UNITED STATES PATENT AND TRADEMARK OFFICE
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`___________________
`
`
`
`
`FORESIGHT DIAGNOSTICS INC.,
`Petitioner,
`
`v.
`
`PERSONALIS, INC.,
`Patent Owner.
`
`
`
`____________________
`
`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,033)
`____________________
`
`
`
`DECLARATION OF JOHN QUACKENBUSH, PH.D. IN SUPPORT OF
`PETITIONER’S REPLY TO PATENT OWNER RESPONSE
`
`
`
`
`
`
`
`11286694
`
`
`
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`TABLE OF CONTENTS
`
`2.
`
`3.
`
`I.
`II.
`
`B.
`
`Page
`Introduction ................................................................................................... 3
`Response to Dr. Furneaux’s Opinions Regarding Claim
`Construction .................................................................................................. 3
`A.
`“whole genome sequencing” .............................................................. 5
`1.
`Application of the preliminary construction. ........................... 7
`2.
`Express construction of the term “whole genome
`sequencing” is not required. ................................................... 25
`“capture probes” ............................................................................... 27
`1.
`The preliminary construction comports with the
`intrinsic evidence. .................................................................. 28
`Personalis’s proposed construction contradicts its
`patents and the extrinsic evidence.......................................... 31
`The literature and Mr. West confirm that PCR primers
`perform the function of capturing DNA loci for
`sequencing. ............................................................................. 36
`III. Response to Dr. Furneaux’s Opinions Regarding the ’394 Patent ............. 38
`A. Ground 1 ........................................................................................... 38
`1.
`Step 1(b): “whole genome sequencing” ................................. 38
`2.
`Step 1(c): “capture probes” .................................................... 50
`3.
`Step 1(e): “using said plurality of capture probes to
`generate an additional subset . . . .” ....................................... 52
`Ground 2 ........................................................................................... 54
`B.
`Ground 3 ........................................................................................... 55
`C.
`IV. Response to Dr. Furneaux’s Opinions Regarding the ’033 Patent ............. 58
`
`
`
`
`- i -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`V.
`
`Page
`A. Ground 1 ........................................................................................... 58
`1.
`Step 1(c): “generated from said nucleic acid sample” ........... 58
`Grounds 2 and 3 ............................................................................... 61
`B.
`Response to Dr. Furneaux’s Arguments Regarding Secondary
`Considerations ............................................................................................ 64
`1.
`Dr. Furneaux’s arguments regarding the “sensitivity”
`of Next Personal are irrelevant and inaccurate. ..................... 64
`Dr. Furneaux’s arguments about skepticism are
`confusing and distort the record. ............................................ 66
`VI. Response to Mr. West’s Conception Testimony ........................................ 66
`
`
`2.
`
`
`
`
`- ii -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,033)
`
`I.
`
`Introduction
`1. My name is John Quackenbush. I previously submitted declarations in
`
`support of Foresight Diagnostics Inc.’s Petitions for inter partes review of U.S.
`
`Patent Nos. 11,384,394 and 11,408,033.
`
`2.
`
`I submit this declaration to respond to certain arguments raised in the
`
`declaration of Patent Owner Personalis’s expert Dr. Furneaux. Ex2031. I also
`
`respond to some aspects of the declaration submitted by Mr. West regarding the
`
`alleged conception of the claimed inventions. Ex2032.
`
`II. Response to Dr. Furneaux’s Opinions Regarding Claim Construction
`3.
`Patent Owner and Dr. Furneaux’s response to the Petitions raise issues
`
`regarding the appropriate claim construction of the terms “whole genome
`
`sequencing” (’394 patent) and “capture probes” (’394 and ’033 patents). I briefly
`
`summarize the disputes here.
`
`a. For “whole genome sequencing” the parties dispute whether Leary’s
`
`mate-pair whole genome sequencing technique satisfies the claim. In
`
`short, the contemporaneous literature described Leary (or analogous
`
`techniques) as “whole genome sequencing,” thus in my opinion a
`
`POSA would have understood that Leary discloses WGS regardless of
`
`the precise countours of a formal claim construction for this term.
