Protein Similarity Score: A Simplified Version of the Blast Score as a Superior Alternative to Percent Identity for Claiming Genuses of Related Protein Sequences

Abstract

Recombinant proteins form the basis for most of the products of biotechnology, including drugs, diagnostics, research reagents, genetically modified organisms and industrial enzymes. However, the nature of proteins and the rules of patentability conspire to make it difficult to achieve adequate patent protection for novel proteins and the polynucleotides that encode them. Narrow patent claims limited to protein sequences sharing a high degree of structural identity can generally be designed around by introducing structural changes in the claimed protein, thereby avoiding the patent without substantially altering the protein's function. However, inventors are generally restricted in their ability to broadly claim protein inventions by the patent law's written description and enablement requirements.

In this paper, I propose a novel approach to claiming genuses of related protein sequences by means of what I refer to as a similarity score. The similarity score is essentially a modified version of the BLAST score, a standard approach used by biologists to measure the degree of structural similarity between related protein sequences. The primary advantage of the similarity score approach compared to conventional protein claiming techniques, such as by percent identity or hybridization, is that it is much more predictive of conserved function. The genus of related proteins defined by similarity score will include a substantially higher percentage of variants that retain function than a similarly sized genus defined using the conventional techniques. Policy objectives are served by allowing the inventor to better protect against trivial design around, while at the same time limiting the tendency of unduly broad patent claims to encompass proteins having substantially divergent structure and function.

Keywords: biotechnology, BLAST, similarity, patents, written description, enablement, percent identity, similarity score, proteins, DNA, sequence