Owen Grayson Hosseinzadeh, Law Clerk in San Diego, contributed to the drafting of this post.
The standard for written description in the life sciences seems to have tightened lately, leading patent applicants to wonder what level of written support and working example data is necessary to support genus claims. The recent BASF v. CSIRO decision provides yet another data point on how the Federal Circuit thinks about this issue. Although broad genus claims continue to be disfavored, this case demonstrates how applicants can maximize the effect of their early-stage data and, by extension, the scope of their patent claims.
The patents at issue in this case concern transgenic plants that express a series of enzymes not native to land plants. When introduced in oilseed crops (e.g., canola), these enzymes would allow the plants to transform a short-chain fatty acid they naturally produce (ALA) into the nutritionally beneficial long-chain polyunsaturated fatty acids found in fish and other marine life (EPA and DHA).
When it submitted its initial patent filings, the patentee had successfully implemented the technology in the model organism Arabidopsis, but it had not yet implemented the invention in canola or other oilseed plants.
In BASF, the Federal Circuit upheld a jury determination that three claims directed to a transgenic canola plant capable of producing long chain polyunsaturated fatty acids were adequately supported by working examples in the model organism Arabidopsis, but invalidated three genus claims covering all plants because they lacked sufficient written description.
In upholding the canola claims, the Federal Circuit cited the Arabidopsis working examples together with the prominence of canola in the specification. Three portions of the specification that expressly identify canola were particularly relevant: a list of preferred oilseed plants, a list of “plants of the invention,” and a list of oilseed crop plants.
The disclosure in the specification was supplementethd by trial evidence explaining the connection between the model organism Arabidopsis and the claimed plant, canola. Expert testimony established that (1) Arabidopsis and canola are both oilseeds (i.e., they store oil in the cotyledons), (2) the two species both naturally produce ALA, and (3) their shared ability to metabolize oils is due to their low evolutionary distance.
The Federal Circuit credited this testimony as demonstrating that proof of concept attained in Arabidopsis was “more than just a stepping stone” and “highly predictive of” achieving production of the desired oils in canola.
But the court’s reasoning could logically extend to additional plants. Other plants appear in the preferred embodiment lists and share relevant characteristics with Arabidopsis. The court’s discussion suggests that the Arabidopsis working examples, the specification, and the information explaining the connection between Arabidopsis and other oilseeds could support a subgenus claim covering at least canola, soy, and flax. Of course, there was no such subgenus claim examined by the court, so we cannot say for sure.
We do know that the court was unwilling to extend its reasoning to all plants. Despite its willingness to recognize the Arabidopsis data in support of the canola claims, the court clearly rejected the broad genus claim covering all plants: it found nothing in the specification or presented through expert testimony to suggest that success in Arabidopsis meant possession of the invention in all plant cells.
Although it invalidates yet another genus claim, BASF is not all bad news for patent applicants in the life sciences field. In fact, patent applicants should look to BASF as a guide for maximizing the power of proof of concept data obtained in model organisms.
From prior case law we know that proof of concept in one species can “serve as a stepping stone” to support patent claims directed towards different species, and that actual reduction to practice is not a requirement. But applicants would be wise to enhance the effectiveness of their data, creating “more than just a stepping stone,” by explicitly laying out the connection in the specification between the model organism and the desired species.
BASF also provides a lesson on the importance of drafting multiple layers of claims, particularly when an independent claim is directed to a large genus, such as “all plants.” Applicants should present subgenus and species claims of varying and overlapping scope to provide backup positions if a broad genus claim is found to be invalid.
By maximizing the power of model organism data through proper disclosure, and by including subgenus claims in an application, a savvy prosecutor can strategically avoid the two extremes of either narrowly claiming the exemplified subject matter or claiming a broad, unsupported genus.
 BASF Plant Sci., LP v. Commonwealth Sci. & Indus. Rsch. Org., 28 F.4th 1247 (Fed. Cir. 2022). The opinion was also featured as a “case of the week” on our Federal Circuity blog for its discussion of co-ownership issues arising from a collaboration agreement: https://federalcircuitry.mofo.com/topics/Last-Week-In-The-Federal-Circuit-March-21-2022-Federal-Circuit-Says-Go-Fish.html.
 Id. at 1267.
 Id. at 1265.
 Id. at 1268-69 (noting that CSIRO did not argue that success in Arabidopsis automatically mean success in all plant cells).
 Centocor Ortho Biotech, Inc. v. Abbott Labs., 636 F.3d 1341, 1349, 1353 (Fed. Cir. 2011).