This appeal stems from an interference proceeding involving the claims of United States Patent Application Serial No. 08/443,184 (the "'184 application") and United States Patent No. 5,639,940 (the "'940 patent"). William H. Velander, William N. Drohan, Henryk Lubon (deceased), and John L. Johnson (deceased) (collectively "Velander") are the inventors named on the '184 application. Ian Garner, Donna E. Prunkard, and Donald C. Foster (collectively "Garner") are the inventors named on the '940 patent. Velander appeals the decision of the United States Patent and Trademark Office, Board of Patent Appeals and Interferences ("Board"), that granted Garner's preliminary motion to have all of the allowed claims of the '184 application held unpatentable as obvious over the prior art. Garner v. Velander, Interference No. 104,242 (B.P.A.I. Aug. 16, 2001) ("Velander"). Because the decision of the Board is supported by substantial evidence and is not contrary to law, we affirm.

This case relates to the production of non-human mammals that have been genetically altered ("transgenic animals") so that they produce the enzyme fibrinogen in its biologically active state. After the enzyme is produced, it is recovered from the milk of the mammal. The fibrinogen enzyme consists of two copies of three separate peptides: the Aα chain, the Bβ chain, and the γ chain.

Fibrinogen is normally synthesized by the liver and plays a key role in blood clotting. An absence of fibrinogen in a human can lead to a malfunction in the blood clotting process. The production and isolation of biologically active fibrinogen can provide a source of fibrinogen to treat individuals suffering from a deficiency of the enzyme. The present interference involves competing claims to a transgenic animal (and methods to make such an animal) that produces fibrinogen and secretes it into its milk.

The '184 application (also referred to as the "Velander application") was filed on May 17, 1995. The application was afforded the benefit of the February 18, 1994 filing date of its parent application, United States Patent Application Serial No. 08/198,068. The claims of the Velander application are directed to (i) the production of a transgenic animal that expresses a foreign (or heterologous) biocompetent fibrinogen protein and (ii) a method of producing biocompetent fibrinogen using such an animal.

On March 3, 1994, Garner filed United States Patent Application Serial No. 08/206,176. The '940 patent (also referred to as the "Garner patent") issued from that application. Upon issuance, the '940 patent was assigned to Pharmaceutical Proteins Ltd. and ZymoGenetics, Inc. The claims of the Garner patent, like those of the Velander application, are directed to both the production of a transgenic animal that expresses a foreign (or heterologous) biocompetent fibrinogen protein and a method of producing biocompetent fibrinogen using such an animal.1

On June 30, 1998, the Examiner allowed claims 64-73 of the Velander application, but declared an interference between those claims and the issued Garner patent. The interference was declared on the ground that claims 64-73 of the '184 application and claims 1-33 of the '940 patent either corresponded to the proposed interference counts or would require performance of the method in the counts in order to produce the claimed transgenic animals. The interference counts, which corresponded to all of Velander's allowed claims (64-73) and all of Garner's patented claims (1-33), were as follows:

A method for producing biocompetent fibrinogen comprising:

providing a transgenic female non-human mammal carrying in its germline heterologous DNA segments Aα, Bβ, and γ chains of fibrinogen, wherein said segments are expressed in a mammary gland of said mammal and biocompetent fibrinogen encoded by said segments is secreted into milk of said mammal;

collecting milk from said mammal; and recovering said biocompetent fibrinogen from said milk.

A transgenic non-human female mammal that produces recoverable amounts of biocompetent human fibrinogen in its milk, wherein said mammal comprises:

a first DNA segment encoding a secretion signal operably linked to a heterologous fibrinogen Aα chain,

a second DNA segment encoding a secretion signal operably linked to a heterologous fibrinogen Bβ chain, and

a third DNA segment encoding a secretion signal operably linked to a heterologous fibrinogen γ chain, and

further wherein each chain is derived from the same species and is operably linked to additional DNA segments required for its expression in the mammary gland of a host female mammal.

A non-human mammal carrying in its germline DNA segments encoding Aα, Bβ, and γ chains of fibrinogen, wherein female progeny of said mammal express said DNA segments in a mammary gland to produce biocompetent fibrinogen.

In due course, after the interference was declared, Garner moved to have Velander's claims 64-73 held unpatentable under 35 U.S.C. §§ 102(b) and 103 (2000).2 In its motion, Garner conceded that if it succeeded in having the claims of the Velander application held unpatentable, the claims of the '940 patent, all of which included the use of cDNA for the production of fibrinogen, would also be unpatentable. Garner did, however, maintain that the use of genomic DNA for the production of fibrinogen, as described in Garner's Reissue Application No. 09/232,488, filed January 15, 1999, was still patentable.3 Because Garner did not identify a single reference that anticipated all of the limitations of any of the claims at issue, the Board stated that it understood Garner's argument to be that the subject matter of Velander's claims was unpatentable under section 103 in view of section 102(b) prior art. Since Velander did not argue that its claims were separately patentable and did not dispute Garner's assertion that either claim 64 or 65 of the '184 application would be a representative claim upon which an obviousness analysis could be based, the Board determined patentability on the basis of claim 65. Claim 65 of the Velander application is as follows:

A transgenic non-human female mammal that produces recoverable amounts of biologically active human fibrinogen that is converted to fibrin upon reaction with human thrombin in its milk, wherein said mammal comprises:

a first DNA segment encoding a secretion signal operably linked to a heterologous fibrinogen Aα chain,

a second DNA segment encoding a secretion signal operably linked to a heterologous fibrinogen Bβ chain, and

a third DNA segment encoding a secretion signal operably linked to a heterologous fibrinogen γ chain, and

further wherein each chain is derived from the same species and is operably linked to additional DNA segments required for its expression in the mammary gland of a host female mammal.

Velander, slip op. at 6.

Pursuant to 35 U.S.C. § 103(a),

[a] patent may not be obtained ... if the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains.

The ultimate determination of whether an invention would have been obvious under section 103(a) is a legal conclusion based on underlying findings of fact. In re Kotzab, 217 F.3d 1365, 1369 (Fed.Cir.2000). The underlying factual inquiries include (1) the scope and content of the prior art; (2) the level of ordinary skill in the prior art; and (3) the differences between the claimed invention and the prior art. Graham v. John Deere Co., 383 U.S. 1, 17, 86 S. Ct. 684, 15 L. Ed. 2d 545 (1966). If all the elements of an invention are found in a combination of prior art references,

a proper analysis under § 103 requires, inter alia, consideration of two factors: (1) whether the prior art would have suggested to those of ordinary skill in the art that they should make the claimed composition or device, or carry out the claimed process; and (2) whether the prior art would also have revealed that in so making or carrying out, those of ordinary skill would have a reasonable expectation of success.