`
`Nevertheless, I understand that the Board has preliminarily construed
`
`
`
`
`- 3 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`“whole genome sequencing” to mean “a nontargeted approach to
`
`sequencing all or almost all of the genome.” Under this framework, the
`
`question ultimately becomes whether Leary’s technique – which is a
`
`nontargeted approach – sequences “almost all of the genome.” (There
`
`is no dispute that sequencing “all” of the genome was not possible at
`
`the relevant time (and perhaps even today).) A POSA’s understanding
`
`of “almost all” would have taken into account the techniques that the
`
`contemporaneous
`
`literature –
`
`including Personalis’s patents –
`
`considered to be “whole genome sequencing,” which included low
`
`coverage techniques that performed nontargeted sequencing at 50% or
`
`less (including 10%) sequence coverage. Accordingly, Leary’s
`
`technique, which Dr. Furneaux asserts achieves between 53-90%
`
`sequence coverage, sequences what a POSA would have understood to
`
`be “almost all” of the genome.
`
`b. For “capture probes,” the parties dispute whether PCR primers fall
`
`within the scope of that term as it would have been understood in light
`
`of the specification. In light of the specification, the Board, in my
`
`opinion, was correct to construe the term to encompass, at a minimum
`
`“PCR primer capture probes and solution-based hybridization capture
`
`probes (e.g., biotinylated oligonucleotide baits).” Personalis disputes
`
`
`
`
`- 4 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`this construction, nominally contending that capture probes include
`
`array-based capture probes on a solid support or solution-phase
`
`hybridization capture probes. In application, Personalis seeks to limit
`
`the term to “pulldown probes,” which a POSA would have understood
`
`to encompass biotinylated oligonucleotide baits. Such baits are
`
`comprised of a nucleic acid linked to a biotin moiety. Because the
`
`specification is clear that biotin is a separate, optional component of a
`
`capture probe, e.g., a “label,” in my view a POSA reading the
`
`Specification would not adopt Personalis’s construction. Moreover, as
`
`detailed below, the testimony of Mr. West and the contemporaneous
`
`literature confirm that PCR primers satisfy the claim.
`
`A.
`4.
`
`“whole genome sequencing”
`I understand that it in its Institution Decision for the ’394 patent, the
`
`Board preliminarily construed the term “whole genome sequencing” as used in claim
`
`element 1(b) of the ’394 patent to mean, “a nontargeted approach to sequencing all
`
`or almost all of the genome.” As explained in Section III, below, the grounds of
`
`unpatentability set forth in my initial declaration satisfy this construction.
`
`5.
`
`As discussed further below, Dr. Furneaux’s application of the
`
`preliminary construction adds potentially indeterminate mathematical endpoints to
`
`the term “whole genome sequencing” that are inconsistent with how the term was
`
`
`
`
`- 5 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`understood in the literature (and hence by a POSA) at the time of the priority date
`
`(regardless of whether May 8, 2012 or January 13, 2013 is used). The literature of
`
`the time demonstrates that a POSA would have understood that the term “whole
`
`genome sequencing” encompassed Leary’s PARE technique.
`
`6.
`
`The ’394 patent introduces the term “whole genome sequencing,” in the
`
`“Background of the Invention,” purporting to define the invention (consistent with
`
`the intent of Personalis’s “Accuracy Program”) as techniques that improve known
`
`drawbacks associated with prior art “whole genome sequencing” methods. Ex1001,
`
`1:23-25 (“Current methods for whole genome [] sequencing may be costly and fail
`
`to capture many biomedically important variants.”). The patent does not expressly
`
`define the term “whole genome sequencing,” let alone in a manner that clearly alters
`
`its meaning to refer to the allegedly “improved” techniques.
`
`7.
`
`Accordingly, a POSA would understand that when the patent uses the
`
`term “whole genome sequencing,” it is referring to the methods that existed at the
`
`time, which included the alleged shortcomings identified in the patent. I note this
`
`explicitly because aspects of Dr. Furneaux’s testimony suggest that he applied an
`
`interpretation of “whole genome sequencing” that imports the patent’s alleged
`
`improvements to sequencing that would not have been reflected in the meaning of
`
`the term at the priority date.
`
`
`
`
`- 6 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`8.
