US 5525464

I'll search for authoritative records on this specific patent number, including its claim set (which isn't included in the text supplied to me) and any 2026 litigation.

Now let me check for any 2026 CAFC litigation or docket activity involving this patent, plus verify the claim set and reexamination history.

Let me verify the claim set against another source and check for any reexamination/certificate of correction affecting this patent.

Search summary — US Patent 5,525,464

Scope note: I ran targeted searches for the literal number 5525464 (and the formatted variants 5,525,464 / US5525464) against Google Patents, Espacenet, uspto.report's patent-grant mirror, and general CAFC/litigation sources. I did not find any 2026 Federal Circuit docket, appeal, or district-court activity naming this patent. I also note a date inconsistency: the brief says today is April 26, 2026, but search results in the same session include items dated September 11 and September 27, 2026 — so the stated "current date" appears stale by ~5 months. I flag that rather than silently reconciling it.

---

Bibliographic data (high confidence)

Family/later filings citing it include US 5,695,940, 5,972,619, 6,270,961 and 6,316,191 (all Hyseq, same specification). A 2004 reassignment record shows a security agreement involving Nuvelo, Inc. and Affymetrix, Inc. as assignees of Callida Genomics, Inc. — this is a security-interest record, not a transfer of title in the '464 patent.

Abstract (verbatim, high confidence)

"The conditions under which oligonucleotide probes hybridize preferentially with entirely complementary and homologous nucleic acid targets are described. Using these hybridization conditions, overlapping oligonucleotide probes associate with a target nucleic acid. Following washes, positive hybridization signals are used to assemble the sequence of a given nucleic acid fragment. Representative target nucleic acids are applied as dots. Up to to 100,000 probes of the type (A,T,C,G)(A,T,C,G)N8(A,T,C,G) are used to determine sequence information by simultaneous hybridization with nucleic acid molecules bound to a filter. Additional hybridization conditions are provided that allow stringent hybridization of 6-10 nucleotide long oligomers which extends the utility of the invention. A computer process determines the information sequence of the target nucleic acid which can include targets with the complexity of mammalian genomes. Sequence generation can be obtained for a large complex mammalian genome in a single process."

(The doubled word "to to" is in the official abstract text; I have not corrected it.)

---

Claim-set caveat

The patent text supplied in the prompt stops before the claims (it ends mid-sentence in Appendix A). I therefore relied on the claim text reproduced on uspto.report's grant record (https://uspto.report/patent/grant/5,525,464), which appears to be a faithful transcription of the printed patent. Confidence in claim numbering and substance: moderate-to-high, but not verified against the USPTO PatentCenter/patentimages PDF in this session. Treat the claim paraphrase below as needing confirmation against the official copy.

---

Plain-language overview of the independent claims

Claim 1 — the core sequencing method. A method of sequencing a target nucleic acid of unknown sequence with six steps: (a) use hybridization conditions that differentiate an exactly complementary probe from a probe with a 5′ or 3′ terminal mismatch; (b) contact the target with a plurality of oligonucleotides each at least six nucleotides long; (c) form duplexes; (d) wash; (e) detect the oligonucleotides that positively hybridized; and (f) compile the target sequence from overlapping positively hybridizing oligonucleotides. The novelty here is the combination of end-mismatch discrimination with probes as short as hexamers plus assembly-by-overlap.

Claim 7 — whole-genome SBH. Independently claims sequencing by hybridization of a complete genomic DNA (or large portions) by hybridizing multiple genomic fragments with all or a portion of oligonucleotide probes 8–20 nucleotides long (built from A, T/U, C, G and their derivatives and analogs), under conditions where the probes hybridize to a fully homologous portion, or to a portion with fewer mismatches than would produce ambiguous or erroneous sequence on assembly by maximum mutual overlap. The claim is framed around an error-tolerance criterion rather than a specific temperature/salt regime.

Claim 8 — probe-set selection method. Independently claims a method of selecting a set of non-identical probes of predetermined length such that together they represent a continuous linear sequence of the target: (a) hybridize a set of non-identical probes to the target DNA; (b) identify a first hybridizing probe; (c) step-wise identify subsequent probes that hybridize immediately 5′ or 3′ adjacent to the segment bound by the previously identified probe; and (d) select that set. This is essentially the "walking by overlap" selection logic, claimed as a method in its own right.

Dependent claims 2–6 (all depend from claim 1)

• Claim 2 — target is cloned genomic DNA in three subclone libraries; inserts consist of two parts separated on average by 50–200 kb, sized 0.1–1.0 kb or 3.0–10 kb; spotted on a filter; vector-insert DNA intact or sheared to 20 bp.
• Claim 3 — target is genomic fragments amplified in vitro with DNA polymerase using ~5 to ~200 primers.
• Claim 4 — the compiling step uses cyclic detection of overlapping subclones and orders subfragments with a linear displacement of less than 100 bp.
• Claim 5 — the compiling step uses competitive hybridization with labeled plus saturating unlabeled oligonucleotides to resolve repeated subfragments.
• Claim 6 — extremely specific: ≥1 million bp of mammalian DNA in 1,250 sample groups (avg. 20 × 0.5 kb M13 subclones, 700 × 7 kb M13 subclones, 170 reconnecting/"jumping" M13 subclones spanning ~100 kb), probed with four defined probe-group sets — 1,024 groups of 16 (A,T,C,G)N10(A,T,C,G) probes (N10 = 10-mers lacking G and C); 23,040 groups of 16 (A,T,C,G)N9(A,T,C,G) probes (N9 = 9-mers with one or two G+C); 55,834 groups of 64 (A,T,C,G)(A,T,C,G)N8(A,T,C,G) or 256 (A,T,C,G)(A,T,C,G)N8(A,T,C,G)(A,T,C,G) probes (N8 = 8-mers with ≥3 C+G); and 3,725 groups of 16 (A,T,C,G)Nm(A,T,C,G) probes (Nm = monotonous sequences <18 bp built of 1–7 nt repeat units). This claim is effectively a numerical "recipe" claim and is the most likely to be narrow/avoidable.