In re Vaeck, 947 F.2d 488, 493 (Fed.Cir.1991) (citing In re Dow Chem. Co., 837 F.2d 469, 473 (Fed.Cir.1988)). Both the suggestion and the reasonable expectation of success "must be founded in the prior art, not in the applicant's disclosure." Id. Before the Board, as the moving party, Garner had the burden of establishing by a preponderance of the evidence that the invention of claim 65 of the Velander application would have been obvious as of February 18, 1993, one year before the filing of Velander's 08/198,068 application (the "critical date"). Bruning v. Hirose, 161 F.3d 681, 685-86 (Fed.Cir.1998) (holding that co-pending applications, as in an interference proceeding, invoke the preponderance of the evidence standard for validity challenges); Velander, slip op. at 4, 6.

In support of its motion, Garner argued to the Board that all the elements of the '184 application were in the prior art, and that the prior art would have suggested to one of ordinary skill in the art (i) the production of a transgenic animal that would express biocompetent fibrinogen, and (ii) a method of producing biocompetent fibrinogen using such an animal. Garner also argued that the prior art would have revealed that one of ordinary skill would have had a reasonable expectation of success in the production of such a transgenic animal and in producing biocompetent fibrinogen using that animal. Therefore, Garner contended, Velander's application would have been obvious to one of ordinary skill in the art as of the critical date of February 18, 1993.

In advancing its obviousness argument, Garner contended that the motivation to combine the elements of the invention, which Garner asserted were in the prior art, could be found in a publication by Dr. Lothar Hennighausen4 (the "Hennighausen Review") and in United States Patent No. 4,873,316 (the "Meade patent"). Garner pointed out that the Meade patent disclosed a method for the production of heterologous proteins in the milk of transgenic animals, see Meade patent, col. 1, l. 52-col. 4, l. 60, while the Hennighausen Review suggested the production of commercial quantities of plasma proteins in transgenic animals, see Hennighausen Review at 3-4. Citing the Hennighausen Review and the Meade patent, as well as United States Patent No. 5,304,489 (the "Rosen patent") and Patent Cooperation Treaty application WO 92/00239 of Anthony J. Clark (the "Clark application"), Garner argued that fibrinogen was an obvious target for expression in transgenic animals. This was so, Garner urged, because the low level expression of fibrinogen demonstrated in cell culture systems was exactly the type of problem that expression in the milk of transgenic animals was designed to overcome. Garner also argued that one of ordinary skill in the art would have had a reasonable expectation of success in producing biocompetent fibrinogen in the milk of transgenic animals in view of the prior art showing successful production of transgenic animals capable of expressing heterologous proteins in biologically active form. As support for that proposition, Garner cited several authorities.5 Garner contended that since all the elements of the invention described in the '184 application were in the prior art and since the prior art suggested overcoming the problems associated with fibrinogen production, "Velander's results were neither unexpected nor surprising."

Velander did not dispute that the elements of claim 65 of the '184 application were in the prior art. Neither did Velander dispute that the prior art contained a motivation to combine those elements. Rather, it argued that there are many variables that affect protein expression levels and that, as of February 18, 1993, the critical date, one of ordinary skill in the art would not have had a reasonable expectation of success in practicing the invention claimed in the '184 application. Velander contended that expression in cell cultures is not predictive of expression levels in transgenic animals and that, generally, expression levels of fibrinogen in transgenic expression systems is extremely variable. Velander Paper No. 41, pp. 15-21. Velander urged that, as of the critical date, a skilled artisan would have had a low expectation of success using either cDNA or genomic DNA in the transgenic production of biocompetent fibrinogen in the mammary gland. In support of its position, Velander relied on the declaration and deposition testimony of its experts, Drs. David Toman, Lothar Hennighausen, and Jeffrey Rosen.

On August 16, 2001, the Board granted Garner's motion to have the allowed claims of the Velander application declared obvious. Velander, slip op. at 35. In view of the fact that Velander did not argue that his claims were separately patentable, the Board decided to analyze patentability of all of the claims based upon claim 65 of the Velander application. Id. at 6. The Board noted that claim 65 was not confined to the use of cDNA or genomic DNA and that, as far as the production of fibrinogen was concerned, the only requirement was that production yield a "recoverable" amount. Id. The Board pointed out that all of the elements of claim 65 were incontestably in the prior art and that Garner had demonstrated a motivation to produce serum proteins, including fibrinogen, in the milk of a transgenic animal. Id. at 8. At the same time, the Board accepted Velander's view of the level of one of skill in the art:

The ordinary skilled worker would have had more than just mere `familiarity' with transgenics. Rather, the skilled person would have had experience in expressing heterologous proteins in transgenic animals and in obtaining the expressed heterologous proteins from the body fluids of the animals, particularly from milk. The skilled worker actually should have had experience as a member of a laboratory group that produced, identified and used transgenic animals to express heterologous proteins in the body fluids of transgenic animals, particularly in the mammary glands of transgenic animals.

Id. at 8-9 (quoting Velander Motion No. 41 at 3-4). The Board then moved on to Velander's main point in opposition to Garner's contention that the invention that was the subject of the '184 application was obvious. Turning to Velander's argument that a person having ordinary skill in the art would not have had a reasonable expectation of success based on the teachings in the art as of the critical date, id. at 9, the Board noted that a reasonable expectation of success is to be assessed from the perspective of one of ordinary skill in the art at the time the invention was made. Life Techs., Inc. v. Clontech Labs., Inc., 224 F.3d 1320, 1326 (Fed.Cir.2000). With that in mind, the Board proceeded to analyze the expert declarations and the cited prior art to determine whether, to one of ordinary skill in the art as of the critical date, there would have been a reasonable expectation of success in producing transgenic animals capable of expressing the protein fibrinogen in biologically active form.

Dr. Alan Colman, Garner's expert, testified by declaration that there would have been a reasonable expectation of success. The Board cautioned, however, that the Colman testimony was to be accorded little weight, except where it was supported by cited literature. Velander, slip op. at 11. Addressing the Hennighausen Review, the Board noted its discussion of the advantages of the use of transgenic milk production of proteins over production of proteins from cell culture systems and its identification of serum proteins, especially blood clotting factors, as ideal proteins for production using transgenic animals. Id. at 11-12. The Board stated that "[a]ccording to the Hennighausen review, success had already been achieved for several human proteins, including serum proteins like protein C, but at widely variable and generally low expression levels (pp. 5-6)...." Id. at 11. The Board added that the Review "create[d] the impression that producing transgenic mammalian livestock is expensive, technically challenging, and laborious, but nevertheless a viable option for producing proteins." Id. at 12. In addition, the Board pointed out that Dr. Colman cited the Meade patent as showing a reasonable expectation of success, and it noted that "Meade provides motivation to switch from cell cultures to transgenic milk production, noting cost advantages (1:15-21) and increased reliability (1:41-45), and (vis-à-vis prokaryotes) proper post-translational modifications (1:36-41)." Id. Finally, the Board observed that "Meade identifies many proteins subject to its method, including serum proteins (3:31-40)." Id.