`
`For example, Dr. Furneaux contends that “to cover the all the base pairs
`
`in the whole human genome using the mate pair approach in Leary, one would need
`
`to very accurately sequence about 64 million non-overlapping mate-pairs (i.e., 3.2
`
`billion divided by 50).” Ex2031 ¶96. But there is no requirement in the claims that
`
`the whole-genome sequencing achieve an undefined standard of “very accurate,”
`
`and at most, the patent requires that “[t]he accuracy of the sequencing reaction may
`
`be greater than about 70%.” Ex1001, 7:18-14 (“The accuracy of the sequencing
`
`reaction may be greater than about 70%, 75%, 80%, 90%, 91% . . . .”). Dr. Furneaux
`
`further contends that “one would still need to contend with biases such as . . . G/C
`
`rich sequences.” Ex2031 ¶96. But at his deposition, Dr. Furneaux confirmed that the
`
`claims do not require implementing the ’394 patent’s purported solution to G/C rich
`
`sequences. Ex1223, 103:15-104:20 (Q. And that’s the one that’s not reflected in the
`
`claims of the ’394 patent; correct? A. Absolutely not. And yeah, why [] it isn’t, you
`
`know, I couldn’t say.).
`
`1.
`Application of the preliminary construction.
`The Board’s construction includes the requirements that the “whole
`
`9.
`
`genome sequencing” assay sequence “all” or “almost all of the genome.” This
`
`construction is satisfied by the cited prior art, because a POSA would have
`
`understood that “almost all” of the genome encompassed the shallow paired-end and
`
`mate pair sequencing techniques that were commonly used in the art. See Section
`
`
`
`
`- 7 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`III. I clarify in this section some of the ambiguities injected by Dr. Furneuax’s
`
`application of the Board’s construction.
`
`a)
`“All of the genome”
`10. As Mr. West confirmed, sequencing of “all of the genome,” to the
`
`extent possible at all, became possible only recently, well after the filing of the
`
`patent:
`
`Q. When did it first become possible to sequence a human genome
`consistent with approximately 3.2 billion base pairs without any gaps?
`
`A. In the last few years, quite a long time after we filed these patents
`or filed for the – filed the provisional filing. So – but I don’t know the
`exact date. You could look it up.
`
`Ex1222, 46:15-21; id. 37:17-24 (“It was well known at that time and is still mostly
`
`true that the – there are portions of the human genome which are difficult for current
`
`sequencing technologies and sequencing technologies of that time to capture . . . .
`
`And – and so those are not included in the reference sequence.”). The ’394 patent’s
`
`usage of the term “whole genome sequencing” is consistent with this understanding,
`
`as it repeatedly recognizes that “whole genome sequencing” data has gaps, misses
`
`certain regions, and includes zero and low mean coverage regions.1
`
`
`1 In fact, Mr. West was correct regarding the relative dates of the patent filings
`and the completion of the genome. It was not until March 2022 that a “complete”
`genome
`sequence
`was
`announced
`(Nurk
`2022;
`
`
`
`
`- 8 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`11. Even though no method was capable of sequencing “all of the genome,”
`
`this endpoint of the Board’s construction is still important because it must be known
`
`to apply the “almost all” component of the Board’s construction. For example, Dr
`
`Furneaux testified to the following at deposition:
`
`Q. Your testimony was since your declaration, you’ve pondered the
`question further, including in light of Mr. West’s experience and
`[Illumina’s] experience, and determined that the threshold for almost
`all of the genome was 90 percent; correct?
`
`A. Coverage.
`
`Q. 90-percent coverage?
`
`A. Uh-huh.
`
`Ex1223, 97:21-98:8.
`
`12.
`
`In light of this testimony, Dr. Furneaux was shown the formula below.
`
`Ex1223, 106:12-17; Ex1210. To determine whether a given experiment sequenced
`
`“almost all of the genome,” one must know the appropriate denominator to use.
`
`
`https://www.genome.gov/news/news-release/researchers-generate-the-first-
`complete-gapless-sequence-of-a-human-genome) and even then the Y chromosome
`was not completed until August 2023 (https://www.nih.gov/news-events/news-
`releases/researchers-assemble-first-complete-sequence-human-y-chromosome).
`
`
`
`
`- 9 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`
`13. As Mr. West confirmed at deposition, the value for “all of the genome”
`
`that a POSA would have applied was the size of the known reference genome to
`
`which raw sequence reads are mapped in a sequencing assay. “Generally, the goal
`
`was to sequence as much as possible of the known reference sequence because the
`
`– you could only form a consensus sequence by aligning sequence reads to the
`
`reference.” Ex1222, 38:3-6; see also id. 38:7-8 (“Any portion of the human genome,
`
`which was not in the official reference, couldn’t be included.”).