---

Litigation status

• Historic: the Darts-IP family record flags this family as having worldwide litigation. The best-documented U.S. action is Hyseq, Inc. v. Affymetrix, Inc., complaint filed March 4, 1997 in the Northern District of California, asserting Hyseq's sequencing-by-hybridization patents (Hyseq also contested the Southern/Isis European patent). See the contemporaneous account in Nature Biotechnology 15:406 (May 1997). US 5,525,464 appears in subsequent patent-owner/applicant citation tables tied to Affymetrix and Hyseq/Complete Genomics proceedings (e.g., Complete Genomics v. Illumina, D. Del. 1:19-cv-00970), but only as a cited reference, not as an asserted patent.
• 2026 CAFC dockets: no results. I found no Federal Circuit appeal, docket entry, or 2026 opinion referencing patent 5,525,464. Given that the patent expired in 2013 (and the term after April 13, 2010 was disclaimed), any 2026 appellate activity would be surprising. I cannot rule out an obscure 2026 filing, but I have no authoritative evidence of one and will not assert one exists.

Uncertainty flags

1. Claims text is sourced from a third-party mirror (uspto.report), not the official USPTO PDF — verify numbering/text before relying on it.
2. Claim count of 8 is my best available figure; confirm against PatentCenter.
3. I could not confirm whether any ex parte reexamination certificate issued for 5,525,464 itself. The reexamination certificate I did find (US 6,355,432 C1, issued Dec. 30, 2008, in Fodor et al.) merely cites 5,525,464 as prior art.
4. No claim construction, invalidity holding, or CAFC merits decision on this patent was located in this session.

I'll search for litigation involving US patent 5525464 across patent litigation databases.

Let me dig deeper into the litigation leads, particularly the Affymetrix-related suits.

Let me search for additional details and any other defendants or cases.

Litigation involving U.S. Patent No. 5,525,464

Bottom line: U.S. 5,525,464 ("Method of sequencing by hybridization of oligonucleotide probes," Drmanac et al., Hyseq, Inc.) was asserted in the Hyseq v. Affymetrix litigation saga (1997–2001) in the [U.S. District Court for the Northern District of California, San Jose Division](/courts/northern-district-of-california-san). That litigation consisted of three civil actions, all resolved by a single comprehensive settlement on October 24, 2001 (publicly announced by Affymetrix on November 1, 2001). I found no other judicial litigation asserting or challenging this patent.

Cases

Outcome / settlement terms (relevant to '464)

Under the settlement, the parties agreed to seek dismissal with prejudice of all three pending lawsuits and acknowledged the validity and enforceability of the asserted patents, which included Hyseq's U.S. Patent Nos. 5,202,231, 5,525,464, 5,695,940, 6,018,041 and 5,972,619 and Affymetrix's U.S. Patent Nos. 5,795,716, 5,744,305 and 5,800,992. Hyseq also agreed to abandon its contest in USPTO Interference No. 104,552 (Chee v. Drmanac). Hyseq created a subsidiary, Callida Genomics, Inc., contributed its SBH patents (including the '464 patent) to it, and entered cross-licenses with Affymetrix.

Related (non-litigation) proceedings

• USPTO Interference No. 104,552, Chee v. Drmanac — patent interference involving Hyseq's SBH portfolio (family of the '464 patent); Hyseq filed an abandonment of contest as part of the 2001 Affymetrix settlement.
• 2004 security/assignment record: NUVELO/ AFFYMETRIX security agreement recorded against Callida Genomics (an assignment/security matter, not a lawsuit) — reflected on the patent's public record.

Caveats and verification notes

• The Google Patents page for US 5,525,464 carries a "Family has litigation" flag (Darts-IP worldwide family litigation dataset), which is consistent with the three U.S. cases above. I could not retrieve the Darts-IP record itself, so I cannot confirm whether any non-U.S. (foreign) counterpart litigation exists beyond the U.S. cases listed.
• I was unable to load the Unified Patents litigation portal (portal.unifiedpatents.com/litigation/caselist) for this specific number in my searches; my findings are grounded in SEC filings (Affymetrix Form 8-K, dated November 1, 2001, and Hyseq Form 10-K for FY2001), the settlement agreement excerpts, and secondary reporting (BioCentury).
• Case number C 99-21163 JF — I did not independently confirm the exact filing date within 1999; the case number and plaintiff/defendant orientation are confirmed from the Affymetrix 8-K.
• Some secondary sources describe Hyseq's December 9, 1997 filing as a "second suit"; the formal case list in the 8-K identifies only the three actions above, so the '940 assertion may have been folded into C 97-20188. I have presented the case numbers as listed in the primary settlement/filing documents rather than resolving that discrepancy.

Sources: Affymetrix Form 8-K (Nov. 1, 2001) (capedge.com/filing/913077/0000912057-01-537335/AFFX-8K); Hyseq, Inc. Form 10-K, Item 3 (sec.gov); Hyseq Settlement Agreement (lawinsider.com/contracts/aOyAP0Wdsyz); BioCentury reporting on the Hyseq/Affymetrix Markman order.