The Board performed a similar analysis with respect to the declarations of Velander's experts, Drs. Hennighausen, Rosen, and Toman, all of whom testified as to the lack of a reasonable expectation of success. Again, as it did with Dr. Colman, the Board cautioned against reliance upon expert testimony that was not sufficiently supported. The Board determined that Dr. Hennighausen's conclusions, like those of Dr. Colman, were "entitled to little weight beyond the corroborating references that he cites." Id. at 13. One such reference was an article by Dr. Hennighausen himself (the "1997 Hennighausen article"),6 in which he noted some of the difficulties in expressing proteins in transgenic animals. The Board, however, concluded that the two-page Hennighausen article, published in 1997 over three years after the critical date, was not convincing support for Dr. Hennighausen's declaration because it was not specific enough as to the problems with transgenic milk production. Velander, slip op. at 13. The Board also stated that there are few circumstances in which a later publication can be used to show earlier knowledge. Id. Cf. In re Koller, 613 F.2d 819, 824 n. 5 (C.C.P.A.1980) (discussing the narrow circumstances in which later publications can be used to show earlier knowledge). The Board pointed out that the 1997 Hennighausen article stated that there were two problems associated with the use of transgenic animal milk protein production on a large scale, specifically: (1) variable levels of production, and (2) problems with post-translational modification.7 Velander, slip op. at 13. The Board noted that the article stated that the first problem could be solved in mice, but not livestock, and that the second problem came into play only when a high level of protein was produced because such a level of production could overwhelm the tissue's "enzymatic machinery" for performing post-translational modifications. Id. The Board further noted that claim 65 was not limited to livestock and that claim 65 required only a "recoverable" amount of biocompetent fibrinogen, making both of these problems inapplicable. Id.

In his declaration, Dr. Hennighausen expressed the view that purification of fibrinogen could be a problem. Hennighausen Declaration ¶ 16 ("In particular, it had been known that milk contains proteases which could destroy proteins during their purification process."). But Gordon et al. (the "Gordon article"),8 upon which Dr. Hennighausen relied, stated that proteins expressed by transgenic milk production were "sufficiently stable in milk." Gordon article at 1185. The Board concluded that the potential difficulties with successfully using transgenic animal milk to produce proteins postulated by Dr. Hennighausen, as well as the potential difficulties with variable levels of protein production, were entitled to little weight in assessing likelihood of success, in view of the references cited for corroboration and the language of claim 65 ("[a] transgenic non-human female mammal that produces recoverable amounts of biologically active human fibrinogen... in its milk"). Velander, slip op. at 13-14.

Likewise, the Board determined that Dr. Rosen's declarations were entitled to little weight beyond the corroborating references cited. Id. at 17. The Board, however, did find Dr. Rosen's statements credible when he stated, consistent with prior-art publications, that

in 1993, it would [have been] obvious to combine [the DNA sequences encoding the three fibrinogen chains and regulatory sequences requisite for expression in the mammary gland] to target expression to the mammary gland. The only issue would have been which promoter or regulatory sequences to use from the known promoters and regulatory sequences.

Id. The Board found that Dr. Rosen's statement that fibrinogen was "more complex than any transgenic expression previously attempted" was credible to the extent it meant that fibrinogen involved assembly of more protein subunits than other proteins that had been expressed transgenically. Id. at 17-18.

The Board considered, but then discounted, Dr. Rosen's testimony about a lack of reasonable expectation of success because he did not explain how the references that he cited, the Meade patent, the Clark application, and the Rosen patent, supported his opinion that a skilled worker would not have expected to succeed. Id.

The Board also considered Dr. Rosen's testimony that the expression of transgenes was highly variable; it discounted this testimony, however, insofar as it applied to the Velander application. Id. at 18. The Board stated that Velander's claims only require the expression of "more than minuscule trace amounts." Id. at 7. The Board stated that Dr. Rosen's testimony about the "variability" in expression level did not mean that the skilled worker would have expected no transgenic fibrinogen to be produced. Id. at 18.

Additionally, the Board considered Dr. Rosen's testimony that cell lines could not be used to predict the possible adverse effects associated with transgenic expression in the mammary glands of animals. The Board found that Yarus et al. (the "Yarus article"),9 the only prior-art publication upon which Dr. Rosen relied, did not support Dr. Rosen's opinion. Id. First, the Board noted that the Yarus article was published after the critical date. The Board also noted that the Yarus article reported two cases in which adverse effects had been observed in connection with the transgenic expression of hormones.10 The Board concluded that one of skill in the art would not have considered any of these possible problems insurmountable for transgenic fibrinogen production. Id.

Finally, the Board analyzed Dr. Toman's declaration, in which he stated that expression in transgenic animals is variable and that cell lines are not predictive of expression problems in transgenic animals. Id. at 19. The Board concluded that "Toman's conclusion that the art was unpredictable and offered little chance of success ... must be read in light of Toman's focus on predictable expression levels and the lack of correlation with cell lines, neither of which are limitations of claim 65." Id. at 19. The Board concluded that "Toman's conclusion is less credible than the wealth of contemporaneous publications offering a contrary view." Id. Summarizing, the Board noted that

Velander['s] witnesses consistently point to the variability of expression to indicate that the art was not predictable, but also rely on that variability to discount Garner's contention that genomic DNA yields unexpectedly good results. The cited publications collectively indicate that, at the critical date, a person having ordinary skill in the art would have expected considerable variability in yields (including no yield), even among animals subjected to the same protocol. A person having ordinary skill in the art would have viewed such variability as an indication of the expense, time, and effort involved in producing transgenic mammals and selecting those with "recoverable" yields, but not as an indication that success would be unlikely.

Id. at 20.

Based upon its analysis, the Board granted Garner's motion, concluding that Velander's claims 64-73 from the '184 application directed to transgenic mammals capable of expression of fibrinogen in their milk were unpatentable as obvious. The Board stated:

For this motion, the preponderance of the evidence supports the findings that the level of skill in the art was high; all of the elements of the invention, including the DNA, the promoters, and the experience with transgenic animals existed in the prior art; and the art provided ample motivation for one skilled in the art to attempt to produce transgenic animals engineered to express human fibrinogen. One skilled in the art would have expected the process to be challenging, expensive, time-consuming, and tedious, but would not have expected the process to require undue experimentation. Furthermore, one skilled in the art would have been well acquainted with the inadequacy of cell cultures as a basis for predicting success in transgenic animals and thus would not have counted on such results to provide much encouragement or discouragement. Finally, absolute predictability is not a requirement for obviousness.

Id. at 21.

Velander timely moved for reconsideration of the Board's decision. In its motion, it argued that the Board had "misapprehended or overlooked" its duty to specify the prior-art references upon which it was relying in coming to its conclusion of obviousness. Velander also argued that the Board had failed to identify references that would support a motivation to combine the elements of the invention found in the prior art. Velander further argued that the Board had erred because it "did not point to any prior art publication that could be construed as evidencing the reasons for expecting success in the transgenic production of fibrinogen, as claimed."