`
`14. The ungapped length of the reference sequence at the time the ’394
`
`patent was filed, hg18, rounded to 2.9 Gb. Ex1209; Ex1222 45:6-9 (“[W]hen we say
`
`that the length of the human reference sequence at that time was about 2.85 billion
`
`bases, I believe that did not include any of – gaps.”). These datapoints are consistent
`
`
`
`
`- 10 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`with the ’394 patent itself, which uses a genome size of “2,861 Mbp,” (2.86 Gbp) as
`
`the size of the “DegappedGenome” in the few examples of laboratory preparation
`
`steps for sequencing experiments noted in its Specification. See ’394 patent, Table
`
`13.
`
`15. Accordingly, a POSA would have understood “all of the genome,” to
`
`mean the ~2.86 Gbp recognized in the degapped reference genome known at the
`
`priority date.
`
`b)
`
`“Almost all of the genome”
`(1) Dr. Furneaux’s 90% Opinion
`I note that Dr. Furneaux added further confusion to his 90% opinion,
`
`16.
`
`testifying in response to the question “Did the 90-percent number change over
`
`time?” that “I think the Board is extremely insightful here – is that they are well
`
`aware that all of this is a moving target.” Ex1223, 94:21-25. Thus, it appears that Dr.
`
`Furneaux considers that both the threshold percentage for “almost all,” as well as the
`
`denominator “Y” to use in the equation referenced above, are “a moving target.” I
`
`understand that this is legally incorrect, as the meaning of a claim term is the
`
`meaning it would have had to a POSA at the time of the alleged invention.
`
`17. Setting aside Dr. Furneaux’s testimony that his 90% opinion is a
`
`“moving target” in multiple respects, it is difficult to apply his understanding of the
`
`
`
`
`- 11 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`term because he would not even identify what he believed to be the appropriate
`
`denominator.
`
`18.
`
`In his declaration testimony, Dr. Furneaux readily divided 2.9 billion
`
`base pairs sequenced in Leary by the known length of the human genome (3.2 billion
`
`bases) to arrive at 91% sequence coverage. Ex2031 ¶97. In that analysis, he asserted
`
`that 2.9 billion bases was “far less than the 3.2 billion bp acknowledged to be present
`
`in the reference genome at this priority date.” Id. I note first that Dr. Furneaux
`
`incorrectly called 3.2 billion base pairs the length of the known reference genome.
`
`No reference genome today, let alone in 2012-2013, lacked gaps. See Section
`
`II.A.1.a), above. As noted above and confirmed by the patent itself, the reference
`
`genome at the relevant time period was about 2.86 billion bases. See id.
`
`19. However, Dr. Furneaux could not explain how to use his new 90%
`
`cutoff in practice. For example he suggested that the number of bases required to be
`
`sequenced in order to satisfy “almost all of the genome” could vary depending on
`
`the experiment:
`
`Q. And I just wanted to see again, are you able to tell me what numbers
`you would plug in for X and Y in this equation to get to the genome?
`
`A. So let’s make sure that we’re talking about the same thing. We’re
`talking about an inspection assay that determines whether are residue is
`covered by a read.
`
`
`
`
`- 12 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`Are we agreed on that?
`
`Q. I believe so.
`
`A. Okay. And so here you’re referring to Y being the size of genome
`in base pairs.
`
`Q. And what is Y?
`
`A. What is Y? Well I’m sure there’s going to be discussions about what
`Y may or may not be, and what’s important is what the investigators
`who were carrying out the experiment might have thought it might
`be.
`
`Ex1223, 106:1-17. Dr. Furneaux further testified that he understood this number to
`
`be a moving target, and that it couldn’t be understood from the Board’s construction:
`
`“So the construct that the Board provided merely indicated nearly all, and what we’re
`
`saying here is 90 percent corresponds to nearly all. There isn’t anything in the
`
`Board’s construct that indicates how one arrives at that number which, for our
`
`purposes, is to clarify what ‘nearly all’ means.” Id., 107:3-13.
`
`20.