I'll verify the absence of PTAB activity against public sources before writing the report.

Proceedings overview

Zero AIA trial proceedings — the canonical USPTO Open Data Portal list for US 5,525,464 contains no IPR, PGR, or CBM petitions (no institutions, no denials, no settlements, nothing to invalidate or sustain), and targeted web searches for a proceeding on this patent number returned no hits; the bottom-line defensive posture is therefore not "the patent has been hardened by surviving IPRs" but rather "the patent is expired and has never been tested by the PTAB — you will not be facing an AIA-trial record either way." The absence is explainable rather than anomalous (see the reasons below), so do not read it as evidence that the claims are strong or weak; it is evidence that nobody had a reason to pay for an IPR.

No proceedings to enumerate

The task template assumes at least one proceeding. There is none. Rather than invent filler entries, here is what I actually verified and what the closest analogues are.

Verification performed:

• Canonical source (ODP): the structured "PTAB proceedings on file" block returns no AIA trial proceedings. This is the controlling list.
• Web search for a proceeding on US 5,525,464: no IPR/PGR/CBM petition, institution decision, final written decision, or appeal surfaced under any of IPR "5,525,464", "5525464" PTAB, or petitioner-name queries against the likely challengers (Affymetrix, Illumina, Complete Genomics, Applied Biosystems).
• Sibling SBH patents searched for spillover: I also searched for AIA proceedings against the closely related Drmanac/Hyseq/Callida patents in the same family (US 5,202,231; US 5,695,940; US 5,972,619; US 6,316,191; US 6,451,996). No AIA proceeding was identified against any of them on this search. Flag: this sibling check was not exhaustive, and I did not query PTAB E2E party-by-party for Callida Genomics or Hyseq, so treat the sibling result as unverified.

Caveats / do not fabricate:

• I could not directly query PTAB E2E (https://ptab.uspto.gov/) or the USPTO patent-specific AIA trial API from this environment. The negative result rests on the ODP block plus web search. Given the patent's status below, a newly-filed-but-unindexed proceeding is close to impossible, but I cannot state that as a certainty.
• I did not enumerate the claim set of US 5,525,464 (independent vs. dependent claim count) from the retrieved record, and the Google Patents text fetched for this analysis did not reproduce the claims verbatim. Do not rely on this report for claim-number granularity on '464. If claim-level precision matters, pull the PDF from https://patents.google.com/patent/US5525464/en.

Close analogues that are not AIA trials (do not conflate)

These came out of the previously generated sections and my verification pass. They are pre-AIA or Article III proceedings and carry no § 315(e) estoppel:

1. USPTO Interference No. 104,552 — Chee v. Drmanac. A pre-AIA interference under 35 U.S.C. § 135, not an IPR/PGR/CBM. Hyseq filed an abandonment of contest as part of the 2001 Affymetrix settlement (per the litigation summary in this file). An interference loss/abandonment affects priority, not patentability over prior art in the § 102/§ 103 sense, and creates no IPR estoppel.
2. Hyseq v. Affymetrix, N.D. Cal. (C 97-20188 RMW; C 99-21163 JF; C 00-20050 RMW). Affymetrix's December 11, 1997 counterclaim challenged validity as a defense/counterclaim in district court, and all three cases were dismissed with prejudice under the 2001-10-24 settlement in which both sides acknowledged the validity and enforceability of the asserted patents, including the '464 patent. That acknowledgment is a contractual estoppel running between those parties only. It is not a PTAB outcome, not a Federal Circuit affirmance, and not binding on a defendant today.
3. Statutory expiration: 2013-06-11 ("Anticipated expiration," legal status "Expired - Lifetime" on the Google Patents record; the term was capped by the earlier-expiring US 5,202,231 and US 5,525,464 per the terminal-disclaimer-style notice carried on US 5,695,940). Because the '464 patent expired roughly nine months after the first AIA petitions were even permitted to be filed (AIA trials available from 2012-09-16), the practical window for a cost-justified IPR on '464 was nil.

---

Strategic summary

Claim status. This is the important divergence from the usual IPR report. There are no PTAB-canceled claims and no PTAB-sustained claims on US 5,525,464 — the entire claim set is UNTESTED by the PTAB. It is not "narrowed by IPR." It is also not "hardened by surviving IPRs." The correct label is untested and expired. The only adjudicative-style outcomes on the record are (i) Affymetrix's 1997 invalidity counterclaim, which was never tried to judgment and was resolved by the 2001 settlement in which Affymetrix acknowledged the '464 patent's validity and enforceability, and (ii) the abandoned interference contest. I am not able to state which individual claims (independent vs. dependent) were asserted in the 1997–2001 litigation beyond the patents named; the litigation summary above identifies patents, not claims, and I have not verified claim numbers.

Estoppel landscape. There is no § 315(e)(2) IPR estoppel to map, because there has been no IPR petitioner. No party is barred from anything by AIA estoppel. Two non-AIA constraints do exist: (a) the 2001 settlement's validity acknowledgment binds Hyseq/Callida and Affymetrix and their successors as a matter of contract (terms largely confidential; a security agreement recorded in 2004 against Callida Genomics naming Nuvelo/Affymetrix appears on the public record) — check whether you are a successor/privy before assuming you can raise invalidity against an affiliate of those parties; and (b) issue-preclusion principles from the interference, if any determinations were reached, but the contest was abandoned, so I would not expect a preclusive effect. For an ordinary third-party defendant today, every prior-art ground — § 102, § 103, and § 112 — remains fully available, subject only to the ordinary rules of evidence and estoppel from your own prior litigation.