Addressing Velander's motion, the Board concluded that "[t]he findings Velander submits are lacking in the decision on motions are in fact present...." Garner v. Velander, Reconsideration Decision slip op. at 9. Turning to Velander's contention that there was no "prior art publication that could be construed as evidencing reasons for expecting success in the transgenic production of fibrinogen," the Board noted that the decision on the motions contained the following findings based on the statement of material facts in Garner's preliminary motion 2 (Paper No. 22):

m. As of the critical date, biocompetent human fibrinogen had been produced using cultured mammalian cells transfected with the cDNAs encoding the three [fibrinogen] chains (Paper No. 22 at 7, ¶ 12 (uncontested)).

n. As of the critical date, several introduced proteins had been produced in the milk of transgenic mammals. The prior art included teachings or suggestions for the production of blood and serum proteins in the milk of transgenic mammals (Paper No. 22 at 10-11, ¶ 16).

Id. at 5. The Board noted that the references cited in Garner's Paper No. 22 in paragraph 16 of the statement of facts were the Meade patent, the Rosen patent, and the Clark application. On reconsideration, the Board again found that all of the elements of Velander's claims were present in the prior art. The critical inquiry then reduced to whether there was, from the prior art, a motivation to combine these elements and a reasonable expectation of success in producing transgenic animals capable of expressing the protein fibrinogen in biologically active form if those elements were combined. Velander noted the Board's discussion of the Hennighausen Review and the Meade patent but urged that these references related only to "motivation" to combine. Id. at 9. The Board responded that motivation to combine and reasonable expectation of success are "logically associated and thus it should not be surprising that they are discussed together." Id. The Board went on to point out that the Hennighausen Review touted the use of transgenic animals for the production of other blood serum proteins analogous to fibrinogen, although admittedly less complex, and that the Meade patent described the production of a blood protein in transgenic mammal milk. In light of this analysis, the Board found that "Velander has not justified relief from the order granting Garner preliminary motion 2, which held all of Velander's involved claims to be unpatentable." Id. Velander now appeals to us. We have jurisdiction pursuant to 28 U.S.C. § 1295(a)(4)(A).

Velander makes essentially three arguments on appeal. First, it contends that the Board erred in its allocation of the burden of proof. Velander asserts that the Board improperly placed on it, with respect to the prior art, the burden of proving "an expectation of failure," rather than requiring Garner to prove a reasonable expectation of success to substantiate obviousness. Second, it contends that the Board erred in refusing to consider the testimony of its independent experts. According to Velander, their testimony provided substantial evidence in support of the proposition that the prior art did not give rise to a reasonable expectation of success in expressing fibrinogen in mammary gland cells in vivo. Third, Velander argues that substantial evidence does not support the Board's finding that, to one of ordinary skill in the art, the prior art provided a reasonable expectation of success in expressing fibrinogen in mammary gland cells in vivo. Accordingly, Velander urges, the Board's ultimate conclusion of obviousness is incorrect as a matter of law.

As noted above, before the Board, Garner was required to establish by a preponderance of the evidence that the claims of the Velander application were unpatentable. The preponderance of the evidence standard requires the trier of fact "to believe that the existence of a fact is more probable than its nonexistence before [he] may find in favor of the party who has the burden to persuade the [judge] of the fact's existence." Bosies v. Benedict, 27 F.3d 539, 542 (Fed.Cir.1994) (quoting In re Winship, 397 U.S. 358, 371-72, 90 S. Ct. 1068, 25 L. Ed. 2d 368 (1970) (alterations in original) (citations omitted) (Harlan, J., concurring)). Velander argues that instead of requiring Garner to prove that the prior art showed a reasonable expectation of success, the Board improperly shifted the burden of proof by requiring Velander to establish that the prior art showed an expectation of failure. Velander asserts that the Board "clearly did not require Garner to prove that the prior art showed a reasonable expectation of successfully expressing fibrinogen in mammary cells in vivo, for nowhere does the Board state in its decision that Garner did so."

Velander's burden of proof argument is without merit. To begin with, the Board clearly understood that Garner had the burden of proof and what that burden was. In its August 16, 2001 decision, the Board stated, "Garner bears the burden of justifying the relief it seeks." Velander, slip op. at 21. Then, in the very next sentence, the Board stated that "the preponderance of the evidence" supported the pertinent findings of fact bearing on obviousness. Id.

Neither did the Board, while correctly articulating the burden of proof, nevertheless misapply it, as Velander suggests. As Garner points out, the Board specifically found that Garner had proven the following facts by a preponderance of the evidence: (1) the level of skill in the transgenic art was high; (2) all of the elements of the '184 application existed in the prior art; and (3) the prior art provided ample motivation for the skilled worker to produce transgenic animals engineered to express human fibrinogen. Id. In addition, we do not think that the Board required Velander to prove an expectation of failure, as argued by Velander. We agree with Garner that what happened was that the Board simply found unpersuasive Velander's rebuttal testimony that there was no reasonable expectation of success. See WMS Gaming, Inc. v. Int'l Game Tech., 184 F.3d 1339, 1359 (Fed.Cir.1999) (Objective evidence of non-obviousness may be used by a patentee "to rebut a prima facie case of obviousness based on prior art references.").

For example, the Board considered the 1997 Hennighausen article and determined that it was of limited probative value in assessing nonobviousness, except to the extent that it showed either (i) prior art knowledge of some of the problems associated with transgenic milk protein expression or (ii) that the problems it described actually applied to the transgenic production of fibrinogen. Velander, slip op. at 13. The Board concluded that the article failed to describe any specific problem applicable to the transgenic expression of fibrinogen, and that the two problems mentioned in the article (i.e., large-scale production in livestock and problems with post-translational modifications at high expression levels) were not relevant to the claim at issue in the Velander application because the claim was not limited to livestock or a high level of production. Id. at 13-14. Based upon its analysis, the Board concluded that the article did not rebut Garner's case. The Board did not misplace the burden of proof on Velander.

Additionally, the Board did not require Velander's experts to prove an expectation of failure, as argued by Velander. Rather the Board simply looked for an objective basis for the opinions of the experts with respect to the asserted lack of a reasonable expectation of success. Because the Board found no objective basis for the opinions of Velander's experts, the Board gave these declarations little probative weight as rebuttal to Garner's evidence.

Velander argues that the Board erred by refusing to consider the testimony of its expert witnesses, Drs. Rosen, Hennighausen, and Toman. According to Velander, this error was prejudicial because the expert testimony that the Board disregarded provided "substantial evidence" that there was no reasonable expectation in the prior art that fibrinogen could be successfully expressed in mammary gland cells in vivo. In making this argument, Velander points, inter alia, to the Board's statement that the "broad conclusions" of Velander's experts were "entitled to little weight beyond the corroborating references" the experts cited. Id. at 17. Velander also points to the Board's statement that the "litigation-driven testimony" of one of its experts was to be accorded less weight than his "ante litem motam writings."11 Id. Garner responds that the Board did not improperly disregard any of the expert testimony that was offered by Velander. It asserts that the Board did exactly what a fact finder must do, consider all of the evidence before it and accord the proper and appropriate weight to the evidence.

