PTAB

IPR2025-01355

Guardant Health Inc v. Cold Spring Harbor Laboratory

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Varietal counting of nucleic acids for obtaining genomic copy number information.
  • Brief Description: The ’589 patent discloses methods for determining genomic copy number information from nucleic acid samples. The invention purports to overcome amplification distortion by tagging nucleic acid segments with unique molecular tags prior to Polymerase Chain Reaction (PCR) amplification, which allows for a more accurate count of the original molecules in the sample.

3. Grounds for Unpatentability

Ground 1: Obviousness over Lo and POSA Knowledge - Claims 1-3, 6-8, 11, and 17-18 are obvious over Lo in view of the knowledge of a POSA, as reflected in McCloskey, Brenner, Chee, and Porreca.

  • Prior Art Relied Upon: Lo (Application # 2009/0029377), with POSA knowledge reflected in McCloskey (a 2007 journal article), Brenner (Patent 7,537,897), Chee (Patent 9,085,798), and Porreca (Patent 11,840,730).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Lo teaches the core method of determining copy number information: obtaining nucleic acid segments from a sample, sequencing them, mapping the sequences to a reference genome, and counting the sequences corresponding to specific genomic locations. Petitioner asserted that Lo discloses all elements of independent claim 1 except for the step of uniquely tagging nucleic acids before amplification to correct for PCR bias.
    • Motivation to Combine (for §103 grounds): Petitioner contended that correcting for amplification bias using unique molecular tags was a well-known and routine solution to a known problem in the field of molecular counting. The secondary references (McCloskey, Brenner, Chee, and Porreca) were presented as evidence of this common knowledge, each describing the use of unique tags or barcodes to enable accurate quantification of molecules post-amplification. A POSITA would combine this standard technique with Lo’s counting method to improve its accuracy and sensitivity, an obvious and desirable improvement.
    • Expectation of Success (for §103 grounds): A POSITA would have had a high expectation of success because molecular tagging was a straightforward, widely used technique. The ’589 patent itself acknowledges prior art methods for tagging, and the inventor of the ’589 patent admitted in a separate publication that modifying molecules to facilitate counting was "not a new idea."

Ground 2: Obviousness over Lo and Quake - Claims 4 and 5 are obvious over Lo in view of Quake.

  • Prior Art Relied Upon: Lo (Application # 2009/0029377) and Quake (Patent 8,195,415).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner asserted that Lo primarily teaches comparing sequence counts from a test sample against an external standard, such as data from a separate pool of normal genomes. Claims 4 and 5, in contrast, recite using an internal standard by comparing counts from a first genomic region to counts from a second genomic region within the same sample. Petitioner argued that Quake explicitly teaches this internal normalization method, where the genome is divided into bins and the number of sequence reads in a target region is compared to the median number of reads across other, presumed normal, regions of the same sample’s genome.
    • Motivation to Combine (for §103 grounds): A POSITA would combine Quake's internal standard with Lo's general framework to make the analysis more efficient and robust. Using an internal standard obviates the need to prepare and sequence separate external control samples, allowing all sequencing capacity to be dedicated to test samples and avoiding potential batch-to-batch variability associated with external controls.
    • Expectation of Success (for §103 grounds): Success was reasonably expected because Quake demonstrated that the internal standard method was effective, providing actual experimental data in its specification to prove the concept.

Ground 3: Error Correction Obviousness - Claim 13 is obvious over Lo in view of Chee or Porreca.

  • Prior Art Relied Upon: Lo (Application # 2009/0029377), Chee (Patent 9,085,798), and Porreca (Patent 11,840,730).
  • Core Argument for this Ground:
    • Prior Art Mapping: Claim 13 adds the limitation of ignoring tagged molecules that differ by only a single nucleotide during the counting step. Petitioner characterized this as a basic error-correction technique to account for mutations that may arise during PCR or sequencing. Petitioner argued that both Chee and Porreca teach methods for improving accuracy by generating consensus sequences and tuning the stringency for what constitutes a valid read, which includes protocols for handling or discarding reads with minor variations or errors.
    • Motivation to Combine (for §103 grounds): A POSITA would be motivated to incorporate such an error-correction step into the Lo-based method to improve the overall accuracy of the final molecular count. Discarding reads that are likely erroneous (e.g., single-nucleotide variants of a highly abundant tag) is a logical and obvious step to enhance the reliability of the quantitative data.
    • Expectation of Success (for §103 grounds): The technique amounts to a straightforward data-processing step of comparing sequences and filtering outliers based on a defined threshold. A POSITA would have had no trouble implementing such a commonly known error-correction strategy.
  • Additional Grounds: Petitioner asserted additional obviousness challenges, including that claim 9 (tags differing by more than one nucleotide) is obvious over Lo in view of Brenner or McCloskey; claims 12 and 14-16 (use of "sample tags" for multiplexing) are obvious over Lo in view of Brenner, McCloskey, or Porreca; and claim 10 (using ligation to attach tags) is obvious over Lo in view of Chee, Porreca, or Brenner.

4. Relief Requested

  • Petitioner requests institution of inter partes review and cancellation of claims 1-18 of Patent 10,947,589 as unpatentable.