Pattern signals. Three signals worth noting and one non-signal: (1) Same-petitioner pattern: none — no petitioner ever filed on '464, so there is no serial-IPR filer to watch. (2) Patent owner PTAB aggression: none — Hyseq/Callida never had a PTAB proceeding to appeal, hence no Federal Circuit PTAB appeals from this patent; the only Federal Circuit-adjacent activity would be from the 1997–2001 district court case. (3) Defensive aggregator: no evidence. The Unified Patents portal surfaced in search results for a different patent bearing Hyseq as an assignee (US 6,943,034 B1, priority 1987-03-31), which is a listing page rather than a challenge; I found no Unified Patents IPR, CBM, or ex parte reexamination filing on '464. Treat this as a weak signal only — I could not confirm it either way. The actual pattern signal here is expiration, not litigation.

The point that dominates everything. US 5,525,464 expired 2013-06-11. Patent infringement liability attaches only to acts before expiration, and damages under 35 U.S.C. § 286 reach back only six years — i.e., no earlier than approximately 2020-09-30 as of today (2026-09-30). Those two windows do not overlap. Consequently, there is no live infringement theory available against the '464 patent at all today, regardless of claim construction, invalidity, or PTAB posture. An IPR is a mechanism to knock out claims you are being sued on; nobody needs to knock out claims they cannot be sued on. That is almost certainly why the Darts-IP "Family has litigation" flag on the Google Patents page reflects only the 1997–2001 Hyseq/Affymetrix cases and no PTAB or later actions.

---

Recommended next steps

• If you are a defendant or a demand letter recipient: the dispositive response is not an IPR petition — it is the expiration date. Put the "Anticipated expiration: 2013-06-11 / Expired - Lifetime" record and the § 286 six-year limitation side by side and ask the asserting party to identify (a) any accused act occurring on or before 2013-06-11 and (b) the basis for damages given the § 286 bar. Request that in writing before spending anything on prior art.
• Do not file an IPR on '464. It would be an economic nullity: even a full win cancels claims you cannot infringe, and 35 U.S.C. § 315(e)(2) estoppel would then constrain your other district-court invalidity arguments in any case where an affiliate of the patent owner is asserting a live sibling patent. Preserve your grounds.
• Redirect the invalidity workstream to live family members. Callida Genomics / Hyseq's SBH portfolio (e.g., US 6,451,996; US 7,232,093; the random-array lineage including US 8,785,127 and its continuations) is where assertion pressure could actually exist. Those are unexpired or expired far more recently and are the proper IPR candidates. I have not verified whether any of them has pending litigation or PTAB activity — that requires a fresh, patent-specific check on PTAB E2E (https://ptab.uspto.gov/) and the Unified Patents litigation portal.
• If an active proceeding on a sibling patent is pending: PTAB trial milestones run on the statutory clock — institution decision due within 6 months of the preliminary response (35 U.S.C. § 314(b)), and final written decision within 12 months of institution (§ 316(a)(11)), extendable up to 6 months for good cause. Verify actual due dates in PTAB E2E rather than estimating.
• Verify before relying: (i) the full claim set of '464 (not reproduced in the fetched text), (ii) whether any collateral estoppel attaches from the abandoned interference, and (iii) the confidentiality status of the 2001 settlement terms if you are any successor-in-interest to Hyseq, Callida, Affymetrix, Nuvelo, or their privies. My report cannot substitute for those three checks, and I am explicitly flagging them as unresolved rather than guessing.

Sources consulted: USPTO Open Data Portal "PTAB proceedings on file" block (canonical; no AIA trials); Google Patents US 5,525,464 (https://patents.google.com/patent/US5525464/en, expiration and legal-status fields); PTAB E2E landing page (https://ptab.uspto.gov/, not directly queried from this environment); the previously generated litigation summary and its cited Affymetrix 8-K and Hyseq 10-K. No PTAB opinion or Federal Circuit opinion exists to link for this patent, because none was issued.

I'll research the assignment record for US 5525464 using USPTO Assignment Search, Google Patents legal events, and SEC filings.

I'll search for the authoritative citation records for US 5,525,464, focusing on its front-page prior-art references.

Let me pull the front-page reference list directly from patent record mirrors.

Let me retrieve the front-page reference list from the official patent-images PDF and the Google Patents citation tables.

Search status — important limitation first

I must be upfront: I could not retrieve the actual front-page "References Cited" block (the USPTO patent-document citation list) for US 5,525,464. This matters because your request is specifically "look at each patent citation for 5525464."

What I did and what happened:

So the authoritative data I can ground is the citation apparatus that is inside the specification body, which the full patent text in your prompt reproduces verbatim (i.e., the non-patent literature the applicant cited as background). I will analyze that set carefully and be explicit about where I am inferring versus stating.

I will not fabricate a front-page U.S. patent-document list. If you need the literal "U.S. Patent Documents" column, it must be pulled from the PDF (patentimages.storage.googleapis.com/.../US5525464.pdf) or PatentCenter — see verification guidance at the end.

---

⚠️ STRICT-IDENTIFIER WARNING — a same-number trap

Per the strict rule, I must flag this rather than silently reconcile it: my searches surfaced extensive PTAB/ITC material referring to "the '464 Patent." In those documents — e.g. IPR2020-00813 and an ITC complaint excerpt (ptacts.uspto.gov/.../1547091, and 1534921) — "the '464 Patent" is U.S. Patent No. 9,859,464 (a light-emitting-device patent, LED chip on a transparent plate/lead frame, claims to "an LED chip" and "a phosphor for converting the light"). That is a completely different patent with no relationship to US 5,525,464.