We agree with Garner that the Board did not fail to consider the testimony of Velander's expert witnesses. As seen above, in the case of both Garner's and Velander's experts, what the Board consistently did was accord little weight to broad conclusory statements that it determined were unsupported by corroborating references. It is within the discretion of the trier of fact to give each item of evidence such weight as it feels appropriate. See Newell Cos., Inc. v. Kenney Mfg. Co., 864 F.2d 757, 787-88 (Fed.Cir.1988) ("Determining the weight and credibility of the evidence is the special province of the trier of fact. The trier of fact must not only identify the prior art, its scope and content, but it must also weigh all the evidence, impose the viewpoint of the person of ordinary skill, and determine if the burden of proof has been met.") (citation omitted). In giving more weight to prior publications than to subsequent conclusory statements by experts, the Board acted well within that discretion. See Ashland Oil, Inc. v. Delta Resins & Refractories, Inc., 776 F.2d 281, 294 (Fed.Cir.1985) ("Lack of factual support for expert opinion going to factual determinations, however, may render the testimony of little probative value in a validity determination."); Carella v. Starlight Archery & Pro Line Co., 804 F.2d 135, 138 (Fed.Cir.1986) ("Although in some circumstances unsupported oral testimony can be sufficient to prove prior knowledge or use, it must be regarded with suspicion and subjected to close scrutiny.").

Velander asserts that the Board ignored Dr. Rosen's declaration and deposition testimony that the prior art did not show a reasonable expectation of success of transgenically expressing fibrinogen in the milk of animals. The Board, however, did consider all of Dr. Rosen's declaration and oral testimony, finding those portions of his testimony that were consistent with the scientific literature to be probative and credible. Velander, slip op. at 17. Because it was consistent with the scientific literature, the Board did find credible Dr. Rosen's statement that

in 1993, it would [have been] obvious to combine [the DNA sequences encoding the three fibrinogen chains and regulatory sequences requisite for expression in the mammary gland] to target expression to the mammary gland. The only issue would have been which promoter or regulatory sequences to use from the known promoters and regulatory sequences.

Id. The Board also found that Dr. Rosen's statement that fibrinogen was "more complex than any transgenic expression previously attempted" credible to the extent it meant that "fibrinogen involve[d] the assembly of more protein subunits than the other proteins" that had been expressed transgenically. Id. at 17-18.

The Board considered, but then discounted, Dr. Rosen's testimony about a lack of reasonable expectation of success because he did not explain how the references that he cited, the Meade patent, the Clark application, and the Rosen patent, supported his opinion that a skilled worker would not have expected to succeed.

The Board also considered Dr. Rosen's testimony that the expression of transgenes was highly variable but discounted this testimony as it applied to Velander's claims. Id. at 18. Velander's claims only require the expression of "more than minuscule trace amounts." Id. at 7. The Board stated that Dr. Rosen's testimony about the "variability" in expression level did not mean that a skilled worker would have expected no transgenic fibrinogen to be produced. Id. at 18.

Finally, the Board considered Dr. Rosen's testimony that cell lines could not be used to predict the possible adverse effects associated with transgenic expression in the mammary glands of animals. The Board found that the Yarus article, the only document relied on by Dr. Rosen, did not support his opinion. Id. The Board noted that the Yarus article reported two cases in which adverse effects had been observed, each relating to the transgenic expression of hormones. Because fibrinogen is not a hormone and has no potent biological activity,12 the Board concluded that one of skill in the art would not have considered any of these possible problems insurmountable for transgenic fibrinogen production. Id.

Velander also asserts that the Board disregarded Dr. Hennighausen's testimony in deciding that there was a reasonable expectation of success in expressing fibrinogen in transgenic animals as of the critical date. The Board did consider Dr. Hennighausen's testimony, however, accepting it to the extent that it was corroborated by the scientific literature. The Board considered Dr. Hennighausen's testimony that protein recovery could be difficult for some proteins, but found that he did not explain why the skilled worker would have expected the recovery of fibrinogen to be problematic. In addition, the Board noted that Dr. Hennighausen's own review article as well as Houdebine13 suggested that any possible problem in protein recovery would be readily solvable. Id. at 14.

Additionally, the Board considered Dr. Hennighausen's testimony that transgenes could cause deleterious effects on the transgenic animal or its offspring. Id. at 15. The Board, however, discounted this testimony because it was not supported by the documents to which Dr. Hennighausen pointed: Burdon et al.,14 Wall et al.,15 and Shamay et al.16 Each of these references taught that the expression of a specific milk protein, whey acidic protein ("WAP"), could lead to impairment or loss of lactation. The Board found Dr. Hennighausen's testimony regarding the deleterious effects of transgenes unpersuasive because none of the cited references taught that "the technology was unworkable" and because Dr. Hennighausen failed to provide "any basis for generalizing [the WAP] problem to other proteins." Id. at 14-15.

The Board also considered Dr. Hennighausen's testimony that cell cultures do not necessarily correlate with transgenic animal results. Id. at 15-16. Again, however, the Board discounted the testimony because it determined that the references cited by Dr. Hennighausen, Furth et al.17 and Whitelaw et al.,18 did not support his opinion. Those references taught that expression enhancers do not necessarily operate in the same way in cell culture and transgenic animals. In his deposition, however, Dr. Hennighausen acknowledged that in vitro cell culture data are at least predictive of protein expression, if not activity and secretion.

Finally, the Board considered Dr. Hennighausen's testimony that purification of a protein transgenically produced in milk may be problematic due to the presence of proteases in milk. Id. at 16. The Board gave this testimony little weight because the Gordon article, see footnote 8 above, the reference to which Dr. Hennighausen pointed, did not support his declaration. Id. at 16. The Gordon article taught that tissue plasminogen activator ("tPA") was successfully expressed and "sufficiently stable in milk." Similarly, according to the Hennighausen Review, various other blood proteins, including blood coagulation factor IX, α1-antitrypsin, and protein C, had been successfully produced transgenically in milk. Finally, Clark et al.19 had reported that several blood proteins of biomedical importance, such as α1-antitrypsin, tPA, and antithrombin III, existed naturally in milk. We cannot say the Board abused its discretion in giving more weight to the scientific literature than to Dr. Hennighausen's testimony.

Finally, Velander contends that the Board incorrectly criticized Dr. Toman's declaration and ignored Dr. Toman's deposition testimony. As in the case of Drs. Rosen and Hennighausen, however, the Board did consider Dr. Toman's testimony. It concluded that the testimony was "starkly at odds in tone with the contemporaneous literature" and "less credible than the wealth of contemporaneous publications...." Id. at 19. The Board noted that although Dr. Toman indicated that one of skill in the art would not have had an expectation of success due to the relative complexity of fibrinogen, he did not point to anything in the scientific literature supporting that view.

