Selection inventions are concernedwith the selection of one or more spe-cific embodiments from a general dis-closure in the prior art. A generaldisclosure in the prior art is not con-

sidered to take away the novelty of a later claim toa specific embodiment. For example, where theprior art discloses metal, a later claim to copperwould still be novel.

However, the assessment of novelty of a selec-tion invention is less straightforward when theprior art relates to lists or ranges.

In Europe, case law concerning the novelty ofselection inventions relating to lists or ranges hasevolved over time and has come to impact se-verely on how applicants for patents in Europe 2

WWW.MANAGING IP.COM

Selectioninventions Louise Tottie and Anette Romare of Valeaoutline important aspects of selectioninventions relating to lists and ranges, toconsider when drafting patent applications

EUROPE

L I F E SC I ENCES I P FOCUS

EUROPE

may amend their patent applica-tions during prosecution, as wellas on the validity of a priorityclaim from an earlier filed appli-cation. In all cases, the crucialquestion to be answered is whatinformation can be directly andunambiguously derived fromthe disclosed lists or ranges.

While recent decisions fromthe Boards of Appeal of the Euro-pean Patent Office (EPO) con-firm and clarify established caselaw relating to lists, the situationappears to be less clear for selec-tion inventions relating toranges.

Two list principleA well-established principle inEurope is that a selection isnovel when it results from a se-lection of elements from twolists of some length. In contrast,no novelty is recognised whenthe selection is made from a sin-gle list of individualised ele-ments. This principle derivesfrom T 12/81, later confirmed inT 7/86 and has since beenadopted by the Boards of Appealconsistently.

For instance, when a docu-ment discloses a compound hav-ing two substituents R1 and R2,wherein R1 is methyl, ethyl orpropyl and R2 is chloro, bromoor iodo, the specific combinationof R1 being methyl and R2 beingiodo is not considered disclosed.Accordingly, if the documentwere a prior art document, sucha specific selected combinationwould be novel if claimed in a

new patent application. How-ever, if the document were apending patent application thereis no basis in the application forselecting and claiming the spe-cific combination.

It should be pointed out thatthe above principle only appliesto lists relating to structural in-formation. Accordingly, if onelist discloses starting materialsand the other list disclosesprocess conditions, the in-evitable product resulting from acombination of any one of thestarting materials with any oneof the process conditions is con-sidered disclosed. Thus, no selec-tion invention is possible sincethe product lacks novelty. On theother hand, in a patent applica-tion under prosecution, the listscan be used as a basis for claim-ing a product resulting from thecombination.

The structural informationin the lists need not necessarilyrelate to starting substances re-quired to prepare end products,but may be entities for the prepa-ration of mixtures. In T 401/94,the Board found the claimedcomposition novel since it corre-sponded to a specific combina-tion of constituents which hadbeen selected from a relativelylong list in a prior art document.

The following decisionsfrom the Board of Appeal furtherillustrate the view of the EPOwith respect to selections fromlists.

Two components in asingle list

In T 1374/07 a claim was directed to a bread im-prover including particles made of at least fat andenzymes. The patent was revoked by the Opposi-tion Division, holding that it contravened article123(2) EPC, since the feature “made of at least fatand enzymes” lacked basis in the application asfiled.

The patentee lodged an appeal against the de-cision arguing that the selection was made fromone list, and that the case law relating to selectionsfrom two lists was not applicable.

The Board of Appeal did not agree with thepatentee. The Board pointed out that although thedescription stated that “the particles are made ofat least two different active ingredients” as well as“the improver according to the invention can fur-4

WWW.MANAGING IP.COM

EUROPE

Louise TottieLouise is a Swedish patent attorney. Shehas a PhD in organic chemistry.

Louise has expertise within the areasof chemistry, pharmaceuticals and poly-mer technology. She has experience of thepharmaceutical industry where sheworked for AstraZeneca as a research sci-entist in medicinal chemistry for 10 yearsbefore joining the patent profession in 2003in the company�s IP department.

