UNITED STATES PATENT AND TRADEMARK OFFICE
`___________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`___________________
`
`TWINSTRAND BIOSCIENCES, INC.
`Petitioner,
`v.
`GUARDANT HEALTH
`Patent Owner.
`
`___________________
`
`Inter Partes Review Case No. IPR2022-01115
`
`U.S. Patent No. 10,801,063
`___________________
`
`DECLARATION OF PAUL T. SPELLMAN, Ph.D.
`
`Mail Stop "PATENT BOARD"
`Patent Trial and Appeal Board
`U.S. Patent and Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`EX1002
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`
`
`TABLE OF CONTENTS
`
`- i -
`
`B.
`
`C.
`
`INTRODUCTION .......................................................................................... 1
`I.
`II. MY BACKGROUND AND QUALIFICATIONS ........................................ 2
`III.
`SUMMARY OF OPINIONS .......................................................................... 4
`IV. LIST OF DOCUMENTS CONSIDERED ................................................... 10
`V.
`PERSON OF ORDINARY SKILL IN THE ART ....................................... 16
`VI. STATE OF THE ART .................................................................................. 17
`A.
`The protocols for DNA sequencing were well known in the art ........ 17
`1.
`Genetic information comprises the building blocks of life ..... 17
`2.
`DNA sequencing was commonly performed to determine
`the order of nucleotides in an individual’s DNA ..................... 17
`Next-generation sequencing involved well-known steps ........ 19
`a)
`Template preparation and tagging ................................. 20
`b)
`Amplification ................................................................. 31
`c)
`Enrichment ..................................................................... 32
`d)
`Sequencing and detection .............................................. 33
`e)
`Sequence alignment and assembly ................................ 34
`Prior to December 28, 2013, cell-free DNA in blood had
`attracted interest for the diagnosis of cancer, fetal gender, and
`many other inherited disorders and was commonly sequenced
`using NGS platforms ........................................................................... 35
`1.
`Cell-free tumor DNA was a well-known cancer
`biomarker ................................................................................. 36
`Circulating cell-free fetal DNA was also a well-known
`biomarker for prenatal screening and diagnostics ................... 41
`Commercially available kits were routinely used to
`extract and isolate cfDNA from blood samples ....................... 43
`Prior to December 28, 2013, Duplex Sequencing was developed
`to dramatically improve the accuracy of next-generation
`sequencing ........................................................................................... 44
`
`3.
`
`2.
`
`3.
`
`

