very interesting decision T 784/06 of an EPO Technical Board of Appeal, relating to the rather new scientific discipline of  bioinformatics, i.e. the application of computing and mainly statistical methods to DNA sequencing, e.g within the framework of the Human Genome Project.

Not only that this decision presumably is the first bioinformatics case dealt with by a Technical Board of Appeal, it also demonstrates the EPO approach towards computer-implemented inventions (although with some "twists"), as presented in T 641/00 (Comvik, 1992) and recently confirmed in G 3/08 by the Enlarged Board of Appeals (see also here and here). Some supplementary information on patenting of bioinformatics inventions can be found on Phamalicensing.com.

Originally filed and granted Claim 1 of EP 0 736 107 relates to a method of determining the genotype at a locus within genetic material obtained from a biological sample, the method comprising:
  1. reacting the material at the locus to produce a first reaction value indicative of the presence of a given allele at the locus;
  2. forming a data set including the first reaction value;
  3. establishing a distribution set of probability distributions, including at least one distribution, associating hypothetical reaction values with corresponding probabilities for each genotype of interest at the locus;
  4. applying the first reaction value to each pertinent probability distribution to determine a measure of the conditional probability of each genotype of interest at the locus; and
  5. determining the genotype based on the data obtained from step (D).
An essentially parallel device claim 22 was directed to device for determining the genotype at a locus within genetic material obtained from a subject executing the method according to claim 1. 

Technical features: The proprietor (Beckman Coulter, Inc.) held that all steps A to E are technical, while the opponent (Roche Diagnostics GmbH) argued that steps B to E relate to pure mathematical operations and thus have to be considered non-technical.

The Board came to the conclusion that claim 1 represents a five-step method (steps B to E) for determining the genotype at a locus within genetic material, operating on the data acquired in leading reaction step A to produce a first reaction value. Each of steps B to E was assessed as to only relate to mental activities, so that the claim represents a mixture of technical (step A) and non-technical features (steps B to E).  

Patentability: As to the general question of patent-eligibility according to Art. 52 (2), (3) EPC, such a claim has to be considered having the required technical character, since for determining whether or not a claim exhibits a technical character (i.e. the claim does not relate to a mathematical method, mental act, or computer program as such) the presence of non-technical features is irrelevant.

Adapted problem/solution approach: It is established case law that non-technical features relating e.g. to mental activities, mathematical operations, or commercial steps cannot contribute to novelty and inventiveness of a claim because such features are deemed to be in the prior art. Since, by definition, non-technical features are known by the skilled person, however, they may be considered when deriving the objective technical problem from the technical differences between the instant claim and the closest prior art. This so-called adapted problem/solution approach was developed in T 208/84 (Vicom) and T 641/00 (Comvik) and recently confirmed by the Enlarged Board of Appeal's opinion on case G 3/08.

Tangible technical result: For assessing inventive step, the Board of Appeal had to observe whether the mental activities of steps B to E are non-technical features or whether theses steps interact with technical step A "so as to yield a tangible technical result", whereas the term "tangible result" appears a bit strage in a Board of Appeals' argumentation regrading non-technical features of a computer-implemented invention, since this term has not yet been used in the pertinent case law.

In fact, it was the famous US CAFC decision "State Street Bank" which found that "a claim is eligible for protection by a patent in the United States if it involved some practical application and it produces a useful, concrete and tangible result". The so called "useful, concrete and tangible result test" for assessing patent-eligible subject mater was later substituted by the "machine or transformation test" in the CAFC decision In re Bilski which, in turn, was rejected by the US Supreme Court in its Bilski vs. Kappos ruling (see my earlier postings here and here). 

Non-technical features: The Board of Appeal observed that steps B and C of are indeed very generally formulated, whereas the specification is in this respect no more generous than the claims in terms of information and the “Summary of the Invention” is no more that a mere repetition of the claims and an outline of the flow of data treatment with no concrete details.