`
`In light of the discussion above, however, the record is clear that the
`
`appropriate denominator in such a calculation is 2.86 billion base pairs, the known
`
`length of the degapped reference genome. For example, Table 13 of the ’394 patent
`
`purports to show Personalis’s data using its “improved” techniques. The data for
`
`“whole genome” use a “DegappedGenome” length of 2,861 Mbp (2.86 billion base
`
`
`
`
`- 13 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`pairs), and report an “improved” whole genome sequencing assay which covers at
`
`most 98.7% of those bases, confirming that even Personalis’s “improvements” did
`
`not achieve sequencing “all” of the genome. Table 13 further shows that the patent
`
`considers sequencing 95.3% of 2.8 billion bases to be “improved” whole genome
`
`sequencing.
`
`
`(2) Personalis’s contemporaneous patents set the
`cutoff no higher than 50%.
`21. Dr. Furneaux’s 90% cutoff is far higher sequence coverage than a
`
`POSA would have required to consider a particular sequencing assay to be “whole
`
`genome sequencing.” Indeed, Personalis’s contemporaneous patents define “whole
`
`
`
`
`- 14 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`genome sequencing” as “untargeted sequencing,” and do not impose a hard
`
`boundary on the required sequence coverage. Indeed, Personalis’s contemporaneous
`
`patents reflect the knowledge and understanding of a POSA at that time, which
`
`recognized that even so-called “low coverage” whole genome sequencing
`
`techniques constituted “whole genome sequencing.” While I understand these
`
`patents may not necessarily qualify as prior art, I understand that non-prior art
`
`evidence of what was known can be relied on for indicating the level of ordinary
`
`skill in the art, what certain terms would mean to one with ordinary skill in the art,
`
`and how one with ordinary skill in the art would have understood a prior art
`
`disclosure.
`
`22. For example, in Personalis’s U.S. Patent No. 11,456,058, which claims
`
`priority to August 30, 2013, the inventors (which include John West), repeatedly use
`
`the term “whole genome sequencing” as synonymous with “untargeted” sequencing,
`
`as distinguished form “targeted” sequencing like “exome sequencing.” Ex1211
`
`(’058 patent). For example:
`
`• “Methods provided herein advantageously employ untargeted (e.g., whole
`genome) sequencing and target-specific sequencing to generate an output
`that indicates the presence or absence of one or more polymorphisms in the
`sample of the subject.” Id., 1:45-49.
`
`• “In another embodiment, the whole genome sequencing data comprises
`paired-end reads. In another embodiment, the whole genome sequencing
`
`
`
`
`- 15 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`data comprises mate-pair reads. In another embodiment, the paired-end reads
`have insert-sizes of larger than about 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9,
`1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0,
`9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 17.0, 18.0, 19.0, 20.0, 30.0, 40.0,
`50.0, 60.0, 70.0, 80.0, 90.0, or 100.0 kilobasepairs. In another embodiment,
`the mate-pair reads have insert-sizes of larger than about 0.1, 0.2, 0.3, 0.4,
`0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3, 1.4, [ 2] 1.5, 1.6, 1.7, 1.8, 1.9, 2.0, 3.0,
`4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 17.0, 18.0,
`19.0, 20.0, 30.0, 40.0, 50.0, 60.0, 70.0, 80.0, 90.0, or 100.0 kilobasepairs.”
`Id., 3:37-49.
`
`• “Untargeted sequencing (i.e., whole genome sequencing) can determine
`the complete DNA sequence of the genome at one time. Untargeted
`sequencing (i.e., whole genome sequencing or
`the non-exonic portion
`of whole exome sequencing) can cover sequences of almost about 100
`percent, or about 95%, of the sample's genome. In some cases, the
`untargeted sequencing (i.e., whole genome sequencing or
`non-exonic
`portion of
`the whole exome sequencing) can cover
`sequences of
`the whole genome of the nucleic acid sample of about or at least about
`99.999%, 99.5%, 99%, 98%, 97%, 96%, 95%, 94%, 93%, 92%, 91%, 90%,
`89%, 88%, 87%, 86%, 85%, 84%, 83%, 82%, 81%, 80%, 79%, 78%, 77%,
`76%, 75%, 74%, 73%, 72%, 71%, 70%, 69%, 68%, 67%, 66%, 65%, 64%,
`63%, 62%, 61%, 60%, 59%, 58%, 57%, 56%, 55%, 54%, 53%, 52%, 51%,
`or 50%.
`
`
`2 As I explain further in Section III, Leary’s whole-genome sequencing
`comprised mate-pair reads with insert lengths of 1,400 bp (1.4 kilobasepairs).
