Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 1 of 36 PageID #: 409
`
`IN THE UNITED STATES DISTRICT COURT
`FOR THE DISTRICT OF DELAWARE
`
`
`TWINSTRAND BIOSCIENCES, INC., &
`UNIVERSITY OF WASHINGTON,
`
`
`
`
`
`GUARDANT HEALTH, INC.,
`
`
`
`Plaintiffs,
`
`v.
`
`
`
`
`
`
`C.A. No. _________________
`
`JURY TRIAL DEMANDED
`
`Defendant.
`
`
`
`
`
`COMPLAINT
`
`Plaintiffs TwinStrand Biosciences, Inc. (“TwinStrand”) and University of Washington
`
`(“UW”) file this Complaint against Defendant Guardant Health, Inc. (“Guardant”), alleging as
`
`follows:
`
`NATURE OF THE ACTION
`
`1.
`
`This is an action for infringement of U.S. Patent Nos. 10,287,631 (“the ’631
`
`patent”); 10,689,699 (“the ’699 patent”); 10,752,951 (“the ’951 patent”); and 10,760,127 (“the
`
`’127 patent”) (collectively, “the Asserted Patents”) arising under the patent laws of the United
`
`States, 35 U.S.C. § 1 et seq. Guardant has infringed and continues to infringe the claims of the
`
`Asserted Patents by using, offering for sale, and selling its genetic-sequencing services in the
`
`United States.
`
`PARTIES
`
`2.
`
`TwinStrand is a corporation organized and existing under the laws of the State of
`
`Delaware and having its principal place of business at 3131 Elliott Ave., Suite 750, Seattle, WA,
`
`98121. At all relevant times, TwinStrand has been the exclusive licensee of the Asserted Patents.
`
`
`
`00001
`
`EX1078
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 2 of 36 PageID #: 410
`
`3.
`
`UW is a public institution of higher education and an agency of the State of
`
`Washington. Its principal place of business is in the city of Seattle, Washington. At all relevant
`
`times, UW has owned all right, title, and interest in the Asserted Patents.
`
`4.
`
`Guardant is a corporation organized and existing under the laws of the State of
`
`Delaware and having its principal place of business at 505 Penobscot Dr., Redwood City, CA
`
`94063.
`
`JURISDICTION AND VENUE
`
`5.
`
`This action arises under the patent laws of the United States, 35 U.S.C. §§ 1, et
`
`seq., and this Court has jurisdiction over the subject matter of Plaintiffs’ claims pursuant to
`
`28 U.S.C. §§ 1331, 1338(a), 2201, and 2202.
`
`6.
`
`This Court has jurisdiction over Guardant at least because Guardant is a Delaware
`
`corporation.
`
`7.
`
`This Court also has jurisdiction over Guardant because Guardant has purposefully
`
`availed itself of the rights and benefits of Delaware law by engaging in systematic and
`
`continuous contacts with Delaware, including by Guardant selling and offering for sale its
`
`infringing genetic-sequencing products in Delaware.
`
`8.
`
`Venue is proper in this District pursuant to 28 U.S.C. § 1400(b) because Guardant
`
`resides in Delaware as a consequence of its incorporation in the state.
`
`THE PATENTS-IN-SUIT
`
`9.
`
`On May 14, 2019, the United States Patent and Trademark Office lawfully issued
`
`the ’631 patent, entitled “Methods of Lowering the Error Rate of Massively Parallel DNA
`
`Sequencing Using Duplex Consensus Sequencing.” A true and correct copy of the ’631 patent is
`
`attached hereto as Exhibit A.
`
`- 2 -
`
`00002
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 3 of 36 PageID #: 411
`
`10.
`
`On June 23, 2020, the United States Patent and Trademark Office lawfully issued
`
`the ’699 patent, entitled “Methods of Lowering the Error Rate of Massively Parallel DNA
`
`Sequencing Using Duplex Consensus Sequencing.” A true and correct copy of the ’699 patent is
`
`attached hereto as Exhibit B.
`
`11.
