PTAB

IPR2025-01434

Tempus Ai Inc v. Guardant Health Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Methods for Detecting Genetic Variants
  • Brief Description: The ’306 patent discloses methods for detecting genetic variants in cell-free DNA (cfDNA). The methods involve tagging double-stranded cfDNA molecules with duplex tags containing molecular barcodes, amplifying and sequencing the tagged molecules, and grouping the resulting sequence reads into families of paired and/or unpaired reads to improve accuracy.

3. Grounds for Unpatentability

Ground 1: Obviousness over Bielas and Vogelstein - Claims 1-2, 7-8, 12-13, 17-18, 20-21, 24-26, and 29 are obvious over Bielas in view of Vogelstein.

  • Prior Art Relied Upon: Bielas (WO 2013/123442) and Vogelstein (WO 2012/142213).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that the combination of Bielas and Vogelstein taught all limitations of independent claims 1 and 17. Bielas allegedly disclosed the core method of accurately detecting mutations by tagging double-stranded nucleic acids with duplex molecular barcodes (called "cyphers") attached to both ends of a molecule. Bielas further taught grouping sequence reads having identical barcode pairs, along with their reverse complements, into "families" of paired reads to generate a consensus sequence, thereby reducing errors. To the extent Bielas did not explicitly disclose processing unpaired reads, Vogelstein supplied this teaching by describing methods to reduce sequencing errors by grouping reads into families based on unique identifiers (UIDs), which a person of ordinary skill in the art (POSITA) would understand includes families of unpaired reads, particularly since not all original DNA strands are perfectly recovered during amplification.
    • Motivation to Combine: A POSITA would combine Bielas and Vogelstein to improve the accuracy of next-generation sequencing (NGS). Both references addressed the known problem of sequencing errors. A POSITA would have been motivated to apply Bielas’s superior error-correction method for paired reads while also incorporating Vogelstein’s techniques for handling the inevitable unpaired reads to maximize the use of sequencing data and further improve overall accuracy.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success, as both references described general error-correction techniques for NGS using standard molecular biology tools. Combining methods for processing both paired and unpaired reads was a predictable way to enhance data analysis without requiring unconventional chemistry or instrumentation.

Ground 2: Obviousness over Bielas, Vogelstein, and Forshew - Claims 3, 9-11, 19, and 22-23 are obvious over Bielas in view of Vogelstein and Forshew.

  • Prior Art Relied Upon: Bielas (WO 2013/123442), Vogelstein (WO 2012/142213), and Forshew (a 2012 journal article).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground builds on the base combination of Bielas and Vogelstein. Petitioner asserted that Forshew rendered obvious the dependent claims related to specific cfDNA input amounts (claims 3 and 19) and target enrichment (claims 9-11 and 22-23). Forshew, which focused on detecting cancer mutations in cfDNA, explicitly disclosed using cfDNA amounts (0.9 ng to 19.7 ng) that fall within the claimed range of 1 ng to 100 ng. Forshew also taught enriching for specific cancer-associated genes (e.g., TP53, EGFR, BRAF, KRAS) prior to sequencing, a well-known technique to improve efficiency when analyzing known regions of interest.
    • Motivation to Combine: A POSITA implementing the error-correction methods of Bielas and Vogelstein would have been motivated to incorporate the teachings of Forshew to optimize the process for cfDNA analysis. Since cfDNA is present in low concentrations, a POSITA would look to prior art like Forshew for established input amounts and would use well-known targeted enrichment techniques to focus sequencing on relevant cancer genes, making the overall method more efficient and cost-effective.
    • Expectation of Success: Success was expected because targeted enrichment and using specific nanogram-level inputs for cfDNA were routine and well-documented techniques in the field of cancer genomics by 2013.

Ground 3: Obviousness over Bielas, Vogelstein, and Hicks - Claims 14-16 and 27-28 are obvious over Bielas in view of Vogelstein and Hicks.

  • Prior Art Relied Upon: Bielas (WO 2013/123442), Vogelstein (WO 2012/142213), and Hicks (Patent 9,404,156).
  • Core Argument for this Ground:
    • Prior Art Mapping: Building on the base combination, Petitioner argued Hicks rendered obvious the dependent claims related to quantitative analysis (claims 14-15 and 27-28) and detecting copy number variation (CNV) (claim 16). Hicks taught methods for obtaining genomic copy number information by tagging DNA molecules, mapping sequence reads to a reference genome, and counting the number of different tagged molecules mapped to specific locations ("bins"). This counting provides a quantitative measure used to determine CNV.
    • Motivation to Combine: A POSITA using the methods of Bielas and Vogelstein to accurately identify mutations would have been motivated to incorporate the quantitative methods of Hicks. The ability to not only detect mutations but also to quantify the number of molecules and determine CNVs from the same sequencing data would provide a more comprehensive and powerful diagnostic tool for genetic conditions.
    • Expectation of Success: A POSITA would have expected success in applying Hicks's quantification methods to the families of reads generated by the Bielas/Vogelstein method, as both systems relied on using barcodes to track original molecules, making the integration of a counting step a predictable extension of the process.
  • Additional Grounds: Petitioner asserted an additional obviousness challenge against claims 4-6 based on Bielas, Vogelstein, and either Hendricks (WO 2012/099832) or Diehn (’081 application). This ground argued it was obvious to use a molar excess of tags to achieve specific ligation efficiencies, as taught by Hendricks and Diehn.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued that discretionary denial would be inappropriate because the petition raises new arguments and relies on prior art not substantively considered before. Specifically, Petitioner noted that a prior IPR on the ’306 patent (IPR2022-01400) resulted in a Final Written Decision (FWD) affirming patentability. However, Petitioner contended that the primary reference in this petition, Bielas, was not considered in that FWD. Petitioner argued Bielas directly teaches the key limitations the Board previously found missing in other art, namely attaching duplex tags to both ends of a cfDNA molecule and sorting reads into families of paired reads.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-29 of the ’306 patent as unpatentable.