EP 3 243 073 B1 relates to a “Method for controlling continuous chromatography and multi-column chromatography assembly”.
Brief outline of the case
The OD decided maintenance in amended form.
Opponent 1 withdrew its opposition, but Opponent 2 appealed the decision.
The board decided revocation of the patent for lack of clarity.
The clarity problem
Claim 1 contains the step of calculating a process parameter “by applying a chemometric method comprising a trained and validated mathematical model” (feature 1.4).
The feature “comprising a trained and validated mathematical model” was included in claim 1 during opposition proceedings and is therefore to be examined for the requirement of clarity.
The proprietor’s point of view
The objection of lack of clarity is not admissible.
The proprietor was of the opinion that in the first instance proceedings the opponent raised an objection of clarity only against claim 12 of the MR, but not against claim 1. Therefore, the current objection of clarity against claim 1 is late and not admissible.
Clarity is given
The skilled person in the present case has in-depth knowledge of chemistry and mathematics as well as extensive experience with chemometric calculations.
This knowledge is also described in the patent specification, e.g. in paragraphs [0049], [0051], [0070], [0072] and [0077]. Based on his relevant technical knowledge, he knows exactly what a “trained and validated mathematical model” is. In particular, he clearly recognizes the difference between a “trained and validated” and an “untrained and unvalidated” mathematical model.
The proprietor referred to the case law of the BA, according to which the claims must be clear for reasons of legal certainty because they serve to define the scope of protection of the patent (Case Law of the Boards of Appeal, II.A.1.1).
However, a certain amount of effort for the skilled person in defining the scope of protection must be accepted. According to case law, the scope of protection must “only” be recognizable for the skilled person without unreasonable effort.
Since the “trained and validated mathematical model” is always better than the “untrained and unvalidated mathematical model”, the skilled person can recognize this without unreasonable effort.
In addition, the skilled person is aware that mathematical models of this type are usually purchased, so that the manufacturer of the model can be contacted to find out whether the model has been trained.
Contrary to the opponent’s and the board’s view, the “trained and validated mathematical model” was not trained and validated before the claimed method was carried out.
Claim 1 does not define a finished device, but a method with a method step in which the mathematical model is explicitly used. Therefore, the training and validation of the mathematical model are part of the claimed method.
The opponent’s point of view
The feature in claim 1 “comprising a trained and validated mathematical model” is unclear.
The training and validation per se of an untrained and unvalidated mathematical model does not take place during the claimed process, but before it.
The mathematical model is already in a trained and validated, completed state before the claimed method “for controlling a multi-column chromatography arrangement” is carried out.
“However, it cannot be “’seen” from the mathematical model created whether it has been checked [validated] or not. The situation is similar with a trained mathematical model. If the mathematical model is available, it is not recognizable whether the mathematical model has been trained or not.
However, in the claimed state, a trained and validated mathematical model cannot be distinguished from an untrained and unvalidated model found by skilful modelling.
As a result, the scope of protection cannot be precisely identified. The independent claims 1 and 12 therefore do not meet the requirements of Art 84.
The board’s decision
Admissibility of the objection of lack of clarity
The clarity objection concerning the “trained and validated mathematical model” was first raised by the opponent against claim 12 at the OP before the OD.
Since the OD considered that the feature of the “trained and validated mathematical model” was clear, “it was not appropriate to raise the same objection again on the same feature, but this time on a different claim”.
The objection of lack of clarity against the feature in claim 1 “trained and validated mathematical model” was thus at least implicitly raised by the opponent during first instance proceedings.
The lack of clarity
That the skilled person would recognize on the basis of his expertise whether the mathematical model before him is already trained and validated or not is a mere assertion that is not based on verifiable facts or other evidence.
The skilled person cannot recognize ad hoc whether the model has been trained and validated or not. In order to recognize this clearly, he must be able to compare the model before him with the corresponding untrained and unvalidated model.
Without this comparison, he is left in the dark as to whether the model before him is already trained and validated or whether it is untrained and unvalidated but was originally developed by the skilled person on the basis of relevant experience with chemometric calculations of process parameters in purification processes of biopharmaceuticals in such a way that it already comes very close to a trained and validated mathematical model.
Whether the present mathematical model has already been trained and validated is not a question of the extent or reasonableness of the effort involved, but of knowledge of the initial, untrained and unvalidated state of the mathematical model. However, the initial state is not defined in claim 1.
Claim 1 does not define a device but a method. Nevertheless, the wording of feature 1.4 “… by applying a chemometric method comprising a trained and validated mathematical model …” only defines the step of “applying a chemometric method”.
This chemometric method used clearly comprises a product, the mathematical model, in a specific, finished and completed state. This “comprising” is not a method step, since claim 1 does not define an active connection between the chemometric method and the mathematical model. The mathematical model hangs,” so to speak, “in the air. It is simply present and encompassed by the chemometric method.
As the MR and all valid AR suffered the same defect, the patent had to be revoked.
Comments
I would have thought that the notion of unreasonable or undue effort applies to sufficiency, not to clarity.
When introducing a mathematical model in a process or a corresponding product claim, it appears of utmost importance to precisely define whether the model is previously trained or if the training is part of the process.
It is somehow surprising that the OD considered the inclusion of the mathematical model fulfilled the requirements of clarity.
Comments
Leave a comment