`
`On August 25, 2020, the United States Patent and Trademark Office lawfully
`
`issued the ’951 patent, entitled “Methods of Lowering the Error Rate of Massively Parallel DNA
`
`Sequencing Using Duplex Consensus Sequencing.” A true and correct copy of the ’951 patent is
`
`attached hereto as Exhibit C.
`
`12.
`
`On September 1, 2020, the United States Patent and Trademark Office lawfully
`
`issued the ’127 patent, entitled “Methods of Lowering the Error Rate of Massively Parallel DNA
`
`Sequencing Using Duplex Consensus Sequencing.” A true and correct copy of the ’127 patent is
`
`attached hereto as Exhibit D.
`
`BACKGROUND
`
`13.
`
`The Asserted Patents cover groundbreaking duplex sequencing methods invented
`
`at UW by Jesse Salk, M.D., Ph.D., then a medical student, and two of his academic colleagues.
`
`Among many other applications, these duplex sequencing methods, for the first time, allowed for
`
`reliable, early, non-invasive cancer detection and post-treatment cancer monitoring in patients
`
`simply by analyzing blood plasma, without the need for biopsies of solid tumors. The inventions
`
`of the Asserted Patents can detect mutations in DNA target molecules that are present in
`
`extremely low abundance relative to the DNA from healthy cells—an elusive feat that previous
`
`sequencing methods could not achieve. The inventions of the Asserted Patents overcome the
`
`shortcomings in the prior art, offering unprecedented accuracy without sacrificing the high
`
`throughput of modern DNA sequencing approaches.
`
`- 3 -
`
`00003
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 4 of 36 PageID #: 412
`
`14.
`
`Following their invention, Dr. Salk and his co-inventors founded TwinStrand to
`
`make their inventions available to clinicians and researchers. TwinStrand exclusively licenses the
`
`Asserted Patents from UW and practices Duplex Sequencing through its sale of kits and services
`
`under its TwinStrand Duplex SequencingTM technology platform.
`
`A.
`
`15.
`
`The Need for High-Accuracy, High-Throughput Sequencing Methods
`
`Genetic mutations are the hallmark of cancer and other significant diseases
`
`affecting human health. Detecting the presence of these mutations in an individual was thought
`
`to be a promising way to screen for or diagnose cancers and other illness before individuals
`
`became symptomatic. But, in many instances, these hallmark genetic mutations are at an ultra-
`
`low frequency relative to the presence of DNA from healthy cells in a given sample, requiring
`
`the use of highly sensitive and specific genetic methods that did not exist before the inventions of
`
`the Asserted Patents.
`
`16.
`
`Conventional genetic-sequencing methods generally involve trade-offs among
`
`accuracy, throughput, and expense. For example, the Sanger sequencing method allowed
`
`scientists to complete the Human Genome Project, but that effort took decades and cost many
`
`millions of dollars. Sanger sequencing’s low throughput and high expense make it unsuitable for
`
`many applications. Moreover, Sanger sequencing approaches simply report the average sequence
`
`of a collection of many grouped molecules, obscuring low frequency mutations.
`
`17.
`
`Next Generation Sequencing (“NGS”) approaches sequence millions of individual
`
`DNA molecules at a time and offer much higher throughput at a fraction of the cost per DNA
`
`base compared to Sanger sequencing. But, conventional NGS approaches are still notoriously
`
`inaccurate. Indeed, conventional NGS approaches generate error rates of 0.1%–1%—meaning up
`
`to one in one hundred DNA bases are miscalled, and the presence of real biological mutations
`
`are obscured.
`
`- 4 -
`
`00004
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 5 of 36 PageID #: 413
`
`18.
`
`For many applications, the existing sequencing methods offered by Sanger and
`
`conventional NGS approaches were adequate. But before Dr. Salk’s inventions, neither could be
`
`used effectively in applications where the target DNA is at an ultra-low frequency, as is the case
`
`with early cancer detection using blood plasma or when looking for residual disease in a patient
`
`following a treatment course. Indeed, DNA from cancer cells is only present in blood plasma in
`
`extremely low concentrations; the overwhelming majority of DNA present comes from non-
`
`cancerous cells. To detect a target cancer mutation in that case, a sequencing method was needed
`
`that achieves high throughput and high sensitivity—something that conventional approaches
`
`simply could not deliver.