`
`
`
`
`- 16 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`
`In some cases, the output can have a coverage of about or at least about
`99.999%, 99.5%, 99%, 98%, 97%, 96%, 95%, 94%, 93%, 92%, 91%, 90%,
`89%, 88%, 87%, 86%, 85%, 84%, 83%, 82%, 81%, 80%, 79%, 78%, 77%,
`76%, 75%, 74%, 73%, 72%, 71%, 70%, 69%, 68%, 67%, 66%, 65%, 64%,
`63%, 62%, 61%, 60%, 59%, 58%, 57%, 56%, 55%, 54%, 53%, 52%, 51%,
`or 50% of the whole genome of the nucleic acid sample from a subject.” Id.
`at 17:33-55.
`
`• “The paired-end reads and/or mate-pair reads may sequence 100 basepairs
`(bp) reads on both ends of the insert. For example, when an untargeted
`sequencing data (e.g. whole genome sequencing data) is comprised of either
`paired-end reads with an insert size of larger than 1 kilobasepairs (kbp) or
`mate pairs with a separation of larger than 2 kbp, then even with low (˜1×)
`read depth, there may be at least about 10 molecules spanning any particular
`position on the genome, which may be sufficient to provide corroborating
`evidence of large CNVs detected by methods and systems presented herein.
`In another example, methods and systems of the present disclosure, such as
`the HMM method, can be used to detect a 50 kbp heterozygous deletion.
`When paired-end reads are used, about half of the paired-end molecules that
`span the detection breakpoints may have insert-sizes that are 50 kbp larger
`than normal. The insert-sizes may appear larger because they may be
`mapped to a reference with the sequence segment that is deleted in the
`sample. In some examples, when 1 kbp insert sizes are used, then 10 reads
`span the breakpoint, 5 of which can have anomalous insert size. Five reads
`may provide sufficient statistical power and information to generate accurate
`and/or reliable results. In some cases, if an insert-size of only about 300
`
`
`
`
`- 17 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`basepairs is used, then there may be only about 3 reads that span the
`breakpoint on average, and half of that may be about 1 read, which may not
`provide sufficient statistical power and information to generate accurate
`and/or reliable results.” Id. at 25:14-41.
`
`• Figure 3 “illustrates a statistical model on sensitivity vs. coverage of
`genomic regions with single run whole genome sequencing data,” id. 9:36-
`37, and depicts data confirming that ½x coverage “whole genome
`sequencing” was within the scope of the term. See Fig. 3 (annotated below).
`
`
`In Personalis’s U.S. Patent No. 11,584,968, which claims priority to
`
`23.
`
`10/30/2014, the inventor Mr. West similarly defined “whole genome sequencing” to
`
`include “low coverage whole genome sequencing.” For example, the ’968 patent
`
`repeatedly states that its whole genome sequencing comprises “untargeted
`
`sequencing data (e.g., low coverage whole genome sequencing data) and one or
`
`more target-specific data.” Ex1212 (’968 patent), 25:54-67; see also id., 26:6-9 (“In
`
`some cases, the methods as disclosed herein comprise untargeted sequencing data
`
`
`
`
`- 18 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`(e.g., a low coverage whole genome sequencing data).”); id., 18-20 (“In some cases,
`
`the untargeted sequencing (e.g., whole genome sequencing) in the methods as
`
`disclosed herein . . . .”). The ’968 patent cites on its face at page 4 “Pasaniuc, et al.,
`
`Extremely low-coverage sequencing and imputation increases for genome-wide
`
`association studies,” Ex1213, which describes “sequencing libraries appropriate for
`
`whole-genome sequencing” that are sequenced “at ultra-low coverage (0.1-0.5x[)],”
`
`consistent with the patent’s understanding that “whole-genome sequencing” refers
`
`to untargeted, as opposed to targeted (e.g., exome or specific sets of genes)
`
`sequencing.
`
`24. Personalis’s U.S. Patent No. 11,640,405, which claims priority to
`
`October 3, 2013, also used “whole genome sequencing” synonymously with
`
`“untargeted” sequencing.3 For example, the ’405 patent explains that “[t]he low
`
`coverage untargeted sequencing (i.e., single run whole genome sequencing data) can
`
`be fast and economical, and can deliver genome-wide polymorphism sensitivity in
`
`addition to the target-specific sequencing data. In addition, variants detected in the
`
`low coverage untargeted sequencing data can be used to identify known haplotype
`
`blocks and impute variants over the whole genome with or without targeted data.”