Louise handles patent work in thewhole field of chemistry. She specialises inchemical and life science inventions, suchas general chemistry, pharmaceuticals,chemical engineering and polymers.Louise files and prosecutes patent applica-tions. She works closely with companies onidentifying inventions and proposingpatent portfolio management and also pro-vides infringement and validity investiga-tions, freedom-to-operate analyses and duediligent work.

The crucial question is whatcan be directly andunambiguously derived fromthe disclosed lists or ranges

ther comprise one or more ingre-dients selected from the groupconsisting of emulsifiers, fat, en-zymes, sugar, organic acids,minerals, polysaccharides, pro-teins and/or a mixture thereof”,this provides no basis for theclaimed combination of fat andenzymes. The Board stated thatthe selection of two ingredientsfrom a single list is equivalent tomaking a selection from twoidentical lists and that the ex-pression “and/or a mixturethereof” cannot serve as a basisfor the combination of “at leastfat and enzymes” since it coversnot only the 28 possible combi-nations of two ingredients butalso the combination of three in-gredients, of four ingredients,and so on. In particular, theBoard stated that no teaching inthe application as filed guidesthe skilled person towards thespecific combination of fat andenzymes and therefore it consti-tutes added matter.

A learning point from thiscase is that it is important thatthe application explicitly recitesall embodiments for which theapplicant can envisage the needfor useful protection. Merely list-ing possible constituents with-out any guidance as to howthese may be combined affordslittle or no help when amendingthe application during prosecu-tion.

Specific combinationsfrom two listsThis example relates to combi-nations of dipeptidylpeptidase-IV inhibitors with antidiabetic agents. Theclaimed combination of the dipeptidylpeptidase-IV inhibitor LAF237 and an antidiabetic agent se-lected from rosiglitazone, pioglitazone, andtroglitazone was considered by the Opposition Di-vision to contravene article 123(2) EPC. TheBoard of Appeal disagreed, pointing out that thedescription discloses a particularly preferred em-bodiment in which a dipeptidylpeptidase-IV in-hibitor is selected from LAF237 and DPP728 andan antidiabetic agent is selected from a list of 22compounds. The Board concluded that the skilledperson would directly and unambiguously derive44 individual combinations, among them thethree claimed combinations indicated above. All44 combinations were considered equally highly

preferred and useful, and the se-lection of three of these combina-tions was considered to bemerely a deletion of the other 41combinations. Such a deletion inorder to improve patentabilityover prior art is admissible in ac-cordance with established caselaw.

Thus, the specific combina-tion of a dipeptidylpeptidase-IVinhibitor selected from LAF237and an antidiabetic agent se-lected from rosiglitazone, piogli-tazone, and troglitazone wasconsidered by the Board of Ap-peal to be allowable. In contrastto the example of the bread im-prover, the Board concluded thatthe disclosure of the original ap-plication provided guidance tospecific combinations and there-fore such combinations were dis-closed and could provide a basisfor amendment.

Two lists, claimingpriority In decision T 0077/97 the Boardheld that the priority claim wasinvalid. The application was di-rected to specifically substitutedtaxoïd compounds. The priorityfounding application disclosedsubstituted taxoïd compounds,but the combination of sub-stituents was not specificallymentioned. Consequently, theBoard found the priority claim tobe invalid. As there had been anintermediate publication of thespecifically substituted taxoidcompounds, the new applicationwas found to lack novelty.

5

L I F E SC I ENCES I P FOCUS

EUROPE

Anette Romare Annette is a European patent attorney, andan authorised patent attorney (SE). Shehas an MSc in chemical engineering andbiotechnology.

Anette joined the patent profession in1985 and specialises in absorbent materialsand products, paper, cellulose and fibretechnology, polymers, adhesives, generalchemistry, medical technology, packagingand packaging technology.

Anette has expertise in opposition andappeal proceedings at the European PatentOffice and provides freedom-to-operateanalyses and infringement opinions, aswell as drafting and prosecution of patentapplications. Anette has been involved inthe CEIPI program as a tutor since 1994.