`
`B.
`
`The Inventions
`
`19. While at UW, Dr. Salk, then a medical student, and his colleagues invented
`
`breakthrough sequencing methods that achieved a 10,000-fold increase in accuracy over standard
`
`NGS approaches, without sacrificing throughput, eliminating nearly all technical errors
`
`introduced by NGS sequencing.
`
`20.
`
`These new sequencing inventions avoid the errors inherent in conventional NGS
`
`approaches by leveraging the duplicated information stored in each complementary strand of
`
`DNA. By labeling each original double-stranded target molecule, Dr. Salk and his team found
`
`that they could track each sequenced strand back to its original template molecule. And, by
`
`separately and uniquely labeling each complementary strand of that molecule, one strand of each
`
`molecule could be differentiated from the other strand of that molecule. This innovative labelling
`
`strategy allows for the comparison of complementary strands of the original target molecule.
`
`Because true mutations in a molecule are duplicated on both complementary strands, this
`
`comparison allows true mutations to be distinguished from sequencing errors, which occur on
`
`only one strand. Free of the performance trade-offs of prior sequencing approaches, Dr. Salk and
`
`- 5 -
`
`00005
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 6 of 36 PageID #: 414
`
`his co-inventors’ novel TwinStrand Duplex SequencingTM methods, claimed in the UW patents,
`
`enable the identification of rare genetic mutations that have very low frequencies among a
`
`population of target DNA molecules.
`
`21.
`
`The Duplex Sequencing methods conceived by Dr. Salk and his co-inventors
`
`eliminate essentially all of the background noise generated by sequencing errors in prior-art NGS
`
`sequencing methods—allowing for accurate identification of mutations that are present at an
`
`ultra-low frequency. The charts below compare the same gene sequenced with standard NGS
`
`sequencing with TwinStrand Duplex SequencingTM technology. With standard NGS sequencing,
`
`every position in the sequenced gene appears mutated in 0.1 to 1% of the molecules sequenced.
`
`In contrast, UW’s patented methods, embodied by the TwinStrand Duplex SequencingTM
`
`technology, remove this NGS noise to reveal the previously hidden, low-frequency true
`
`mutation.
`
`Standard NGS Sequencing
`
` TwinStrand Duplex SequencingTM
`
`“true” mutation
`previously lost
`in the “noise”
`
`
`
`TwinStrand Biosciences, “TwinStrand Duplex Sequencing™” technology brochure (2020)
`
`(Exhibit E).
`
`22.
`
`TwinStrand’s technology—built on the inventions of the Asserted Patents—
`
`uniquely provides the sensitivity and specificity necessary for accurate cancer detection and
`
`monitoring with non-invasive blood draws—liquid biopsies—resulting in dramatic
`
`- 6 -
`
`00006
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 7 of 36 PageID #: 415
`
`improvements in oncology patient care. By using UW’s patented Duplex Sequencing methods,
`
`the risk to the diagnostic patient is lower, cancer can be detected earlier (sometimes even before
`
`a tumor mass is identified), optimal treatments can be identified and prescribed quickly, and the
`
`costs to the patient and healthcare system are significantly reduced while personalized care
`
`improves health outcomes. Additionally, with much greater sensitivity, recurrent or residual
`
`cancer can be detected at previously undetectable levels to allow medical intervention at stages
`
`when they are most effective.
`
`23.
`
`For patients undergoing cancer treatment, the inventions can detect the emergence
`
`of drug-resistant cancer cells, allowing clinicians to select appropriate therapeutics.
`
`24.
`
`In addition to cancer applications, the patented technology offers the ability to aid
`
`in crime-scene forensics, to identify the emergence of drug-resistant microbes, and to sequence
`
`fetal DNA from maternal blood for non-invasive prenatal diagnostics, to name just a few
`
`applications.