I am excluding all of that material from the prior-art analysis below. It is a false positive produced by the shared trailing digits "464." Any analyst searching "464 + patent" will hit it.

---

Forward citations ≠ prior art (disambiguation)

The same searches surfaced patents that cite US 5,525,464 (i.e., it is their prior art, not the reverse):

• US 9,523,125 (Drmanac) — lists "5,525,464 A 6/1996 Drmanac et al." in its References Cited (patentimages.storage.googleapis.com/.../US9523125.pdf)
• US 10,726,942 (Drmanac et al.) — same (.../US10726942.pdf)
• US 9,650,673 (.../US9650673.pdf)

These are later documents (post-1996) and therefore cannot be §102 prior art against the '464 patent. They are listed here only so they are not misclassified if you see them in a citation graph. Direction of the arrow matters.

---

The prior art actually cited in the '464 specification (grounded in your prompt's authoritative text)

Below is every reference the '464 specification itself cites by name, as it appears in the patent text you supplied. Full citations are reproduced as printed (including the patent's own typographical errors, which I flag rather than correct, per the strict rule).

A. Hybridization / discrimination art — most relevant to Claim 1

B. Sequencing methodology art — background to Claim 1's "assembly" element

C. Genome-mapping art (cited in §2.4) — relevant to Claims 2, 4, 7 (whole-genome claims)

---

§102 anticipation analysis — the honest answer

No single one of the cited references anticipates any claim of US 5,525,464 standing alone. Anticipation under 35 U.S.C. § 102 requires one reference disclosing every element arranged as in the claim. The independent claims are combination claims, and none of the cited art supplies the combination:

• Claim 1 requires, in combination: (a) conditions discriminating an exactly complementary probe from a probe with a 5′ or 3′ terminal mismatch; (b) probes at least six nucleotides long; (e) detection; and (f) compilation of the target sequence by overlapping positives.

  • Wallace (1979) supplies only internal-mismatch discrimination at 11–17 nt — the spec itself (§2.2) says eleven-mers are the shortest previously workable. It fails elements (a) [end vs. internal; length] and (f) [no overlap assembly]. → Anticipates nothing. Closest §103 art for element (a).
  • Wood (1985) supplies length-controlled Tm at 11–20 nt but no discrimination of a terminal mismatch, no sub-11-mer teaching, no assembly. → Anticipates nothing.
  • Poustka et al. (1986, CSH) supplies hybridization of ~100 oligos to a cosmid array to identify clones, but expressly does not yield sequence (the spec: "Such identification however does not provide desired and useful sequence information"). → Does not anticipate Claim 1; relevant only to the array/format concept.
  • Maxam-Gilbert (1977) / Sanger (1977) are the alternate sequencing paradigm; they disclose no hybridization step. → Anticipates nothing.
  • Saiki (1986) discloses PCR amplification (relevant to Claim 3's "amplified in vitro with DNA polymerase") but nothing about hybridization-based sequence compilation. → Does not anticipate Claim 3; single-element relevance only.

• Claim 7 (whole-genome SBH, 8–20 nt probes, mismatch tolerance framed by assembly error). Wood (1985) reaches 8–20 nt on the low end only at 11; Poustka (1986) uses arrays of oligos but not for sequencing. No single reference discloses the whole-genome-assembly criterion. → No anticipation.

• Claim 8 (method of selecting a set of non-identical probes by step-wise identifying 5′/3′ adjacent hybridizing probes). This "walking-by-overlap selection" logic is not disclosed by any cited reference. → No anticipation.

Bottom line: the cited art is best characterized as §103 (obviousness) material, not §102 material, and even then only as a mosaic. The classic §102-type risk to this family came not from the cited references but from post-priority parallel disclosures — see below.

---

The prior art the examiner did not cite (context you should know)

Because the '464 patent's effective priority is April 1, 1987 (YU P-570/87), the following independent co-inventions of sequencing-by-hybridization all published after that date and therefore are not anticipatory §102(a)/(b) art against the earliest priority — but they are the references that matter for priority/validity attacks and for the interference history already summarized in your prior section:

• Bains & Smith, "A novel method for nucleic acid sequence determination," J. Theor. Biol. 135: 303–307 (1988)
• Lysov et al., Dokl. Akad. Nauk SSSR 303: 1508 (1988)
• Southern, "Analyzing polynucleotide sequences," WO 89/10977 / UK priority 1988

None of these appear in the specification citations above. I flag them as my own domain knowledge, not as items verified from the record in this session — confirm each against the file wrapper before relying on it.

---

Verification guidance (since I could not pull the front page)

To obtain the literal "References Cited — U.S. Patent Documents" column for 5,525,464, use (in order of authority):

1. USPTO PatentCenter → Application 08/203,502 → "Documents & Transactions" → the printed patent PDF front page.
2. The patent images PDF — the Google Patents record links a "Download PDF" whose storage URL follows the patentimages.storage.googleapis.com/…/US5525464.pdf pattern; the front page carries the examiner's citation list, the Primary Examiner, and the Attorney, Agent or Firm field.
3. FreePatentsOnline (freepatentsonline.com/5525464.html) — renders the front-page reference table in HTML.
4. The Certificate of Correction flagged on uspto.report (referenced in your prior summary) — retrieve the text; it may renumber or amend claims, which would change any §102 mapping.