Recent decisions from theBoard of Appeal show thatthe three-step test is notapplied in a consistent way

This shows the importance of disclosing thespecific combination of features, in this case a sub-stituted taxoïd compound, for which the applicantdesires patent protection.

Numerical rangesIn situations where the prior art is a numericalrange for which only the end points are disclosed,the EPO takes the view that values between the endpoints are not specifically disclosed. For instance, ifthe disclosure is an alkyl group with a chain lengthfrom one to four carbons, that is, a chain from C1to C4, a C2-chain is not considered disclosed.

In cases where a sub-range is selected froma broader numerical range of the prior art thethree-step test is applied. This test derives from de-cisions T 198/84 and T 279/89 and is described inthe Guidelines for Examination in the EuropeanPatent Office (part G, chapter VI-8). A sub-rangeselected from a broader numerical range of theprior art is considered novel, if each of the follow-ing criteria is satisfied:

the selected sub-range is narrow compared to•the known range;the selected sub-range is sufficiently far re-•moved from any specific examples disclosedin the prior art and from the end-points of theknown range;the selected range is not an arbitrary specimen•of the prior art, that is, not a mere embodimentof the prior art, but another invention (purpo-sive selection, new technical teaching).Accordingly, a selection invention from a nu-

merical range appears to differ from other typesof selection inventions in that it should be a pur-posive selection (a technical effect is needed to ful-fill the requirement of novelty). For all other typesof inventions this criterion is dealt with separately,in the assessment of inventive step.

However, some recent decisions from theBoard of Appeal show that the three-step test isnot applied in a consistent way.

In T 1233/05 relating to a non-azeotropiccomposition, the Board decided not to take the

third criterion into account in the assessment ofnovelty. Instead, the third criterion of a purposiveselection was dealt with in the assessment of in-ventive step. Two further decisions (T 0230/07 andT 1130/09) took the same approach and movedthe third criterion from the assessment of noveltyto the assessment of inventive step. Nevertheless,in decision T 126/09 the Board again applied allthree criteria in the assessment of novelty.

In decision T 1827/08, the patent had a claimdirected to a pipe provided with a barrier layerhaving a thickness of less than 1µm. The prior artdisclosed a pipe with a barrier layer between 0.1and 200µm, and an example was provided wherethe barrier layer was 9µm. According to the Board,the claimed pipe fulfilled the three-step test. Theprior art range was justified by the provided ex-amples. The prior art example of 9µm was faraway from the claimed range of less than 1µm.

Clearly, there is no consistent EPO practicefor assessment of novelty for selection inventionsrelating to numerical ranges. It remains to be seenif a change in practice will take place or if therewill be a referral to the Enlarged Board of Appeal.

Priority from a broad rangeThe following example shows how the novelty ofa European patent can be destroyed by the appli-cation from which it claims priority. In decision T680/08, the Board held that the priority claim of aEuropean patent directed to the range 0.330 to0.415 kWh/kg was invalid since the range dis-closed in the priority document was 0.325 to 0.415kWh/kg. The priority document was a Europeanpatent application that had been published, andtherefore became prior art for novelty purposesunder article 54(3) EPC. Using the three-step testthe Board found the claimed range of 0.330 to0.415 kWh/kg to lack novelty.

This shows that it may be advisable not to filepriority founding applications with the EPO since,in contrast to national patent applications, theseapplications may be held to be comprised in thestate of the art.

Finding a fall backFinally, we advise applicants for selection inven-tions relating to lists or numerical ranges to becareful in providing fall-back positions in the ap-plication for all embodiments for which the appli-cant is interested in obtaining patent protection.As can be seen from the discussion above, it willnot be possible to obtain patent protection for com-binations of features or for ranges which are notexplicitly disclosed or to which there is no guid-ance in the application as filed. On the other hand,if the prior art discloses lists or ranges with no in-dication of how features may be combined or aspecific range selected, it may be worthwhile toconsider filing a patent application directed tosuch a selection invention.6

WWW.MANAGING IP.COM

EUROPE

It is important that theapplication explicitly recitesall embodiments for whichthe applicant can envisagethe need for usefulprotection