`
`25.
`
`Realizing the enormous value of their breakthrough, the inventors, in
`
`collaboration with UW, sought patent protection for their inventions starting in early 2012. And
`
`in 2015, Dr. Salk and his colleagues co-founded TwinStrand with a Small Business Innovation
`
`Research grant and seed funding to develop and commercialize the patented duplex sequencing
`
`methods.
`
`26.
`
`Today, TwinStrand applies the patented duplex sequencing methods to
`
`applications in clinical medicine and life sciences, among others. TwinStrand’s customers
`
`include researchers, academic institutions, government and private laboratories, federal agencies,
`
`health systems, regulatory bodies, pharma and biotech companies, and others, whose work
`
`benefits from highly accurate sequencing techniques. TwinStrand provides services for nucleic
`
`- 7 -
`
`00007
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 8 of 36 PageID #: 416
`
`acid analysis using the patented Duplex Sequencing methods and provides customers with
`
`Duplex Sequencing kits. These kits include a DNA library prep kit containing the reagents,
`
`adapters, and other components necessary to practice its Duplex Sequencing process.
`
`TwinStrand also provides access to bioinformatics software to process raw sequence read files
`
`and produce error-corrected sequences according to the patented processes.
`
`C.
`
`Guardant’s Willful Infringement of the Asserted Patents
`1.
`
`Guardant’s infringing products and services
`
`27.
`
`Starting in 2014, Guardant began selling a number of products and services to
`
`monetize a Guardant-performed sequencing method that infringes the Asserted Patents. Guardant
`
`markets this sequencing method under the moniker “Digital Sequencing Technology.” Guardant
`
`IPO Prospectus, 96, 105, 108, and 120 (2018) (“Guardant IPO Prospectus”) (Exhibit F).
`
`28.
`
`In particular, Guardant sells kits for diagnostic purposes to customers around the
`
`world, including the Guardant360 lab developed test (“LDT”), Guardant360 CDx (“CDx”),
`
`GuardantOMNI (“Omni”), Guardant Reveal1 (“Reveal”), Guardant LUNAR-2 (“LUNAR-2”),
`
`Guardant360 Response, and Guardant360 TissueNext (collectively, “the Guardant Kits” or
`
`“Accused Products”). The Guardant Kits are used by Guardant’s customers to collect tissue
`
`samples and return them to Guardant. Guardant then performs its infringing sequencing method
`
`at its Redwood City, California laboratory. Each of these kits uses the same or essentially the
`
`same underlying sequencing technology, which Guardant calls its Digital Sequencing
`
`Technology. Solutions, GUARDANT HEALTH, https://guardanthealth.com/solutions/ (last visited
`
`Aug. 2, 2021) (Exhibit G).
`
`
`1 Guardant previously marketed “Reveal” as the “LUNAR-1” test.
`
`- 8 -
`
`00008
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 9 of 36 PageID #: 417
`
`29.
`
`In 2017, Dr. Rick Lanman, Guardant’s Chief Medical Officer at the time,
`
`described the methods that Guardant performs, stating: “We actually barcode. You have double
`
`stranded DNA—two strands. Each one is going to get a digital bar code attached to it. After we
`
`sequence it, if those two strands don’t match—the Watson allele and Crick allele, they should be
`
`complementary—then we correct it.” The Lung Cancer Living Room – Molecular Testing,
`
`Bonnie J. Addario Lung Cancer Foundation, YOUTUBE, 0:56:12–0:56:36 (Feb. 21, 2017).
`
`30.
`
`Guardant further monetizes its infringing sequencing method by selling certain
`
`data services, including Guardant Connect and Guardant Inform (collectively, “the Guardant
`
`Services” or “Accused Services”), to its customers. With Guardant Connect, Guardant sells
`
`access to a real-time database of patients receiving Guardant360 liquid biopsy assays so that
`
`customers can identify patients who may be eligible for clinical trials. And, with Guardant
`
`Inform, Guardant sells access to clinical information and genomic data collected from
`
`Guardant’s infringing Guardant360 liquid biopsy test.