`

`1.
`
`D.
`
`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`Tagging of DNA with adapters comprising molecular
`barcodes ................................................................................... 46
`Amplification ........................................................................... 51
`2.
`Target enrichment .................................................................... 52
`3.
`Sequencing ............................................................................... 53
`4.
`5. Mapping the sequence reads to a reference sequence ............. 53
`6.
`Grouping the sequenced strands into paired families .............. 53
`7.
`Generating a single-strand consensus sequences ..................... 56
`8.
`Comparing the single strand consensus sequence to its
`complementary strand-mate (generating a Duplex
`Consensus) ............................................................................... 56
`Prior to December 28, 2013, the art taught that Duplex
`Sequencing could be used to sequence cfDNA................................... 58
`VII. THE ’063 PATENT SPECIFICATION AND CLAIMS ............................. 60
`VIII. PROSECUTION HISTORY ........................................................................ 69
`IX. THE MEANING OF CLAIM TERMS ........................................................ 74
`X.
`LEGAL BASIS FOR MY ANALYSIS ........................................................ 75
`XI. KEY PRIOR ART ........................................................................................ 76
`A. Narayan (EX1082) .............................................................................. 77
`B.
`Schmitt (EX1009), Schmitt-623 (EX1083) ......................................... 77
`C. Meyer (EX1005) .................................................................................. 83
`D. Kivioja (EX1006) ................................................................................ 84
`E.
`Craig (EX1007) ................................................................................... 84
`XII. GROUND 1: CLAIMS 1-7, 9-11, 15-18, AND 22-28 WOULD
`HAVE BEEN OBVIOUS OVER NARAYAN, SCHMITT, AND
`MEYER ........................................................................................................ 85
`A.
`Claim 1 ................................................................................................ 85
`1.
`A method for classifying consensus sequences generated
`from sequencing reads derived from double-stranded
`cell-free deoxyribonucleic acid (cfDNA) molecules from
`a sample of a human subject, the method comprising ............. 86
`
`- ii -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`(a) non-uniquely tagging a population of double-stranded
`cfDNA molecules from the sample with more than a 10x
`molar excess of adapters comprising molecular barcodes,
`relative to the double-stranded cfDNA molecules in the
`population, to generate non-uniquely tagged parent
`polynucleotides ........................................................................ 87
`wherein the double-stranded cfDNA molecules that map
`to a mappable base position of a reference sequence are
`tagged with a number of different molecular barcodes
`ranging from at least 2 to fewer than a number of double-
`stranded cfDNA molecules that map to the mappable
`base position, and ..................................................................... 91
`wherein at least 20% of the double-stranded cfDNA
`molecules are non-uniquely tagged with the adapters
`comprising the molecular barcodes at both ends of a
`molecule of the double-stranded cfDNA molecules ................ 93
`(b) amplifying a plurality of the non-uniquely tagged
`parent polynucleotides to produce progeny
`polynucleotides; ....................................................................... 94
`(c) enriching a plurality of the progeny polynucleotides
`for target regions of interest to generate enriched progeny
`polynucleotides; ....................................................................... 95
`(d) sequencing a plurality of the enriched progeny
`polynucleotides to produce a set of sequencing reads; ............ 96
`(e) mapping a plurality of sequencing reads from the set
`of sequencing reads to the reference sequence; ....................... 97
`(f) grouping a plurality of the mapped sequencing reads
`into families of mapped sequencing reads based at least
`on (i) sequence information from the molecular barcodes
`and (ii) a beginning base position and an ending base
`position of the mapped sequencing reads .............................. 100
`(g) generating a consensus sequence for each family from
`among one or more of the families to produce a set of
`consensus sequences; and ...................................................... 101
`(h) classifying one or more consensus sequences from
`among the set of consensus sequences as (1) paired
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`
`8.
`
`9.
`
`10.
`
`11.
`
`- iii -
`
`

`

`2.
`
`3.
`
`4.
`
`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`consensus sequences generated from sequencing reads
`representing a Watson strand and a Crick strand of a non-
`uniquely tagged parent polynucleotide or (2) unpaired
`consensus sequences generated from sequencing reads
`representing only one of either a Watson strand or a
`Crick strand of a non-uniquely tagged parent
`polynucleotide. ....................................................................... 102
`B. Motivation to combine ......................................................................103
`1.
`A POSA would have been motivated to use Schmitt’s 3-
`mer hybrid tagging approach to “non-uniquely” tag a
`population of cfDNA molecules. ........................................... 105
`A POSA would have been motivated to use “more than a
`10x molar excess.” ................................................................. 108
`A POSA would also have been motivated to generate “at
`least 20%” adapter-ligated cfDNA molecules. ...................... 111
`A POSA would have been motivated to use a number of
`different molecular barcodes ranging from at least 2 to
`fewer than a number of double-stranded cfDNA
`molecules that map to the mappable base position. ............... 113
`A POSA would have been motivated to amplify, enrich,
`sequence, map, group, generate a consensus sequence,
`and classify the consensus sequence as recited in claim 1. ... 113
`Reasonable expectation of success ....................................................118
`1.
`A POSA would have reasonably expected to successfully
`sequence cfDNA disclosed in Narayan with Schmitt’s
`Duplex Sequencing method because of knowledge in the
`art that a routine blood draw contains more than
`sufficient quantities of cfDNA. .............................................. 119
`A POSA would have reasonably expected to successfully
`“non-uniquely” tag cfDNA molecules. .................................. 121
`A POSA would have reasonably expected success in
`using “more than a 10x molar excess” of adapters. ............... 122
`A POSA would also have reasonably expected success in
`generating “at least 20%” adapter-ligated cfDNA
`molecules. .............................................................................. 122
`
`C.
`
`5.
`
`2.
`
`3.
`
`4.
`
`- iv -
`
`