As can be seen from the discussion of the Board's consideration of the testimony of Velander's experts, Velander's complaint really is with the way in which the Board, as the trier of fact, weighed and assessed the evidence. That complaint goes to the sufficiency of the evidence. It is to that issue that we now turn.

On appeal, we review the Board's ultimate conclusion of obviousness without deference, while we review the Board's underlying factual determinations for substantial evidence. In re Gartside, 203 F.3d 1305, 1316 (Fed.Cir.2000). Substantial evidence "means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion." Consol. Edison Co. v. NLRB, 305 U.S. 197, 229, 59 S. Ct. 206, 83 L. Ed. 126 (1938). Our review of the Board's factual findings for substantial evidence examines the record as a whole, taking into account evidence that supports as well as detracts from those findings. Gartside, 203 F.3d at 1312. "The possibility of drawing two inconsistent conclusions from the evidence," however, will not render the Board's findings unsupported by substantial evidence. Consolo v. Fed. Mar. Com'n, 383 U.S. 607, 620, 86 S. Ct. 1018, 16 L. Ed. 2d 131 (1966). In other words, if the evidence of record will support several reasonable but contradictory conclusions, we will not find the Board's decision unsupported by substantial evidence because the Board chose one finding over another plausible alternative. In re Jolley, 308 F.3d 1317, 1320 (Fed.Cir.2002). Here, the Board determined that the preponderance of the evidence supported the conclusion that claims 64-73 of the '184 application were obvious as of the critical date based on the following findings: (1) all the elements of Velander's claim 65, and by extension all of Velander's involved claims and all of Garner's claims, were in the prior art; (2) there was a motivation in the prior art to combine those elements; and (3) one of ordinary skill in the art would have had a reasonable expectation of success in generating a recoverable amount of biologically active fibrinogen.

Velander's main argument on appeal is that the decision of the Board is not supported by substantial evidence. Velander does not dispute that all of the elements of claim 65 were in the prior art and that there was a motivation to combine those elements. Velander argues, however, that as of February 18, 1993, one of ordinary skill in the art would not have had a reasonable expectation of success in producing a recoverable amount of biologically active fibrinogen from a "transgenic non-human female mammal that produces recoverable amounts of biologically active human fibrinogen ... in its milk," as required by claim 65 of the '184 application. Velander argues that the Hennighausen Review and the Meade patent only showed a motivation to combine and not a reasonable expectation of success, and that the Rosen patent, the Clark application, and the Greenberg article did not establish a reasonable expectation of success.

Velander argues that the Hennighausen Review does not demonstrate a reasonable likelihood of success for several reasons: (1) because fibrinogen is not listed in the Review in the "table of foreign proteins that had been expressed in the milk of transgenic animals before 1990"; (2) because "[n]one of the proteins listed in [the table] is comparable to fibrinogen in size or complexity"; and (3) because the article teaches that there are "numerous problems, complications and uncertainties in producing foreign proteins in transgenic animals." Essentially, as it did before the Board, Velander urges that fibrinogen is more complex than other proteins successfully expressed in transgenic systems and that, as of the critical date, there were various problems in connection with the transgenic expression of proteins in milk.

As far as the Meade patent is concerned, Velander argues that the patent only provides a general disclosure for producing foreign proteins in the milk of transgenic animals and that its only example is tPA, which Velander argues would give no reasonable expectation of success because tPA is five times smaller than fibrinogen. Velander concedes that Meade suggests the use of the transgenic technique for a number of proteins, including serum proteins, but points out that fibrinogen is not specifically mentioned. Velander also argues that Meade cannot teach an expectation of success in combination with the Hennighausen Review, as stated in the Board's reconsideration decision, because the original decision on the motions acknowledged that cell culture expression of fibrinogen was not predictive of the success of the expression in transgenic animals. Velander, slip op. at 16, 20, 21.

Velander contends that the remaining references relied on by Garner also do not provide substantial evidence to support the Board's decision. In arguing that the Rosen patent, which teaches expressing proteins in a transgenic animal generally, does not provide substantial evidence for the Board's finding of a reasonable likelihood of success, Velander argues that the patent "nowhere states or suggests that multi-chain proteins such as fibrinogen could be successfully expressed." Velander also argues that Dr. Rosen's statement that he "had no idea of making fibrinogen or any other protein like it in this patent" shows that this reference does not provide evidence for the Board's conclusion. Velander contends that the Clark application, like both the Meade and Rosen patents, teaches expressing proteins in a transgenic animal generally, that the only examples in the application are directed to preparing DNA constructs encoding Factor IX and ¥1-antitrypsin, as well as fusion proteins containing the β-lactoglobulin gene, and that nowhere in the reference is there the suggestion that fibrinogen could be successfully expressed. Finally, Velander asserts that the Greenberg article, co-authored by Dr. Rosen, provides support, along with Dr. Rosen's deposition testimony and declaration, for why Dr. Rosen, in light of his experience with FSH (a multimeric protein expressed transgenically), believed that the transgenic expression of FSH in milk was not predictive of a successful outcome for transgenically expressing fibrinogen.

Garner responds that the decision of the Board is supported by substantial evidence. It states that, prior to the critical date, the prior art suggested that transgenic animals would be useful in producing plasma proteins in their milk. Garner cites to the Meade patent, the Rosen patent, the Clark application, and the Greenberg article as demonstrating that, as of the critical date, a number of heterologous proteins had been expressed in transgenic animals using generally accepted transgenic techniques and that successful transgenic production of proteins was normal in the art. Garner also points to human proteins with varying structures being produced in the milk of transgenic animals prior to the critical date. These proteins included blood coagulation Factor IX, α1-antitrypsin, interleukin 2, tPA, growth hormone, protein C, cystic fibrosis transmembrane conductance regulator protein, serum albumin, CD4, γ-interferon, urokinase, and FSH.20

Garner's principal point is that one of skill in the art would have expected success because fibrinogen shares a number of important characteristics with many of the previously produced proteins that had been expressed in the milk of transgenic animals as of the critical date. Thus, Garner notes (1) that fibrinogen is a blood protein, similar in that respect to Factor IX, α1-antitrypsin, tPA, protein C, serum albumin, and urokinase; (2) that fibrinogen is a multimer requiring post-translational modifications, including multimer assembly, in order for it to be active, similar to FSH and Protein C; and (3) that fibrinogen requires assembly from several different monomers, which are encoded by several different DNA sequences, to form its biologically active multimeric complex, similar to FSH. All of these proteins had been successfully expressed in the milk of transgenic animals prior to the critical date. Garner urges that there was no reason for one of ordinary skill in the art to expect anything other than success in the production of at least some (i.e., a recoverable amount) of transgenic fibrinogen in mammary gland cells in vivo.