---

Explicit uncertainty flags

1. I did not obtain the front-page U.S.-patent-document citation list for 5,525,464 in this session. Everything in the tables above comes from the specification body, which your prompt supplies as authoritative text. Do not treat my tables as the examiner's citation list.
2. Claim text remains unverified (carried forward from your prior section's caveat). My §102 mapping is against the paraphrased independent/dependent claims; confirm against the official claims before filing anything.
3. Two references contain printed author-name typos ("Bresslauer," "Tangiuchi") that I reproduced literally rather than correcting, per the strict-identifier rule. The conventional renderings are Breslauer and Taniguchi, but the patent prints them as shown.
4. The "date inconsistency" flagged in your prior section persists: the brief states today is April 26, 2026, while session search results carried later-2026 datelines. I note it rather than reconcile it.
5. No 2026 CAFC activity on 5,525,464 was found (consistent with your prior litigation section and with the patent's 2013 expiry). Nothing in this session changes that.

If you can supply the front-page PDF text (or grant the equivalent), I will re-run the §102 mapping reference-by-reference against the examiner's actual citation column — that is the one deliverable I could not complete here.

I'll gather and verify the prior art references before building the §103 analysis. Note: the "Prior Art section" results referenced in the task were not actually included in my context, so I'll flag that and ground the analysis in (a) the references cited in the patent's own specification (authoritative, from the full text provided) and (b) targeted searches to verify those references and identify additional art.

Now let me verify additional references cited in the patent's specification and probe for any validity challenge (reissue/reexam/court invalidity holding) affecting the '464 claims.

Obviousness Analysis — U.S. Patent No. 5,525,464 under 35 U.S.C. § 103

0. Scope note — the referenced "Prior Art section" was not provided

The task instructs me to "use the results from the Prior Art section of this page." No Prior Art section was included in my context. I will not fabricate one. Instead, this analysis is grounded in:

1. The prior art cited within the '464 specification itself — authoritative, because the full patent text was supplied and the Background section is a written admission of what the art taught as of 1987. Under pre-AIA § 103, an applicant's own specification background is usable as evidence of the state of the art (and cited references are admissions).
2. Search-verified bibliographic data for each of those references, so dates and teachings are grounded rather than recalled.
3. Three references not cited in the '464 specification but critical to any modern invalidity theory: Bains & Smith (1988), Lysov et al. (1988), and Khrapko et al. (1989).

I also continue to flag the date inconsistency raised in the earlier Patent-summary section (the stated "current date" of April 26, 2026 is contradicted by same-session search results dated September 2026). I do not reconcile it silently.

---

1. Legal framework and the single dispositive issue: the priority gate

1.1 The claim set

From the earlier Claim-set section (sourced from the uspto.report grant mirror — still a moderate-to-high, not verified, transcription), the '464 patent carries 8 claims, three independent:

• Claim 1 — a method of sequencing an unknown target, comprising (a) end-mismatch-discriminating hybridization conditions, (b) contact with a plurality of probes each at least six nucleotides long, (c)(d) duplex formation and washing, (e) detection, and (f) compiling the target sequence from overlapping positively hybridizing oligonucleotides.
• Claim 7 — whole-genome/large-fragment SBH using probes 8–20 nt under an error-tolerance criterion (mismatches permitted so long as assembly by maximum mutual overlap stays non-ambiguous).
• Claim 8 — a method of selecting a probe set by step-wise identification of 5′/3′-adjacent hybridizing probes.
• Claims 2–6 — dependent: sub-clone/jumping library preparation (2), PCR amplification (3), cyclic overlap ordering (4), competitive hybridization (5), and a very specific numerical recipe (6).

1.2 The priority gate

This is where the entire § 103 analysis turns.

Two consequences follow:

(a) The pre-1987 references are prior art regardless. Wallace 1979, Wood 1985, Estivill (published 1987-02-25), Ikuta 1987, Smith 1983, Besmer 1972, Porschke & Eigen 1971, Craig et al. 1971, Breslauer 1986, Asseline 1984, and Saiki 1986 are all § 102(b) art measured from any of the possible effective dates (1987, 1988, 1990, or 1994).

(b) Bains & Smith and Lysov et al. are prior art only if the claims lose the 1987–1988 priority.

• Bains & Smith, A novel method for nucleic acid sequence determination, J. Theor. Biol. 135:303–307, published 7 December 1988 (https://doi.org/10.1016/S0022-5193(88)80247-9).
• Lysov, Florentiev, Khorlin, Khrapko, Shik & Mirzabekov, Dokl. Akad. Nauk SSSR 303(6):1508–1511 (1988) (https://pubmed.ncbi.nlm.nih.gov/[3250844](/patent/3250844)/).
• *Khrapko et al., FEBS Lett. 256:118–122 (1989)* — the experimental reduction to practice of the Lysov scheme.

For Bains & Smith to be § 102(b) art, the effective U.S. filing date must be later than 7 December 1989. That is satisfied only if the claims are entitled to no earlier than the 1990-08-31 (07/576,559) filing — i.e., if the 1987/1988 disclosures do not provide § 112 support for the claimed subject matter. This is a CIP chain (07/175,088 → 07/576,559 → …), which means new matter plausibly entered the chain — the 6-mer/7-mer/8-mer empirical data (Examples 6.3–6.4) reads as later-added material.