According to the invention, the probability distributions of steps C and D are generated using a program written in C whose listing is shown in Appendix A of the specification. With regard to this computer program, the Board observed that it concerns
an unspecified embodiment wherein the data treatment is carried out using computer processing employing a computer software called "GetGenos". This section does not provide a truly useful example of how to proceed within the framework of the outlined method. In fact, the mathematical reasoning starting from an actual experimental value determined according to step A and ending with the determination of a precise genotype according to step E is not described in detail. Moreover, it fails to supply a reasonably complete and sufficient description of the software "GetGenos" especially developed by the inventors to produce GBA data [...].

This deficiency is remedied neither by the mathematical formulae referred to […] nor by the mere code lines written in C language contained in Appendix A […], which […] serve the purpose of generating probability distributions. No informational content can be attributed to the constants, variables and functions mentioned therein.
Based on this observations the Board of Appeal took the position that, based on the disclosure, a skilled person cannot "understand how to proceed from the first reaction value collected in step A through steps B, C and D to the determination on a probabilistic basis of the genotype of step E" and concluded that "no interaction can be established between the technical activity of step A with the mental activities of steps B to E leading to a tangible technical result, as required by the case law".

Lack of inventive step: The Board concluded that steps B to E only relate to a general manner of data analysis and are thus to be ignored in assessing inventive step, so that the claimed method of determining the genotype was technically only limited by "the general and broad wording of step A". Such a method, however, could easily be identified in the cited prior art. The Board thus concluded that the claimed subject matter does not meet the requirements of Art. 56 EPC and revoked the patent.

Comments: It comes as a surprise, that the Board argued that steps B to E are non-technical based on the finding that no reasonably complete and sufficient disclosure exists that may clarify an interaction with technical step A "so as to yield a tangible technical result". In my opinion, this argumentation rather relates to insufficient disclosure according to Art. 83 EPC, which even was raised by the opponent as a ground for opposition pursuant Art. 100 b EPC. If the lack of disclosure really is decisive for features B to E being non-technical, the immediate question arises whether a more complete disclosure could have potentially prevented this objection.

It appears that the Board somewhat relied on the case law for patentability of mathematical methods (cf. Case Law of the Boards of Appeal, 6th ed. 2010, section I.A.2.2.2), according to which
a basic difference between a mathematical method and a technical process can be seen [...] in the fact that a mathematical method or a mathematical algorithm is carried out on numbers [...] and also provides a result in numerical form, the mathematical method or algorithm being only an abstract concept prescribing how to operate on the numbers. [...] In contrast, if a mathematical method is used in a technical process, that process is carried out on a physical entity (which may be a material object but equally an image stored as an electric signal) by some technical means implementing the method and provides as its result a certain change in that entity. [...] Even if the idea underlying an invention may be considered to reside in a mathematical method, a claim directed to a technical process in which the invention's method is used does not seek protection for the mathematical method as such (see also T 208/84, T 1161/04).
That is, mathematical operations are to be considered technical if  the claim specifies what physical entity actually is represented by the mathematically processed data, i.e. if the processed data itself is technical, such as control parameters of a technical process or the like. 

A richer disclosure that clearly specified the concrete relationship between technical step A and mathematical steps B to E as well as some technical effects of those steps could certaily have helped to argue that steps B to E interact with step A in a technical way ("so as to yield a tangible technical result") and thus have to be considered when assessing inventive step.

Similar points were raised by the IPKat, who thinks that "the Board's reasoning relating to the non-technical parts of the claims being insufficiently described seems to be a bit unusual and tends to put the whole decision in a strange light" as well as by well-know CII specialist Stefan Steinbrenner (see his pdf summary on the EPLAW blog):
This finding arrived at by reasoning an absence of interaction from a lack of detailed disclosure is somewhat surprising, in particular since it seems to lose sight of the intended focus on the solution of a technical problem. The decision thus [...] does not answer the question of whether such mathematical steps more or less necessarily performed by computer processing might in principle make technical contributions in biotechnology [...]. Nor do we know for the moment how a "tangible technical result" can reliably be achieved or how specific the disclosure must be to establish the required interaction.