In order to affirm the Board's decision, we must be convinced that substantial evidence supports the Board's conclusion that Garner established by a preponderance of the evidence that the claims of the Velander application were unpatentable. As noted, Velander does not dispute that all of the elements of claim 65 of the '184 application were in the prior art. Neither does it dispute that there was a motivation to combine those elements. Thus, as seen, the case boils down to the question of whether, as of the critical date, one of ordinary skill in the art would have had a reasonable expectation of success in producing a recoverable amount of biologically active fibrinogen from a "transgenic non-human female mammal that produces recoverable amounts of biologically active human fibrinogen ... in its milk," as required by claim 65. What that means for us is that we must decide whether there is (i) relevant evidence that a reasonable mind might accept as adequate to support a conclusion (substantial evidence) that (ii) supports the Board's conclusion that Garner established that it was more probable than not (a preponderance of the evidence) that, as of the critical date, one of ordinary skill in the art would have had a reasonable expectation of success in generating a recoverable amount of biologically active human fibrinogen. This is a close case. However, at the end of the day, we cannot say that Velander has established that the Board's decision is not supported by substantial evidence.

We have set forth in considerable detail the evidence that was before the Board and the contentions of the parties with respect to that evidence. As discussed, it is Velander's position that the complexity of fibrinogen and the difficulties and uncertainties that would necessarily be associated with the production of human fibrinogen in a transgenic system bar the conclusion that, as of the critical date, there was the required reasonable expectation of success. However, the Board came to grips with this point when it stated:

The cited publications collectively indicate that, at the critical date, a person having ordinary skill in the art would have expected considerable variability in yields (including no yield), even among animals subjected to the same protocol. [That person] would have viewed such variability as an indication of the expense, time, and effort involved in producing transgenic mammals and selecting those with "recoverable" yields, but not as an indication that success would be unlikely.

Id. at 20.

As argued by Garner and mentioned above, fibrinogen shares a number of important characteristics with many of the previously produced proteins that had been expressed in the milk of transgenic animals: (1) fibrinogen is a blood protein, similar, in that respect, to Factor IX, α1-antitrypsin, tPA, protein C, serum albumin, and urokinase;21 (2) fibrinogen is a multimer requiring post-translational modifications including multimer assembly in order for it to be active, similar to FSH and Protein C;22 and (3) fibrinogen requires assembly from several different monomers, which are encoded by several different DNA sequences, to form its biologically active multimeric complex, similar to FSH.23 Garner's references present substantial evidence supporting the Board's finding that one of ordinary skill in the art would have had a reasonable expectation of success in the production of at least some (i.e., a recoverable amount) of transgenic fibrinogen in mammary gland cells in vivo. That is because, as of the critical date, proteins with similar characteristics had successfully been produced in transgenic milk systems.

As noted, Velander argues that the Hennighausen Review and the Meade patent cannot be used to show a reasonable likelihood of success because fibrinogen is not specifically listed in either of those publications and because of the significant problems, complications, and uncertainties described in these publications for transgenic milk protein production. The Board recognized the problems associated with transgenic milk protein production, finding that "[t]he review creates the impression that producing transgenic mammalian livestock is expensive, technically challenging, and laborious, but nevertheless a viable option for producing proteins." Id. at 12. Velander also argues that the 1997 Hennighausen article showed that one of skill in the art would not have had a reasonable expectation of success. The Board addressed this argument, stating:

Obviousness, and expectation of success, are evaluated from the perspective of a person having ordinary skill in the art at the time of the invention. While later publications may explain what was known earlier, it would be wrong to impute later-recognized insights — or possible obstacles — to the knowledge available to those skilled in the art at the time of the invention.... One skilled in the art could not be daunted by unknown obstacles although the subsequent publications might be relevant to show that the obstacles actually applied to the specific problems facing the inventor.

Id. at 13.

The Board pointed out the Hennighausen Review's discussion of the advantages of transgenic milk production for proteins over production using cell culture systems, identifying serum proteins, especially blood clotting factors, as ideal proteins for development. Id. at 11-12. The Board stated that, according to the Review, "success had already been achieved for several human proteins, including serum proteins like protein C, but at widely variable and generally low expression levels (pp. 5-6)." Id. at 11. The Board stated that "[t]he review creates the impression that producing transgenic mammalian livestock is expensive, technically challenging, and laborious, but nevertheless a viable option for producing proteins." Id. at 12. While the Board recognized the problems outlined in the Hennighausen Review in connection with transgenic expression systems, especially transgenic expression of complex proteins like fibrinogen, it also recognized that claim 65 of the Velander application only required "`recoverable amounts' [which] in context means more than minuscule trace amounts." Id. at 7.

The Board also noted that Dr. Colman cited the Meade patent as showing a reasonable expectation of success and that "Meade provides motivation to switch from cell cultures to transgenic milk production, noting cost advantages (1:15-21) and increased reliability (1:41-45), and (vis-à-vis prokaryotes) proper post-translational modifications (1:36-41)." Id. at 12. The Board pointed out that "Meade identifies many proteins subject to its method, including serum proteins (3:31-40)." Id. In view of the fact that claim 65 is limited to only a recoverable amount of biologically active fibrinogen, the Board held that one of ordinary skill in the art would have had a reasonable expectation of success of producing a recoverable amount of fibrinogen, given the success with other proteins in transgenic systems, the presence of all the elements of the invention in the prior art, and the motivation to combine conceded by Velander.

Velander also argues that the remaining references relied on by Garner (i.e., the Rosen patent, the Clark application, and the Greenberg article) do not provide substantial evidence to support the Board's finding that one of skill in the art would have had a reasonable likelihood of success principally because these references do not mention fibrinogen or a protein as complex as fibrinogen. It is true that the Rosen patent, the Clark application, and the Greenberg article do not mention fibrinogen. However, it also is true that the Rosen patent, the Clark application, and the Greenberg article do demonstrate that, as of the critical date, a number of heterologous proteins had already been expressed in transgenic animals using generally accepted transgenic techniques, and that successful transgenic production of proteins was normal in the art.

Understandably, Velander directs our attention to the evidence in the record that discusses the difficulties in transgenic expression of complex proteins like fibrinogen. Such material arguably supports a conclusion contrary to the one reached by the Board. At the same time, other evidence in the record supports the conclusion reached by the Board. If the evidence will support several reasonable but contradictory conclusions, we will not find the Board's decision unsupported by substantial evidence simply because the Board chose one conclusion over another plausible alternative. Jolley, 308 F.3d at 1320. That is the case here. In other words, in this complex case, it is not for us to second-guess the Board's assessment of the evidence. Our task is to determine whether substantial evidence supports the conclusion chosen by the Board. We hold that substantial evidence supports the Board's conclusion that Garner established by a preponderance of the evidence that, as of February 18, 1993, one of ordinary skill in the art would have believed that there was a reasonable expectation of producing a recoverable amount of biologically active fibrinogen by successfully generating a "transgenic non-human female mammal that produces recoverable amounts of biologically active human fibrinogen ... in its milk."