`

`D.
`
`E.
`
`F.
`
`G.
`
`H.
`
`I.
`
`J.
`
`5.
`
`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`A POSA would have reasonably expected to successfully
`amplify, enrich, sequence, map, group, generate a
`consensus sequence, and classify the consensus sequence
`as recited in claim 1. .............................................................. 123
`Claim 2: The method of claim 1, wherein the population of
`double-stranded cfDNA molecules comprises 1 nanogram (ng)
`to 100 ng of double-stranded cfDNA molecules. .............................124
`Claim 3: The method of claim 1, wherein the sample is blood,
`plasma, or serum. ...............................................................................126
`Claim 4: The method of claim 1, wherein the adapters
`comprising the molecular barcodes are non-uniquely tagged to
`the double-stranded cfDNA molecules by blunt-end ligation or
`sticky-end ligation. ............................................................................128
`Claim 5: The method of claim 1, wherein at least 40% of the
`double-stranded cfDNA molecules are non-uniquely tagged
`with the adapters comprising the molecular barcodes at both
`ends of a molecule of the double-stranded cfDNA molecules..........130
`Claim 6: The method of claim 1, wherein more than a 90×
`molar excess of the adapters relative to the double-stranded
`cfDNA molecules in the population is used to generate the non-
`uniquely tagged parent polynucleotides. ...........................................132
`Claim 7: The method of claim 1, wherein the molecular
`barcodes are from a set of molecular barcodes having 2 to 1,000
`different molecular barcode sequences. ............................................134
`Claim 9: The method of claim 1, wherein the target regions of
`interest associated with cancer comprise genetic sequences of a
`plurality of genes selected from the group consisting of ALK,
`APC, BRAF, CDKN2A, EGFR, ERBB2, FBXW7, KRAS,
`MYC, NOTCH1, NRAS, PIK3CA, PTEN, RB1, TP53, MET,
`AR, ABL1, AKT1, ATM, CDH1, CSF1R, CTNNB1, ERBB4,
`EZH2, FGFR1, FGFR2, FGFR3, FLT3, GNA11, GNAQ,
`GNAS, HNF1A, HRAS, IDH1, IDH2, JAK2, JAK3, KDR,
`KIT, MLH1, MPL, NPM1, PDGFRA, PROC, PTPN11, RET,
`SMAD4, SMARCB1, SMO, SRC, STK11, VHL, TERT,
`CCND1, CDK4, CDKN2B, RAF1, BRCA1, CCND2, CDK6,
`NF1, TP53, ARID1A, BRCA2, CCNE1, ESR1, RIT1, GATA3,
`
`- v -
`
`

`

`L.
`
`2.
`
`3.
`
`K.
`
`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`MAP2K1, RHEB, ROS1, ARAF, MAP2K2, NFE2L2, RHOA,
`and NTRK1. ......................................................................................135
`Claim 10: The method of claim 1, further comprising
`amplifying a plurality of the enriched progeny polynucleotides
`prior to sequencing. ...........................................................................137
`Claim 11: The method of claim 1, further comprising (i)
`calculating a first quantitative measure of paired consensus
`sequences that map to a locus of the reference sequence, and (ii)
`calculating a second quantitative measure of unpaired
`consensus sequences that map to the locus of the reference
`sequence. ...........................................................................................139
`M. Claim 15 ............................................................................................142
`1.
`A method for classifying unique sequencing reads
`generated from sequencing reads derived from double-
`stranded cell-free deoxyribonucleic acid (cfDNA)
`molecules from a bodily fluid sample of a human subject,
`the method comprising: .......................................................... 142
`(a) tagging a population of double-stranded cfDNA
`molecules from the bodily fluid sample with more than a
`10x molar excess of adapters comprising molecular
`barcodes, relative to the double-stranded cfDNA
`molecules in the population, to generate tagged parent
`polynucleotides wherein at least 20% of the cfDNA
`molecules are ligated with the adapters comprising the
`molecular barcodes at both ends of a molecule of the
`double-stranded cfDNA molecules; ....................................... 143
`(b) amplifying a plurality of the tagged parent
`polynucleotides to produce progeny polynucleotides; .......... 144
`(c) sequencing a plurality of the progeny polynucleotides
`to produce a set of sequencing reads ..................................... 144
`(d) mapping a plurality of sequencing reads from the set
`of sequencing reads to a reference sequence; ........................ 145
`(e) determining unique sequencing reads from the set of
`mapped sequencing reads based at least on the molecular
`barcode sequences, wherein a unique sequencing read
`from among the unique sequencing reads is
`
`4.
`
`5.
`
`6.
`
`- vi -
`
`