Practical consequence: any obviousness theory against the '464 claims resolves into two regimes:

• Regime A (priority holds, effective date ≈ 1987-04-01): Bains & Smith, Lysov and Khrapko are off the table. The invalidity case relies on Wallace + Wood + Ikuta + Estivill + Smith.
• Regime B (priority slips to 1990): Bains & Smith, Lysov and Khrapko are squarely § 102(b) art, and the obviousness case becomes very strong.

I analyze both.

---

2. Verified prior-art references and their teachings

Note the inventors' own pre-priority publication — *Drmanac et al., Nucl. Acids Res. 14:4691–4692 (1986)* — appears in later-Hyseq reference lists. I could not retrieve its content in this session, so I do not rely on it; if it discloses gapped/segmented probe strategies, it does so as an inventors' own § 102(b) publication, which would be a self-inflicted statutory bar.

---

3. Combination 1 — Regime A (priority holds): Wallace + Wood + Ikuta (+ Estivill)

3.1 What this combination reaches

Claim 1 elements (a)–(e) are squarely met:

• (a) "hybridization conditions that differentiate an exactly complementary probe from a probe with a 5′ or 3′ terminal mismatch" → Wallace '79 (adjust hybridization or wash temperature; mismatched Td ~10 °C lower) + Ikuta '87 (kinetic/washing model).
• (b) "a plurality of oligonucleotides each at least six nucleotides long" → Wood '85 (pools of probes, length-based stringency) and Estivill '87 (a single 8-base probe).
• (c)–(e) duplexing, washing, detection → Estivill '87 (dot/blot hybridization, wash at 0 °C then 27 °C, autoradiographic detection) and Saiki '86 (oligo dot-blot detection).

Claim 1 element (f) — "compiling the target sequence from overlapping positively hybridizing oligonucleotides" — is NOT reached by Wallace + Wood + Ikuta + Estivill. None of them performs de novo sequencing by assembly of a hybridizing probe set. Estivill uses the octamer to map rare restriction sites, not to read sequence.

Therefore: in Regime A, Combination 1 does not render claim 1 obvious. That gap is the patent's central inventive contribution, and it is precisely where Bains & Smith / Lysov would have supplied the missing element had they been available.

3.2 The motivation to combine (for whatever it is worth)

Even though the combination fails on element (f), the motivation analysis is worth stating because it shows how close the art was to the claimed method:

• Wood '85 is an express invitation to scale up. It frames its technique as solving the problem of screening "highly complex libraries" with a "pool of oligonucleotide probes" that "usually vary widely in base composition," so that stringency is "controlled as a function of probe length only." A POSA pursuing a massively parallel hybridization readout would seize on exactly this teaching.
• Wallace '79 and Ikuta '87 supply the discrimination mechanism and its optimization parameter (wash temperature/time), and Wallace itself points to kinetics as the lever.
• Estivill '87 closes the length gap toward short probes, demonstrating empirically that an 8-base probe can be washed to discriminate a full match from an end-mismatched duplex — and the '464 patent itself concedes this: "There is no evidence for use of probes shorter than 11-mers in the literature, except for very stable ones [Estivill et al., Nucl. Acids Res. 15: 1415 (1987)]."

3.3 The patentee's answer under Regime A — and it is a good one

The specification pre-emptively distinguishes the very reference a challenger would use:

"It is difficult to obtain a sufficient amount of hybrid with short probes, and to discriminate end mismatch with long probes. There is no evidence for use of probes shorter than 11-mers in the literature, except for very stable ones [Estivill]."

And: "On the other hand, probes longer than 15 bases discriminate end mismatch with difficulty [Wood]."

This is a classic teaching-away posture, and it is reinforced by Smith '83 (you need 16–17-mers for uniqueness in a mammalian genome) and by Wood itself (its Td work spans "11 to >1000 bp" — it does not contemplate 6-mers). The NotI octamer used by Estivill, GCGGCCGC, is 7/8 G+C — near the stability ceiling; the '464 specification expressly characterizes Estivill's success as an artifact of using a "very stable" probe.

Under Regime A, the nonobviousness case is strong: the art taught a minimum of ~11 nt; a POSA would not have expected 6-mers to give usable end-mismatch discrimination; and the highest-stability octamer is the least probative example.

---

4. Combination 2 — Regime B (priority slips): Bains & Smith + Lysov + Khrapko + Estivill + Wallace

This is the combination that would decide the case if the priority gate fails.

4.1 The mapping

4.2 Motivation to combine — articulated under KSR

1. Same field, same problem, same technique. Bains & Smith/Lysov address sequencing, Wallace/Estivill address oligonucleotide hybridization discrimination, Wood addresses probe-pool stringency in complex libraries. All are "familiar elements … according to known methods" (KSR, 550 U.S. 398, 416 (2007)).
2. An express design need. SBH is only tractable if the probe set is finite. With L = probe length, the set size is 4^L. For whole-genome work, 4^8 = 65,536 (workable) versus 4^11 = 4.2 million (not). Bains & Smith and Lysov create the motivation to shorten probes, and Wallace/Estivill supply the discrimination technique that shortening requires. This is a textbook "finite number of identified, predictable solutions" situation: KSR, 550 U.S. at 421; In re Aller, 220 F.2d 454, 456 (CCPA 1955) (routine optimization of a known process parameter); In re Boesch, 617 F.2d 272, 276 (CCPA 1980).
3. Wood '85 is the actual bridge. It teaches that stringency can be made a function of length alone — removing the GC-content obstacle that would otherwise make a heterogeneous short-probe set unmeltable in one uniform wash. A POSA combining Bains/Lysov (SBH) with Wood (length-only stringency) arrives at claim 1's "single condition for all probes" idea directly.
4. Estivill reduces the concept to practice at 8 nt. It is the missing empirical link: a POSA reading Estivill would not need to predict that short probes discriminate end mismatches — Estivill shows it.
5. Predictable result / reasonable expectation of success. Wallace's ~10 °C Td depression per mismatch and the general principle that a shorter duplex has a larger relative destabilization per mismatch make the result mechanistically predictable.