For, the foregoing reasons, the decision of the Board granting the preliminary motion of Garner to have the claims of the Velander application declared unpatentable as obvious is affirmed.

No costs.

AFFIRMED.

GAJARSA, Circuit Judge, dissenting.

Although the majority opinion traces through a very unclear Board decision and tries with a substantial degree of specificity to supply a reasoned basis for the Board's decision, I respectfully dissent from the majority opinion because there is no reasoned basis for the Board's decision and there is no substantial evidence to support the PTO's finding of obviousness. SEC v. Chenery Corp., 332 U.S. 194, 195, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947); Colorado Interstate Gas Co. v. FPC, 324 U.S. 581, 595, 65 S. Ct. 829, 89 L. Ed. 1206 (1945). But see In re Huston, 308 F.3d 1267 (Fed.Cir.2002).

The Board held claims 64-73 of the Velander '184 application unpatentable for obviousness. Velander argues on appeal that there is no substantial evidence to support the Board's finding that one of ordinary skill in the art had a reasonable expectation of success regarding the production of fibrinogen in the milk of a transgenic mammal. Whether an invention is obvious is ultimately a legal conclusion. We must determine if "the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains." 35 U.S.C. § 103. The factual inquiries underlying obviousness include (1) the scope and content of the prior art, (2) the differences between the prior art and the claims at issue, (3) the level of ordinary skill in the art at the time the invention was made, and (4) any objective evidence of nonobviousness. Graham v. John Deere Co., 383 U.S. 1, 17-18, 86 S. Ct. 684, 15 L. Ed. 2d 545 (1966). "The consistent criterion for determination of obviousness is whether the prior art would have suggested to one of ordinary skill in the art that this process should be carried out and would have a reasonable likelihood of success, viewed in the light of the prior art." In re Dow Chem. Co., 837 F.2d 469, 473 (Fed.Cir.1988). Obviousness requires one of ordinary skill in the art have a reasonable expectation of success as to the invention — "obvious to try" and "absolute predictability" are incorrect standards. In re O'Farrell, 853 F.2d 894, 903 (Fed.Cir.1988). The presence of a reasonable expectation of success is measured from the perspective of a person of ordinary skill in the art at the time the invention was made. Life Techs., Inc. v. Clontech Labs., Inc., 224 F.3d 1320, 1326 (Fed.Cir.2000). To this end, the ultimate success of the invention is irrelevant. Id.

The Board's decision should pave the way to meaningful appellate review. In its decision, the Board is obligated not only to make the requisite findings based on the evidence of record, but also to explain the reasoning underlying its findings. In re Lee, 277 F.3d 1338, 1344-45 (Fed.Cir.2002). Our review of the Board's decision for substantial evidence should be restricted to the grounds relied on by the Board. See id. ("[R]eview of an administrative decision must be made on the grounds relied on by the agency. `If those grounds are inadequate or improper, the court is powerless to affirm the administrative action by substituting what it considers to be a more adequate or proper basis.'" (quoting Chenery, 332 U.S. at 196, 67 S. Ct. 1575)); see also In re Zurko, 258 F.3d 1379, 1385 (Fed.Cir.2001) (rejecting the PTO's argument for this court to consider additional references not relied on by the Board in reaching its conclusion of obviousness because the references did not support the Board's conclusion and, moreover, "would constitute a new ground for rejection, not considered or relied upon by the Examiner or the Board").

Here, the Board explicitly supported its finding of reasonable expectation of success on the Hennighausen review and Meade patent. In particular, the Board found persuasive the fact that both references discussed the production of blood proteins via transgenic animals:

The Hennighausen review in particular touts the use of transgenic animals for the very class of pharmaceutical proteins (blood serum proteins) to which fibrinogen belongs. Similarly, Meade obtained a patent for production of a blood protein in transgenic mammal milk. The problems identified in the art are not presented as insurmountable or unique to blood proteins.

The Board thus found that one of ordinary skill in the art1 would have believed, on the critical date,2 that there was a reasonable chance of producing fibrinogen in view of the fact that the expression of other blood proteins in transgenic animals was disclosed in the Hennighausen review and Meade patent.

This finding, however, was unsupported by substantial evidence because it was based on the Board's unsupported assumption, or alternatively, unsupported finding, as to the second Graham factor: the difference between the prior art and the claims at issue, as viewed from the vantage point of one of ordinary skill in the art. The Board's decision reveals its implicit assumption that one of ordinary skill in the art would have perceived the difference between the disclosed blood proteins and fibrinogen to be insignificant. Even assuming that this implicit assumption constituted an actual "finding" by the Board, it was unsupported by any evidence, let alone substantial evidence, that one of ordinary skill in the art would have agreed that the mere disclosure of blood proteins in the prior art would have led one of ordinary skill in the art to believe that fibrinogen could be expressed in transgenic non-human mammals. As we explained in Zurko:

As an administrative tribunal, the Board clearly has expertise in the subject matter over which it exercises jurisdiction.... With respect to core factual findings in a determination of patentability, however, the Board cannot simply reach conclusions based on its own understanding or experience — or on its assessment of what would be basic knowledge or common sense. Rather, the Board must point to some concrete evidence in the record in support of those findings. To hold otherwise would render the process of appellate review for substantial evidence on the record a meaningless exercise.

Zurko, 258 F.3d at 1385-86 (citing Baltimore & Ohio R.R. Co. v. Aberdeen & Rockfish R.R. Co., 393 U.S. 87, 91-92, 89 S. Ct. 280, 21 L. Ed. 2d 219 (1968)).

While we cannot step into the shoes of one of ordinary skill in the art, I believe that the Board's functional definition, accepted by the majority, of fibrinogen as a blood protein was misguided. The Board implies that the mere disclosure of the production of blood proteins, in general, via transgenic animals renders obvious the expression of any other blood protein in the same way, even where the target blood protein is so much more structurally complex that one of ordinary skill in the art would be hard-pressed to determine how to produce it via transgenic means. For example, the blood protein that could be encoded from multiple genes is large and requires post-translational processing that is not controlled by any human mammary genes. Potential problems faced by one of ordinary skill might include co-integration of the three genes encoding the fibrinogen protein into the foreign genome, proper assembly and bonding of the six chains of the protein once expressed, and proper secretion of the protein from the mammary gland into the milk of the lactating animal.3 The Board's classificatory framework of fibrinogen as blood protein seems irrelevant to resolving such potential issues.

The Board failed to determine (1) the difference, if any, between the prior art and the claims at issue as perceived by one of ordinary skill in the art, and (2) whether that difference, if any, is so significant that one of ordinary skill in the art would not entertain a reasonable expectation of success of expressing fibrinogen in transgenic non-human mammals.

For these reasons, I respectfully dissent.