The number of bioinformatics patent applications being filed, although still relatively small, has been steadily increasing over the past few years. The prosecution of these applications provides a unique challenge. Bioinformatics applies computer science and mathematics to biological data and so drafting requires cross-discipline expertise. Patent offices have not yet adopted a multidisciplinary approach to examination, often providing objections that are inappropriate from either the biology or computer science perspective.

This inconsistent approach to bioinformatics applications is reflected in case law from the EPO Boards of Appeal. Although the number of decisions in this field is small, two decisions are of particular interest.

The first is T784/06 – an appeal by both the patentee and the opponent against the interlocutory decision of the opposition division. The contentious issue was inventive step. The application related to a method of determining the genotype at a locus within genetic material obtained from a biological sample. The genotype was determined from a probability distribution using a mathematical program. Importantly, the method was a mix of technical steps (obtaining the data) and non-technical steps (the application of the mathematical method to the data to arrive at the genotype). The Board (from the Biotechnology division) considered that the fundamental question was whether the technical and non-technical steps interacted to yield a “tangible technical result”. This reasoning was consistent with present practice – a mix of technical and non-technical steps are patentable, even if the non-technical features form a dominating part, as long as the overall result provided by the combination is technical. Unusually, however, to answer this question the Board looked at the support for the non-technical steps in the application and found that the non-technical stepsm were too generally formulated and insufficiently described. This deficiency was not remedied by the inclusion of the mathematical formulae or the one page of computer code. On this basis the Board concluded that there was no interaction between the technical and non-technical steps and the patent was revoked.

The second is T2050/07 – an appeal against the decision of the examining division to refuse the application. Again, the application related to a method of analysing a DNA sample to determine a genotype using probability distributions (and a mathematical equation to resolve the data). And again, the claim was a mix of technical and non-technical steps. However, the contentious issue here was novelty. In this case the Board had to decide whether the distinguishing feature, which was non-technical, made a technical contribution and could therefore impart novelty. The Board (including this time the chairman of the computer science division) concluded that it did, and that in contrast to T784/06, the description “made it sufficiently clear how the distinguishing features… should be implemented and how they interact with the remaining steps to provide a common technical result”.

Given the differing Board composition between the two cases, it is tempting to wonder if the T784/06 decision would have been different if the former had included a computer science examiner. Nonetheless, although the second decision is more in keeping with current practice, we can still learn from both. Clearly, when drafting bioinformatics applications, all non-technical steps need to be fully described and explained in their biological context so that the application can be understood by both biotechnology and computing examiners.