`

`7.
`
`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`representative of a tagged parent polynucleotide from
`among the tagged parent polynucleotides; and ...................... 145
`(f) classifying one or more of the unique sequencing
`reads as either (1) paired sequences generated from
`sequencing reads representing a Watson strand and a
`Crick strand of a tagged parent polynucleotide or (2)
`unpaired sequences generated from sequencing reads
`representing only one of either a Watson strand or a
`Crick strand of a tagged parent polynucleotide. .................... 145
`Claim 16: The method of claim 15, further comprising
`selectively enriching a plurality of the progeny polynucleotides
`for target regions of [interest.] ...........................................................147
`Claim 17: The method of claim 16, wherein determining the
`unique sequencing reads comprises grouping a plurality of the
`mapped sequencing reads into families wherein a family of the
`families comprises mapped sequencing reads of progeny
`polynucleotides amplified from a same tagged parent
`polynucleotide from among the tagged parent polynucleotides. ......148
`Claim 18: The method of claim 15, further comprising (i)
`calculating a first quantitative measure of paired sequences that
`map to a locus of the reference sequence, and (ii) calculating a
`second quantitative measure of unpaired sequences that map to
`the locus of the reference sequence. ..................................................151
`Claim 22: The method of claim 15, wherein the molecular
`barcodes are from a set of molecular barcodes having 2 to
`10,000 different molecular barcode sequences. ................................153
`Claim 23: The method of claim 15, wherein the molecular
`barcodes are pre-determined sequences. ...........................................154
`Claim 24: The method of claim 15, wherein at least 40% of the
`cfDNA molecules are tagged with the adapters comprising the
`molecular barcodes at both ends of a molecule of the double-
`stranded cfDNA molecules. ..............................................................155
`Claim 25: The method of claim 15, wherein more than a 90×
`molar excess of the adapters relative to the double-stranded
`cfDNA molecules in the population is used to generate the
`tagged parent polynucleotides. ..........................................................157
`
`N.
`
`O.
`
`P.
`
`Q.
`
`R.
`
`S.
`
`T.
`
`- vii -
`
`