`
`2.
`
`Guardant knew or should have known of the Asserted Patents and that
`its conduct amounted to infringement
`
`31.
`
`Guardant undoubtedly knew or should have known of the inventions claimed in
`
`each of the Asserted Patents before it launched the Accused Products. At or around the time
`
`Guardant launched the first of the Accused Products, Guardant made several unsuccessful
`
`attempts to license the patent family that includes the Asserted Patents. Having failed there,
`
`Guardant attempted and failed to cancel the exclusive license between UW and TwinStrand.
`
`And, Guardant repeatedly faced patentability rejections based on UW’s patent documents when
`
`Guardant was prosecuting its own patent applications directed to its infringing commercial
`
`sequencing method.
`
`- 9 -
`
`00009
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 10 of 36 PageID #: 418
`
`32.
`
`Indeed, on numerous occasions, Guardant cited the Asserted Patents during
`
`Guardant’s prosecution of its own later-filed patents attempting to cover its infringing
`
`technology.2 For example, in a non-final rejection of one of Guardant’s applications, the
`
`Examiner considered the ’631 patent to be one of three “references of interest.”3 In another
`
`instance, the Examiner rejected Guardant’s application as anticipated by a UW application from
`
`which the asserted ’631 and ’699 patents were continuations.4
`
`33.
`
`In inter partes review proceedings, a petitioner challenged the validity of
`
`Guardant’s genetic sequencing patents using as prior-art references a provisional application to
`
`which the Asserted Patents claim priority and a patent family member of the Asserted Patents.5
`
`34.
`
`In European opposition proceedings, UW’s PCT application 2013/032665—
`
`whose national stage application yielded the ’631, ’699, and ’951 patents—was cited to revoke
`
`Guardant’s European Patent Nos. 3087204, 3378952, and 2893040, which all related to genetic
`
`sequencing.
`
`
`2 U.S. Patent App. Nos. 16/593,633 (Non-Final Rejection and Notice of References Cited
`dated January 22, 2020), 15/669,779 (IDS dated April 27, 2020), 16/389,680 (IDS dated April
`16, 2020), 16/601,168 (IDS dated June 3, 2020 and IDS dated July 13, 2020), 16/897,038 (IDS
`dated June 25, 2020 and July 28, 2020), and 17/068,710 (IDS dated October 12, 2020 and
`November 23, 2020).
`3 U.S. Patent App. No. 16/593,633 (Non-Final Rejection and Notice of References Cited
`dated January 22, 2020).
`4 U.S. Patent App. No. 14/712,754 (Non-Final Office Action dated December 4, 2015).
`5 See, e.g., Found. Med., Inc., v. Guardant Health, Inc., PTAB-IPR2019-00130, Patent
`Owner’s Preliminary Response, at 12–14 (March 6, 2019) (citing U.S. Patent No. 9,752,188);
`Found. Med., Inc., v. Guardant Health, Inc., PTAB-IPR2019-00653, Petition for Inter Partes
`Review, at Exhibits 1011 and 1012 (May 20, 2019) (citing U.S. Patent No. 9,752,188 and U.S.
`Provisional App. 61/613,413).
`
`- 10 -
`
`00010
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 11 of 36 PageID #: 419
`
`35.
`
`Yet, despite facing TwinStrand’s licensed intellectual property time and again,
`
`Guardant continued its infringing activities in conscious disregard of UW and TwinStrand’s
`
`intellectual property rights. Guardant’s infringement has gone on long enough.
`
`COUNT I
`
`(Direct Infringement of U.S. Patent No. 10,287,631)
`Plaintiffs re-allege and incorporate by reference Paragraphs 1–35 above, as if
`
`36.
`
`fully set forth herein.
`
`37.