`
`Ex1214, 13:21-25. Like the ’058 patent, the ’405 patent expressly considers 50%
`
`
`3 The inventors of the ’405 patent include Gemma Chandratillake, Sarah
`Garcia, and Richard Chen, each of whom is named as an inventor or the ’394 patent.
`
`
`
`
`- 19 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`coverage whole genome sequencing to fall within the scope of “whole genome
`
`sequencing.” Id., 13:1-43; see id., 13:31-44 (“In some cases, the untargeted
`
`sequencing (i.e., whole genome sequencing or non-exonic portion of the whole
`
`exome sequencing) can cover sequences of the whole genome of the nucleic acid
`
`sample of about or at least about 99.999%, 99.5%, 99%, 98%, 97%, 96%, 95%, 94%,
`
`93%, 92%, 91%, 90%, 89%, 88%, 87%, 86%, 85%, 84%, 83%, 82%, 81%, 80%,
`
`79%, 78%, 77%, 76%, 75%, 74%, 73%, 72%, 71%, 70%, 69%, 68%, 67%, 66%,
`
`65%, 64%, 63%, 62%, 61%, 60%, 59%, 58%, 57%, 56%, 55%, 54%, 53%, 52%,
`
`51%, or 50%.”).
`
`(3) Extrinsic evidence cited by Dr. Furneuax sets the
`cutoff as low as 10%.
`25. Over several paragraphs of his declaration, Dr. Furneaux argues in
`
`effect, that because certain prior art literature did not expressly cite Leary in non-
`
`exhaustive lists of whole-genome sequencing studies, that its technique did not
`
`constitute whole genome sequencing. The logic of that argument does not follow,
`
`and in fact a careful review of the references he cites confirms (a) that Leary’s
`
`technique was understood to be whole genome sequencing; and (b) that the cutoff
`
`for “almost all of the genome” would need to be around 10% in order to encompass
`
`methods considered to be whole genome sequencing as of the priority date.
`
`
`
`
`- 20 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`26. First, references, such as Forshew (Ex1030) and Diehn (Ex1112) (filed
`
`March 15, 2013) affirmatively describe Leary as a whole genome sequencing paper.
`
`Ex1030, 10 (“Previously proposed methods for personalized monitoring of tumor
`
`dynamics relied on . . . identification of rearrangements using whole genome
`
`sequencing (18 [Leary], 19 [McBride]).”); Ex1112, [0003] (“Other studies have
`
`proposed identifying patient-specific chromosomal rearrangements in tumors via
`
`whole genome sequencing (WGS), followed by breakpoint qPCR from cfDNA
`
`(Leary et al. (2010) Sci. Transl. Med. 2:20ra14; McBride et al. (2010) Genes Chrom.
`
`Cancer 49:1062-1069)”).
`
`27. Dr. Furneaux notes that “Meyerson (EX1031) has a table that
`
`summarizes ‘whole-genome sequencing studies of cancer.’” Ex2031 ¶106. He states
`
`that “Leary is cited as reference 99 in Meyerson, but it is not included in the list of
`
`whole genome sequencing studies in Table 1.” Id. Meyerson’s Table 1 is reproduced
`
`below:
`
`
`
`
`- 21 -
`
`Foresight EX1225
`Foresight v Personalis
`IPR2023-00224
`
`

`

`IPR2023-00224 (Patent 11,384,394)
`IPR2023-00317 (Patent 11,408,333)
`
`
`
`
`28. Meyerson’s list does not purport to be exhaustive, and Dr. Furneaux
`
`does not dispute in his declaration that the listed studies were understood at the time
`
`to constitute whole genome sequencing studies. Indeed, as a review article published
`
`in 2010, Meyerson is an informative snapshot into what those skilled in the art
`
`thought at the time. Importantly, Meyerson cites Campbell 2008 and Stephens 2009,
`
`each shallow paired end whole-genome sequencing studies, in its list of “Whole-
`
`genome sequencing studies.”
`
`29. Notably, Campbell 2008 used the same whole-genome sequencing
`
`method as McBride 2010, which Dr. Furneaux agrees used a “similar paired-end
`
`rearrangement screen” to Leary. Ex1015, 3 (“Samples of genomic DNA were
`
`extracted from fresh-frozen