`

`V.
`
`B.
`
`U.
`
`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`Claim 26: The method of claim 15, further comprising
`quantifying a number of the unique sequencing reads
`determined from the set of mapped sequencing reads. .....................159
`Claim 27: The method of claim 15, wherein determining the
`unique sequencing reads is further based on (1) a start base
`position of a given sequencing read from among the set of
`mapped sequencing reads at which the given sequencing read
`starts aligning to the reference sequence, and (2) a stop base
`position of the given sequencing read at which the given
`sequencing read stops aligning to the reference sequence. ...............160
`W. Claim 28: The method of claim 1, further comprising
`determining a quantity of the set of consensus sequences. ...............162
`XIII. GROUND 2: CLAIM 8 WOULD HAVE BEEN OBVIOUS IN VIEW
`OF NARAYAN, SCHMITT, MEYER, AND CRAIG .............................. 164
`A. A POSA would have been motivated to incorporate Craig’s pre-
`determined barcode sequences into Schmitt’s Duplex
`Sequencing method. ..........................................................................166
`A POSA would have reasonably expected to successfully use
`Craig’s 6-mer barcodes in Schmitt’s Duplex Sequencing
`method. ..............................................................................................168
`XIV. GROUND 3: CLAIMS 12-14 AND 19-21 WOULD HAVE BEEN
`OBVIOUS OVER NARAYAN, SCHMITT, MEYER, AND
`KIVIOJA ..................................................................................................... 169
`1.
`Claims 12-13 and 19-20 ......................................................... 169
`2.
`Claims 14 and 21.................................................................... 176
`XV. OBJECTIVE EVIDENCE OF NONOBVIOUSNESS DO NOT
`SUPPORT PATENTABILITY .................................................................. 178
`XVI. CONCLUSION ........................................................................................... 179
`
`
`
`- viii -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`I, Paul T. Spellman, hereby declare as follows.
`
`I.
`
`INTRODUCTION
`1.
`I have been retained as an expert witness on behalf of TwinStrand
`
`Biosciences, Inc. for the above-captioned inter partes review (IPR). I am being
`
`compensated for my time in connection with this IPR at my standard consulting
`
`rate, which is $400 per hour.
`
`2.
`
`I understand that this Declaration accompanies a petition for IPR
`
`involving U.S. Patent No. 10,801,063 ("the '063 patent") (EX1001), which resulted
`
`from U.S. Patent Application No. 16/601,168 ("the '168 application"), filed on
`
`October 14, 2019. I understand that the '063 patent alleges a priority date of
`
`December 28, 2013. I refer to this date throughout this declaration.
`
`3.
`
`In preparing this Declaration, I have reviewed the '063 patent and each
`
`of the documents cited herein, in light of general knowledge in the art before
`
`December 28, 2013. In formulating my opinions, I have relied upon my
`
`experience, education, and knowledge in the relevant art. In formulating my
`
`opinions, I have also considered the viewpoint of a person of ordinary skill in the
`
`art ("POSA") (i.e., a person of ordinary skill in the field of molecular biology, as
`
`defined further below in §V) prior to December 28, 2013.
`
`4.
`
`I further understand that, according to the United States Patent and
`
`Trademark Office (“USPTO”) assignment records, the ‘063 is currently assigned to
`
`- 1 -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`
`Guardant Health, Inc.
`
`II. MY BACKGROUND AND QUALIFICATIONS
`5.
`I am a Full Professor with Tenure in the Department of Molecular and
`
`Medical Genetics at Oregon Health & Science University School of Medicine
`
`(“OHSU”). I have held this position since 2013. I am also co-leader of the
`
`Quantitative Oncology Program in the OHSU Knight Cancer Institute, a position I
`
`have held since 2017. I am also Co-Director of the Cancer Early Detection
`
`Advanced Research (“CEDAR”) Center. From 2018-2020, I was Interim Director
`
`for the Program of Computational Biology at OHSU School of Medicine. From
`
`2011-2013, I served as an Associate Professor at OHSU School of Medicine. From
`
`2003-2011, I was a Staff Scientist at the Life Science Division of Lawrence
`
`Berkeley Lab.
`
`6.
`
`I received a Bachelor’s degree in Biology in 1995 from the
`
`Massachusetts Institute of Technology. In 2000, I received my Ph.D. in Genetics
`
`from Stanford University Medical School. My thesis topic was “Generating and
`
`Analyzing Genome Scale Data.” From 2000-2003, I was a Post-Doctoral Fellow in
`
`the Department of Molecular and Cellular Biology at University of California,
`
`Berkeley.
`
`7. My current research focuses on computational biology, cancer
`
`biology, and cancer genomics. My lab develops technologies and approaches to
`