`
`The ’631 patent is directed to methods of generating high accuracy sequence
`
`reads of a population of double-stranded target nucleic acid molecules. Claim 1 of the patent
`
`recites:
`
`A method of generating high accuracy sequence reads of a population of double-
`stranded target nucleic acid molecules, comprising:
`
`ligating each of the double-stranded target nucleic acid molecules to at least one
`adapter molecule, to form a population of adapter-target nucleic acid
`complexes, wherein each of the adapter molecules comprises—
`
`(a) a degenerate or semi-degenerate single molecule identifier (SMI)
`sequence that alone or in combination with the target nucleic acid
`fragment ends uniquely labels each ligated double-stranded target
`nucleic acid molecule such that each ligated double-stranded target
`nucleic acid molecule is distinguishable from other ligated double-
`stranded target nucleic acid molecules in the population, and
`
`(b) a strand-distinguishing nucleotide sequence that, following the ligation
`step, provides a region of non-complementarity between a first
`strand of each adapter-target nucleic acid complex and a second
`strand of the same adapter-target nucleic acid complex;
`
`for each of the adapter-target nucleic acid complexes—
`
`amplifying each strand of the adapter-target nucleic acid complex to
`produce a plurality of first strand adapter-target nucleic acid
`complex amplicons and a plurality of second strand adapter-target
`nucleic acid complex amplicons;
`
`- 11 -
`
`00011
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 12 of 36 PageID #: 420
`
`sequencing the adapter-target nucleic acid complex amplicons to produce a
`plurality of first strand sequence reads and plurality of second strand
`sequence reads;
`
`grouping the first strand sequence reads and the second strand sequence
`reads into a family of first and second strand sequence reads based
`on the degenerate or semi-degenerate SMI sequence alone or in
`combination with the target nucleic acid fragment ends;
`
`separating the first and second strand sequence reads into a set of first strand
`sequence reads and a set of second strand sequence reads based on
`the region of non-complementarity between the first strand and the
`second strand of the adapter-target nucleic acid complex;
`
`confirming the presence of at least one first strand sequence read and at least
`one second strand sequence read;
`
`comparing the at least one first strand sequence read with the at least one
`second strand sequence read;
`
`identifying nucleotide positions where the compared first and second strand
`sequence reads are non-complementary;
`
`identifying nucleotide positions where the compared first and second strand
`sequence reads are complementary; and
`
`generating a high accuracy consensus sequence read for each of the double-
`stranded target nucleic acid molecules in the population that
`includes only the nucleotide positions where the compared first and
`second strand sequence reads are complementary.
`
`38.
`
`Guardant has infringed and continues to infringe at least claim 1 of the ’631
`
`patent, literally or under the doctrine of equivalents, by performing the methods of the ’631
`
`patent in the United States.
`
`39.
`
`Guardant practices the preamble of claim 1 of the ’631 patent, which provides
`
`“[a] method of generating high accuracy sequence reads of a population of double-stranded
`
`target nucleic acid molecules.” For example, Guardant touts that its sequencing method provides
`
`highly accurate reads of double-stranded DNA molecules. The Guardant360 Assay
`
`Specifications, at 1 (2018) (Exhibit H); see also Guardant IPO Prospectus at 121.
`
`- 12 -
`
`00012
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 13 of 36 PageID #: 421
`
`40.
`
`Guardant also practices the “ligating” step of claim 1 of the ’631 patent. This
`
`element recites “ligating each of the double-stranded target nucleic acid molecules to at least one
`
`adapter molecule, to form a population of adapter-target nucleic acid complexes.” Guardant
`
`practices this element by at least “repair[ing] ends of DNA with a 5’ phosphate prior to ligation
`
`of adapters.” Richard Lanman, et al., Analytical and Clinical Validation of a Digital Sequencing
`
`Panel for Qualitative, Highly Accurate Evaluation of Cell-Free Circulating Tumor DNA, 10
`
`PLOS ONE 10, at 18 (Oct. 16, 2015) (“Lanman”) (Exhibit I) (article authored by Guardant
`
`employees regarding Guardant’s digital sequencing technology). Guardant at least uses “blunt-
`
`end ligation” to attach library adapters to the ends of fragments of cell-free DNA. FDA Summary
`
`of Safety and Effectiveness Data for Premarket Approval No. P200010, at 6 (Feb. 10, 2020)
`
`(“FDA Summary No. P200010B”) (Exhibit J) (FDA summary of data for Guardant360 CDx
`
`product).