4.3 Where the patentee still has room

• Estivill's probe is atypically stable (7/8 GC) and was used to detect known sites, not to generate unknown sequence. Extending it to the full 4^6–4^8 probe space under one uniform condition is not what Estivill teaches.
• Wood's range begins at 11 nt. Nothing in Wood teaches 6-mers.
• Bains & Smith and Lysov are silent on end-mismatch discrimination — they assume error-free hybridization data. The '464 patent's contribution is the recognition that end mismatch, the least destabilizing mismatch type, is nonetheless discriminable at 6–8 nt.
• The specification's own empirical result is the counter-argument: "it is not surprising that the search for stringent temperature conditions for good discrimination in SOH is unjustified … By washing at lower temperatures, one obtains equal or better discrimination" — i.e., the art's instinct (raise stringency) was wrong, and the inventors found the opposite. That is a candidate unexpected result.

---

5. Claim-specific observations

• Claim 1 is the broadest and most exposed. Once (f) is supplied (Regime B), the only remaining differentiator is the ~6-mer floor and end-mismatch discrimination, and Estivill + Wallace cover that.
• Claim 6 (the "1,250 sample groups / 1,024 groups of 16 (A,T,C,G)N10(A,T,C,G) …" recipe) is a numerical implementation claim of the kind that is highly avoidable and, as a practical matter, largely a recitation of design choices following from the disclosed theory. It is unlikely to survive as a broad monopoly even if valid, but it is also the least likely to be anticipated/rendered obvious, precisely because no reference recites it.
• Claim 7 is framed on an error-tolerance criterion rather than a temperature/salt regime — this is the patent's defensive drafting move against Wallace/Wood/Estivill (which are all about specific conditions). A challenger must show the criterion is inherent in any working SBH method (Pevzner '89's error-tolerant reconstruction is the best counter).
• Claim 8 (step-wise adjacent-probe selection) is the closest to pure algorithm. Bains & Smith/Lysov/Pevzner teach ordering by overlap; claim 8's added "step-wise identify the next adjacent probe" is arguably an obvious extension of the same overlap rule.

---

6. Secondary considerations (Graham v. John Deere, 383 U.S. 1 (1966))

Nexus problem: the commercial-success evidence attaches to the SBH platform as a whole (including U.S. 5,202,231, 5,695,940, 5,972,619, 6,018,041), not to the specific claimed method. Expect a challenger to attack nexus.

---

7. Bottom-line assessment

The single most promising invalidity theory is therefore not a reference combination at all — it is a § 119/§ 112 priority attack. If the 6-mer/7-mer/8-mer discrimination data in Examples 6.3–6.4 was added after 07/175,088 (and the CIP chain makes that likely), the claims lose the 1987/1988 dates, Bains & Smith and Lysov cross the § 102(b) critical date, and the obviousness case becomes strong. Conversely, if priority holds, the art as of April 1987 is missing the SBH-assembly element entirely.

Practical caveat on the whole exercise: the patent expired June 11, 2013, with the term after April 13, 2010 disclaimed. There is no live exclusionary right to invalidate, and I found no 2026 litigation or CAFC activity naming 5,525,464 (consistent with the earlier Litigation summary). Any present-day § 103 determination is retrospective — relevant to defensively citing the patent as art, to its res judicata/collateral role in the Hyseq–Affymetrix history, or to academic analysis, not to an injunction or damages posture.

---

8. Uncertainty flags

1. The referenced "Prior Art section" was never supplied. Everything above is built from the patent's own specification (authoritative here) plus independently searched bibliographic/abstract data. A true prior-art search report could materially change the Regime-A analysis (e.g., if it contains a 1985–1987 reference teaching SBH assembly).
2. Claim text is not verified against the official USPTO copy. The elements I mapped (especially claim 1's "(a)–(f)" and claim 7's error-tolerance framing) come from the third-party mirror discussed in the earlier Claim-set section. Verify against PatentCenter before relying on the element-by-element map.
3. I could not confirm whether the Yugoslav priority documents (P-570/87 and 18617-P 570/87) support the 6-mer/7-mer/8-mer disclosure. That is the pivot of this entire analysis and requires the certified copies referenced in the '464 specification's priority paragraph.
4. *Drmanac et al., Nucl. Acids Res. 14:4691–4692 (1986)* — content not retrieved; it is the inventors' own pre-priority publication and could constitute a § 102(b) statutory bar if it discloses the claimed method. Not relied upon here.
5. No verified claim construction, reexamination certificate, or invalidity holding on the '464 patent itself was located in this session (the earlier Patent-summary section reached the same conclusion regarding reexamination).
6. Reference dates I did not independently confirm: Ikuta '87 page/date (relied on the '464 specification's citations), Poustka '86 and Michiels '87 (relied on the '464 Background), and Pevzner '89 (relied on secondary reference lists).

Primary sources for the verified references: Wallace 1979, PMID 158748; Wood 1985, PMID 3856838; Estivill 1987, PMID 3029714; Lysov 1988, PMID 3250844; Bains & Smith 1988, DOI 10.1016/S0022-5193(88)80247-9.