For the half-year to 30 June 2014, the IPKat's regular team is supplemented by contributions from guest bloggers Alberto Bellan, Darren Meale and Nadia Zegze.

Two of our regular Kats are currently on blogging sabbaticals. They are David Brophy and Catherine Lee.

Saturday, 30 October 2010

T 784/06 - "technical effect" reaches bioinformatics

The IPKat has been prompted by the EPLAW Patent Blog to read a recent EPO decision, who in turn have been prompted by none other than Stefan Steinbrener.  This name will be familiar to all those who have been keeping a close watch on developments at the EPO leading up to the recent Enlarged Board decision of G 3/08 (commented on by Tufty here, among many others).

The decision, T 784/06 (currently available via the EPO register for application 95906094.8), relates to an application that was originally filed in 1994 for a method of automatic genotype detection.  The method used probability distributions to determine a particular genotype at a locus within genetic material obtained from a biological sample. The actual invention, however, was really all about a mathematical method carried out on a computer, which allowed the probability distributions to be assessed and a result obtained.  If this is all starting to resemble the invention in Vicom, you are already on the right lines.

(Right: the code in question. Can you make any sense of it?)

As a result of the mixture of 'technical' and 'non-technical' features in the claimed invention, the Board took into account the type of analysis normally carried out for computer-implemented inventions.  The applicant argued that the mathematical parts of the claim contributed to the technical effect of the invention, and should therefore be taken into account when assessing inventive step.  The opponent, however, argued that the mathematical parts "did not require any further technical considerations" and were to be disregarded in the assessment of inventive step.

The Board considered that the correct assessment to be made was that from decisions such as T 641/00 (Comvik) and T 154/04 (Duns Licensing), both of which related to computer-implemented inventions.  The question to be answered was whether the claimed invention in the field of biotechnology having a mix of technical and non-technical features could be inventive.  Non-technical features such as mental activities could be taken into account when assessing inventive step, provided they "interact with the technical subject matter of the claim for solving a technical problem and thereby contribute to the technical character of the claimed subject matter" (reasons, point 4).  For assessing inventive step therefore, the question was whether the mental activity steps of the claimed invention interacted with the technical activity part to yield a tangible technical result.  The applicant argued that they did, and the opponent that they did not.

All this was, as far as the IPKat was concerned, quite conventional and straightforward.  The Board, however, then threw something of a googlie in deciding that the non-technical steps were only generally formulated and not well described in the specification.  The specification did not provide a reasonably complete and sufficient description of the software, and this was not remedied by the mathematical formulae or the extract of code provided.  As a result, no interaction could be established between the mental activity steps of the claims and the technical part, and these steps could be ignored when assessing inventive step.  The claimed invention therefore lacked an inventive step, since the technical part was disclosed in documents cited during proceedings. The patent was revoked.

(Left: Or perhaps the real invention was just too simple?)

The IPKat thinks that this is decision is an important one, but perhaps not as important as it could have been.  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.  The IPKat suspects that the Board had decided that the claimed invention should not be allowed but were struggling to find sound reasons to reject it in light of the arguments presented. However, the decision highlights at least two important points.  Firstly, that the case law relating to computer-implemented inventions can certainly be applied to the field of bioinformatics.  Just because an invention is classified in the biotech field does not mean that the invention has to be assessed according to the case law relating only to that field.  Secondly, if an invention is to rely on computer-based processing steps including mathematical methods then the disclosure needs to be as full as possible, and preferably described in a way that makes it understandable to someone outside the field of software coding.  A few lines of impenetrable computer code and a couple of equations does not necessarily amount to a sufficient disclosure of the invention.  The IPKat, who also could not make much sense of the invention as disclosed in the specification, wonders whether a different outcome could have been obtained if it had been more fully described or simply better explained.

3 comments:

Czentovic said...

"The IPKat, who also could not make much sense of the invention as disclosed in the specification, wonders whether a different outcome could have been obtained if it had been more fully described or simply better explained."

Art. 52(1) EPC is a requirement on the scope of the subject-matter claimed, not on how the subject-matter within this scope is disclosed in the application. So providing more details in the description can normally not overcome an objection under Art. 52(1) EPC. (I know of one exception: it may help if the description mentions the technical effect of a feature or a combination of features. But mentioning technical effects is not the same as providing implementation details.)

So I fully agree that the reasoning of Board 3.3.08 is unusual. Maybe what the Board is trying to say is that even the description does not disclose an embodiment in which the mathematical steps contribute to a technical effect?

Regarding claim 1 of the main request, clearly step A is technical ("reacting the material"). The value produced in step A is then mathematically manipulated (in a rather undefined way) to determine the genotype.

I wonder whether better defined mathematical manipulations could have resulted in a technical contribution, since all they do is provide a cognitive interpretation of the value produced in step A.

However, maybe it could be said that the mathematical processing may in certain cases add some technical information in the sense that the reaction value itself does not uniquely determine the genotype, but the mathematical processing provides some kind of educated guess of the correct genotype based on technical (i.e. biological) principles. In that case, I suppose the claim should somehow reflect the application of these technical/biological principles.

But is an educated guess a technical result? For example, given a set of biological measurements of a patient, does the diagnosis of a particular condition based on these measurements achieve a technical result? Or is it merely a (non-technical) mental act? Does G 1/04, r. 5.2 not say the latter?

Tevildo said...

As code, it's not so bad - I assume COND_NEG_PROB is a macro rather than a function - but we don't really have enough detail of what the actual calculations are to make a decision on obviousness.

Compare SAS v World Programming, para 81, which is a _dreadful_ bit of code. There are so many things wrong with it, but the worst bit is highlighted in orange - the section with the handwritten amendment will never be executed (which is probably why there's a handwritten amendment). If this is the sort of material that people go to court over...

Anonymous said...

Mathematical methods are only excluded as such (A 52(3). If they are part of something else, it has to be assessed whther they contribute to the technical character of the invention (whatever this may be).

For long, mathematical methods have been patented when applied eg to cryptography or signal processing (noise reduction or computer tomography).

I cannot see why in the present case this should be different. A mathematical treatment of the data is in principle patentable if it leads to something having a technical character (whatever this may be).

Insufficient disclosure of the method's steps is something that falls under A 100(b) EPC and has in my eyes nothing to do with technicality. If the method produces merely an educated guess or not, is also that has nothing to do with it being sufficiently disclosed.

I see this decision as mixing two different aspects of patentability which should be kept clearly separated.

Subscribe to the IPKat's posts by email here

Just pop your email address into the box and click 'Subscribe':