Johns Hopkins University, GDF-9 (EPO, T 1329/04, 2005)

Dhenne Avocats · 11 October 2026

Our analysis

Summary

Technical Board of Appeal 3.3.08 dismissed an appeal against the refusal of an application for growth differentiation factor GDF-9. It held that the claimed polynucleotides (SEQ ID NO:3 and the sequence encoding the polypeptide of SEQ ID NO:4) lacked inventive step because the application did not make it at least plausible that the stated technical problem had actually been solved. It added that evidence published after the filing date cannot be the sole basis for that showing. The catchword states the plausibility requirement within the problem and solution approach; the epo.org page lists many later decisions citing it.

Facts and procedure

European application No 94907259.9, filed by The Johns Hopkins University School of Medicine concerned « Growth differentiation factor-9 ». By a decision of 5 August 2004 the Examining Division refused it under Articles 56, 57, 83 and 84 EPC: in its view, cloning a putative new member of a gene family from conserved sequence motifs was routine, tissue-specific expression alone could not elucidate any function, and post-published documents could not be taken into account (point I). At the oral proceedings of 28 June 2005 the Board ruled on two requests filed on 26 May 2005, whose claim 1 covered both sequences (points III and VI).

The applicable law

The only ground relied on was Article 56 EPC. The Board took document (3) as the closest prior art, which identified GDF-1 as a new member of the TGF-β superfamily on structure alone (points 1 to 3). The problem was framed as isolating a further member of the superfamily (point 4). Following T 939/92, the Board accepted that predictions of structure and activity relationships may be taken into account, up to a limit (point 14).

Question

Can an application that attributes merely presumed functions to a protein support inventive step, and can post-published evidence make up for the absence of any showing in the application?

Decision

The Board first asked whether the problem had been plausibly solved (point 6). GDF-9 has only six cysteine residues instead of the seven characterising the superfamily, and its highest homology with another member, BMP-4, is 34%, far from the 70 to 90% found within subgroups; a purely structural approach cannot place it in the family (points 7 and 8). The application discloses only that GDF-9 is expressed in ovarian tissue, which is useful but insufficient as to function (point 9). Referring to Article 60(2) EPC, the Board rejected the first-to-file argument: the application must show that a problem was solved, not merely put forward, at that date; listing every putative function of a compound is not the same as providing technical evidence of a specific one (point 10). As the application did not contain enough to make the solution at least plausible (point 11), the post-published documents showing that GDF-9 was indeed a growth differentiation factor could not be considered. Admitting them would make the recognition of a solution vary over time, contrary to the principle that inventive step is assessed at the effective date; such evidence may supplement a showing but cannot be its sole basis (point 12). Having distinguished T 182/03 (point 13) and found the limit accepted in T 939/92 exceeded (point 14), it refused both requests and dismissed the appeal (point 15 and order).

Key points for practice

• Plausibility of the solution is a threshold question in the problem and solution approach, assessed on the application as filed (points 6 and 13).
• A list of hypothetical functions is not technical evidence (point 10).
• Post-published evidence may support a showing begun in the application, never replace it (point 12).
• Practical point: where a sequence’s structural membership of a family is uncertain, the application must contain functional data at filing.

Relevance before the UPC

Since the UPC assesses validity against the grounds in Article 138(1) EPC, which include Article 56, the Board’s reasoning on plausibility and post-published evidence carries over directly to revocation actions concerning biotechnology patents.

Provisions applied

European Patent Convention

Art. 56; Art. 60(2)

Case law cited

T 939/92 (OJ EPO 1996, 309); T 182/03

Prepared by Dhenne Avocats from the text of the decision (epo.org, official PDF of the decision), consulted on 11 October 2026. Only the official text is authoritative.