`- 2 -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`understand the processes by which cancer develops, monitor disease, and to
`
`identify therapeutic strategies. Within this broad area of research, I have specific
`
`experience in using population genetics to help determine who is at risk for cancer,
`
`how to computationally analyze genomic data to identify early changes in cancers,
`
`and how to accurately screen different populations for the disease. My team
`
`develops new methodologies for identifying changes in the cancer genome,
`
`systematic integration of multiple genomic data types, including copy number,
`
`expression and mutation, to better understand the process by which cancer
`
`develops.
`
`8.
`
`CEDAR is a $314 million effort funded by a gift from Phil and Penny
`
`Knight to detect and treat early cancers. Among other activities, my work as Co-
`
`Director of CEDAR includes identifying populations at risk for developing cancer
`
`and understanding the early biology of breast cancers.
`
`9.
`
`I have over 20 years of experience in the field of genomics and 15
`
`years of experience in cancer genomics. I have taught undergraduate and graduate
`
`level courses at OHSU in the area of sequencing technologies, computational
`
`analysis, and genetics. I have also supervised three completed Ph.D. theses since I
`
`joined the OHSU faculty in 2011.
`
`10.
`
`I have authored or co-authored more than 115 peer-reviewed
`
`publications that discuss methods of DNA manipulation and analysis, including
`
`- 3 -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`DNA sample preparation for sequencing, amplification (e.g. PCR), methods of
`
`DNA sequencing (including NGS and related sequencing methods), and
`
`bioinformatics methods for raw data analysis. In addition, I have presented over
`
`100 invited lectures or conference presentations regarding these topics. My
`
`curriculum vitae includes a sample list of these publications and presentations.
`
`11.
`
`I have received over $16 million in federal grants for my research in
`
`the area of cancer genomic and cancer precision medicine. I am currently one of 30
`
`principal investigators for the Genome Data Analysis Network, sponsored by the
`
`National Cancer Institute. The Genome Data Analysis Network is a consortium of
`
`computational researchers from across the United States focused on developing the
`
`framework for relating the genomics and outcomes of patients in cancer clinical
`
`trials.
`
`12. As a postdoctoral scholar, I was selected for the prestigious National
`
`Science Foundation Biocomputing Fellowship, which covered both my stipend and
`
`significant research costs. In 2017, I was named the inaugural holder of the Penny
`
`and Phil Knight Endowed Professorship for Cancer Research Innovation.
`
`III. SUMMARY OF OPINIONS
`13. Generally speaking, the ’063 patent is directed to methods for
`
`classifying sequences generated from sequencing reads of cfDNA molecules using
`
`commonly known, routinely performed sequencing methods that were taught in the
`
`- 4 -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`art. In particular, the claims recite methods for classifying sequences comprising
`
`the following general steps: (1) polynucleotide isolation; (2) tagging with adapters;
`
`(3) amplifying; (4) sequencing; (5) mapping; (6) grouping; (7) generating a
`
`consensus sequence; and (8) analyzing. There is nothing inventive in the steps of
`
`the ’063 patent claims.
`
`14. The claimed methods would have been obvious to a POSA in view of
`
`disclosures in the art cited in the Grounds below.
`
`15. Narayan teaches sequencing of circulating cell-free tumor DNA in
`
`plasma samples to track treatment-associated changes in circulating tumor DNA
`
`levels in patients with non-small cell lung cancer. EX1082, Abstract, 3492. To do
`
`so, Narayan discloses performing ultradeep sequencing on plasma samples of 30
`
`patients with stage 1-IV non-small cell lung cancer. EX1082, 3493.
`
`16. Schmitt teaches all the steps of Guardant’s claimed methods,
`
`including tagging DNA fragments (EX1009, ¶¶[0022], [0026], [0032]-[0034];
`
`Figs. 1-2; EX10831, ¶¶[0015], [0016], [0024]-[0026], Figs. 1-2), amplifying the
`
`
`1 As discussed below, Schmitt (EX1009) claims priority to provisional
`
`application no. 61/625,623 (“Schmitt-623,” EX1083). Since Schmitt-623 provides
`
`support for at least one claim in Schmitt, Schmitt is prior art to the ’063 patent at
`
`least as of Schmitt-623’s filing date of April 17, 2012. As such, I have provided
`
`- 5 -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`tagged fragments (EX1009, ¶¶[0006], [0010]; EX1083, ¶¶[0005], [0009]),
`
`enriching (EX1009, ¶[00102]; EX1083, ¶¶[0065]), sequencing the DNA (EX1009,
`
`¶¶[0074], [0095]; EX1083, ¶¶[0042], [0059]), mapping (EX1009, ¶¶[0096],
`
`[00103]; EX1083, ¶¶[0060], [0066]), generating consensus sequences (EX1009,
`