`
`41.
`
`Guardant also practices the “SMI sequence” element of claim 1 of the ’631 patent.
`
`This element recites “a degenerate or semi-degenerate single molecule identifier (SMI) sequence
`
`that alone or in combination with the target nucleic acid fragment ends uniquely labels each
`
`ligated double-stranded target nucleic acid molecule such that each ligated double-stranded
`
`target nucleic acid molecule is distinguishable from other ligated double-stranded target nucleic
`
`acid molecules in the population.” Guardant practices this element at least by attaching “library
`
`adapters containing inline barcodes” to the ends of fragments of cell-free DNA. FDA Summary
`
`No. P200010B at 6; see Oliver Zill, et al., The Landscape of Actionable Genomic Alterations in
`
`Cell-Free Circulating Tumor DNA from 21,807 Advanced Cancer Patients, 24(15) Clin. Cancer
`
`Res. 3528-38 & Supp. (Aug. 1, 2018) (“Zill”) (Exhibit K) (article authored by Guardant
`
`employees regarding Guardant’s digital sequencing technology). Double-stranded cfDNA is
`
`- 13 -
`
`00013
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 14 of 36 PageID #: 422
`
`“labeled with oligonucleotide barcodes.” Justin Odegaard, et al., Validation of a Plasma-Based
`
`Comprehensive Cancer Genotyping Assay Utilizing Orthogonal Tissue- and Plasma-Based
`
`Methodologies, 24(15) Clin. Cancer Res. 3539–49, at 3542, Fig. 1 (Apr. 24, 2018) (“Odegaard”)
`
`(Exhibit L) (article authored by Guardant employees regarding Guardant’s digital sequencing
`
`technology). After library preparation, enrichment, and sequencing, “[i]ndividual unique input
`
`molecules are then bioinformatically reconstructed using barcodes and sequence data to suppress
`
`analytic error modes.” Id. Guardant “build[s] double-stranded consensus representations of
`
`original unique cfDNA molecules using both inferred molecular barcodes and read start/stop
`
`positions.” Id. at 3540.
`
`42.
`
`Guardant also practices the “strand-distinguishing nucleotide sequence” element
`
`of claim 1 of the ’631 patent. This element recites “a strand-distinguishing nucleotide sequence
`
`that, following the ligation step, provides a region of non-complementarity between a first strand
`
`of each adapter-target nucleic acid complex and a second strand of the same adapter-target
`
`nucleic acid complex.” Guardant practices this element at least by requiring “each single-
`
`stranded half of the original double-stranded 5-30 ng input cfDNA sample” to be “separately
`
`encoded with oligonucleotide heptamers to create a self-referenced digital sequence duplex
`
`library with properties similar to differential signaling in digital communications.” See Lanman
`
`at 22, Fig. S2. “[N]on-unique oligonucleotide heptamer barcodes are ligated to each half of
`
`individual double-stranded cfDNA.” Id. at 18. Guardant further practices this element at least by
`
`“generat[ing] a duplex library whereby each single-stranded half of the original double-stranded
`
`input cfDNA sample is separately encoded with said oligonucleotides.” Id. Guardant also
`
`practices this element by having “[e]ach strand of a double-stranded cfDNA molecule . . .
`
`- 14 -
`
`00014
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 15 of 36 PageID #: 423
`
`individually tagged, allowing custom software to compare the two complementary strands . . . .”
`
`See id. at 19.
`
`43.
`
`Guardant also practices the “amplifying” element of claim 1 of the ’631 patent.
`
`This element recites “for each of the adapter-target nucleic acid complexes—amplifying each
`
`strand of the adapter-target nucleic acid complex to produce a plurality of first strand adapter-
`
`target nucleic acid complex amplicons and a plurality of second strand adapter-target nucleic
`
`acid complex amplicons.” Guardant practices this element at least by performing a step where
`
`“in-line adapters are ligated immediately after cfDNA isolation, prior to PCR and target capture
`
`steps.” Zill at 1; see FDA Summary No. P200010B at 6 (describing that the cfDNA fragments
`
`ligated to barcoded adapters are amplified by PCR before multiple samples are pooled for
`
`sequencing); see Lanman at 22, Fig. S2 (describing that after generating a duplex library, the
`
`digital sequence libraries are amplified).
`
`44.
`
`Guardant also practices the “sequencing” element of claim 1 of the ’631 patent.
`
`This element recites “sequencing the adapter-target nucleic acid complex amplicons to produce a
`
`plurality of first strand sequence reads and plurality of second strand sequence reads.” Guardant
`
`practices this element at least by “parallel sequencing of amplified target genes to an average
`
`depth of coverage greater than 2,700 unique molecules.” See FDA Summary No. P200010B at 6;
`
`see Lanman at 22, Fig. S2 (describing that the digital sequencing libraries are analyzed using
`
`paired-end sequencing).
`
`45.
`
`Guardant also practices the “grouping” element of claim 1 of the ’631 patent. This
`
`element recites “grouping the first strand sequence reads and the second strand sequence reads
`
`into a family of first and second strand sequence reads based on the degenerate or semi-
`
`degenerate SMI sequence alone or in combination with the target nucleic acid fragment ends.”
`
`- 15 -
`
`00015
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 16 of 36 PageID #: 424
`
`Guardant practices this element at least by having “[e]ach strand of a double-stranded cfDNA
`
`molecule . . . individually tagged, allowing custom software to compare the two complementary
`
`strands . . . .” Lanman at 19. “Processed reads were then aligned to hg19 . . . and used to build
`
`double-stranded consensus representations of original unique cfDNA molecules using both
`
`inferred molecular barcodes and read start/stop positions.” Odegaard at 3540. Moreover,
`
`Guardant practices this element at least by detecting fusions, whereby “overlapping paired-end
`
`reads [are] merged to form a representation of candidate fusion cfDNA molecules that are
`
`mapped to initial unique cfDNA molecules based on molecular barcoding and alignment
`
`information. Zill at 2. “Candidate fusion events are identified as clusters of molecules with
`
`similar directionality and breakpoint proximity . . . .” Id.
`
`46.
`
`Guardant also practices the “separating” element of claim 1 of the ’631 patent.
`
`This element recites “separating the first and second strand sequence reads into a set of first
`
`strand sequence reads and a set of second strand sequence reads based on the region of non-
`
`complementarity between the first strand and the second strand of the adapter-target nucleic acid
`
`complex.” Guardant practices this element at least by having “[e]ach strand of a double-stranded
`
`cfDNA molecule . . . individually tagged, allowing custom software to compare the two
`
`complementary strands . . . .” Lanman at 19.
`
`47.
`
`Guardant also practices the “confirming” element of claim 1 of the ’631 patent.
`
`This element recites “confirming the presence of at least one first strand sequence read and at
`
`least one second strand sequence read.” Guardant practices this element at least by using
`
`sequencing reads “to reconstruct each individual cfDNA molecule present in the original patient
`
`sample with high-fidelity using proprietary double-stranded consensus sequence representation.”
`
`Odegaard at 3542. Guardant also practices this element at least by “measur[ing] the total number
`
`- 16 -
`
`00016
`
`

`

`Case 1:21-cv-01126-UNA Document 1 Filed 08/03/21 Page 17 of 36 PageID #: 425
`
`of unique fragments covering each gene comprised of both halves of the original parent
`
`molecules.” Lanman at 19.
`
`48.
`
`Guardant also practices the “comparing” element of claim 1 of the ’631 patent.
`
`This element recites “comparing the at least one first strand sequence read with the at least one
`
`second strand sequence read.” Guardant practices this element at least by comparing the two
`
`complementary and individually tagged strands of a cfDNA molecule to ascertain any errors. See
`
`Lanman at 18–19.
`
`49.
`
`Guardant also practices the “non-complementary identifying” element of claim 1
`
`of the ’631 patent. This element recites “identifying nucleotide positions where the co