`¶[0099]; EX1083, ¶¶[0063]), and classifying consensus sequences from the
`
`sequencing reads (EX1009, ¶¶[0096], [00105]; EX1083, ¶¶[0060], [0068]) in a
`
`method called Duplex Sequencing.
`
`17. Meyer discloses parallel tagged sequencing as a molecular barcoding
`
`method designed to adapt the 454 parallel sequencing technology for use with
`
`multiple samples. EX1005, Abstract. Meyer discloses an overview of the tagging
`
`protocol and teaches that after each DNA sample is blunt-end repaired and sample-
`
`specific barcoding adapters are ligated to both ends of the molecules, the barcoded
`
`samples are pooled in equimolar ratios and unligated molecule ends are excluded
`
`from sequencing. EX1005, Fig. 1, 272-274. Meyer discloses that “the expected
`
`overall recovery” after quantifying tagged samples “is between 40 and 60%.”
`
`EX1005, 274.
`
`18. Kivioja teaches methods of quantitating unseen DNA molecules using
`
`
`citations to both Schmitt and Schmitt-623 in support of my opinions in this
`
`declaration.
`
`- 6 -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`count statistics by adding random DNA sequence labels (unique molecular
`
`identifiers). EX1006, 1. Specifically, Kivioja teaches how to accurately estimate
`
`the number of molecules generated without actually observing all of them.
`
`EX1006, 74, 16-18.
`
`19. Craig discloses “[a] total of 48 different 6-mer index sequences” that
`
`were “appended to adapter sequence[s]” and used for sequencing on the Illumina
`
`platform. EX1007, Suppl. Table 4, Suppl. Methods. Craig teaches that the
`
`disclosed 6-mer barcode design can “control, tolerate, and measure error during
`
`base-calling.” EX1007, 888. Craig also discloses that the 48 6-mer barcodes that
`
`were disclosed that indexes were designed “so that one, and in some cases two,
`
`sequencing errors could be tolerated without [a barcode] being incorrectly
`
`identified as being a different valid [barcode].” EX1007, 888.
`
`20. As discussed more fully below, claims 1-7, 9-11, 15-18, 22-28 of the
`
`’063 patent would have been obvious to a POSA over the combination of Narayan,
`
`Schmitt, and Meyer. A POSA would have had a reason to combine the teachings of
`
`Narayan, Schmitt, and Meyer to arrive at the claimed methods. This is because (1)
`
`a POSA would have been motivated to apply Schmitt’s Duplex Sequencing error
`
`correction and improved accuracy to Narayan’s cfDNA; (2) a POSA would have
`
`been motivated to non-uniquely tag cfDNA molecules using Schmitt’s hybrid
`
`tagging method; and (3) a POSA would have been motivated to optimize the
`
`- 7 -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`adapter-DNA ligation efficiency to maximize the percentage of adapter-ligated
`
`DNA fragments after ligation.
`
`21. A POSA also would have reasonably expected to use Schmitt’s
`
`Duplex Sequencing method on cfDNA, as disclosed in Narayan, because the prior
`
`art taught that Duplex Sequencing could be used to enhance accuracy when
`
`sequencing cfDNA from cancer patients. EX1008, 7. Moreover, a POSA would
`
`have reasonably expected to successfully non-uniquely tag cfDNA molecules to
`
`achieve a high adapter-DNA ligation efficiency because Schmitt and Meyer
`
`disclose that improving ligation efficiency increases the sensitivity of sequence
`
`detection. EX1009, ¶[0006]; EX1083, ¶[0005]; EX1005, 274. Furthermore, the art
`
`taught that sufficient quantities of cfDNA could be extracted from a routine blood
`
`draw and sequenced using Schmitt’s Duplex Sequencing methods. Thus, a POSA
`
`would have reasonably expected to be able to perform the claimed sequencing
`
`methods based on the disclosures in Narayan, Schmitt, and Meyer.
`
`22.
`
`In addition, claim 8 of the ’063 patent would have been obvious to a
`
`POSA over the combination of Narayan, Schmitt, Meyer, and Craig. A POSA
`
`would have been motivated to combine Narayan, Schmitt, Meyer as discussed
`
`above. In addition, a POSA would have been motivated to use Craig’s 6-mer
`
`barcodes in combination with Schmitt because Schmitt teaches performing Duplex
`
`Sequencing at greater depths for greater sensitivity and Craig discloses barcodes
`
`- 8 -
`
`

`

`Inter Partes Review of US Patent 10,801,063
`Declaration of Paul T. Spellman, Ph.D.
`that are able to mitigate errors in the barcode sequence itself, which is especially
`
`important when sequencing at greater depths. Moreover, a POSA would have
`
`reasonably expected to successfully use Craig’s barcodes in Schmitt’s Duplex
`
`Sequencing method because Schmitt discloses that a 6-mer barcode can be used in
`
`the hybrid tag embodiment, and that the molecular barcodes can be pre-determined
`
`sequences, just like Craig’s.
`
`23. As discussed below, claims 12-14, 19-21 of the ’063 patent would
`
`have been obvious to a POSA over the combination of Narayan, Schmitt, Meyer,
`
`and Kivioja. A POSA would have had a reason to combine the teachings of
`
`Narayan, Schmitt, and Meyer, as alre