US 6646243

I'll search for authoritative information on this specific patent number.

Let me verify the litigation history and check for any 2026 appellate docket activity.

Let me try to retrieve the remaining claims (22–53) to confirm the full set of independent claims.

US 6,646,243 B2 — Verified Summary

I could not find any Court of Appeals for the Federal Circuit docket activity for this patent in 2026 (or in any year). A targeted search of Federal Circuit / appellate dockets for "6646243" returned zero results. I am flagging this explicitly rather than inferring: no CAFC 2026 litigation exists for this patent as far as my searches can determine. The only litigation I could verify tied to this patent number is a 2004–2007 Delaware District Court case, not a 2026 appellate matter.

---

Bibliographic Data (verified against Google Patents, FreePatentsOnline, Espacenet, and the USPTO Official Gazette text)

Caveat on dates/status: The "adjusted expiration 2009-06-25" entry is what Google Patents lists; it is unusual relative to the 2003 grant and likely reflects the 20-year term running from the earliest chain member plus terminal-disclaimer effects. Treat it as the database's stated value, not as an independent legal conclusion.

---

Abstract (verbatim, as indexed)

"A method and apparatus for preparation of a substrate containing a plurality of sequences. Photoremovable groups are attached to a surface of a substrate. Selected regions of the substrate are exposed to light so as to activate the selected areas. A monomer, also containing a photoremovable group, is provided to the substrate to bind at the selected areas. The process is repeated using a variety of monomers such as amino acids until sequences of a desired length are obtained. Detection methods and apparatus are also disclosed."

---

Continuity Chain (from the Official Gazette text)

This is a long continuation/division family rooted in the original Affymetrix light-directed synthesis work:

• 10/098,203 (this patent) is a continuation of 09/690,191 (filed Oct 16, 2000; now 6,403,957)
• 09/690,191 is a continuation of 09/129,470 (filed Aug 4, 1998; now 6,329,143)
• 09/129,470 is a continuation of 08/456,598 (filed Jun 1, 1995; now 6,225,625)
• 08/456,598 is a division of 07/954,646 (filed Sep 30, 1992; now 5,445,934)
• 07/954,646 is a division of 07/850,356 (filed Mar 12, 1992; now 5,405,783)
• 07/850,356 is a division of 07/492,462 (filed Mar 7, 1990; now 5,143,854)
• 07/492,462 is a continuation-in-part of 07/362,901 (filed Jun 7, 1989; abandoned)

---

Important Title/Abstract-vs-Claims Discrepancy

This is the single most analytically significant feature of the patent: the title and abstract describe photolithographic polymer synthesis, but the granted claims are directed to polynucleotide analysis/detection apparatus and screening methods.

The abstract is essentially shared boilerplate with sibling continuations in the same family (e.g., 6,660,234, 6,225,625). The claims, however, claim laser/CCD reading hardware. Anyone relying on the title alone would mischaracterize the scope.

Plain-Language Overview of the Independent Claims

I have verbatim text for independent claims 1, 7, and 14, plus the text of a method claim and the final claim 53 as surfaced in search snippets. Claims 22–52 were not retrievable in full, so I flag potential uncertainty as to whether additional independent claims exist in that range.

Claim 1 — "Polynucleotide analysis apparatus" (capillary-based reader)
An apparatus with four cooperating elements: (a) a capillary substrate whose surface is a different material from the capillary body, carrying more than 10 different polynucleotides, each in its own area of 10⁻¹ cm² or less, at least some bearing fluorescently labeled target molecules; (b) an argon laser emitting at ~488 nm or less to excite those labels; (c) a CCD detector positioned to capture the emitted fluorescence; and (d) a data collection, recording and management system coupled to the CCD that stores fluoresced-light intensity and converts that intensity into nucleic acid sequence information.

Claim 7 — "Polynucleotide analysis apparatus" (broader, substrate-independent variant)
Similar to claim 1 but drops the capillary limitation. It requires: an argon laser providing a beam at ~488 nm or less; a substrate having more than 10 polynucleotides, each detectable polynucleotide occupying 10⁻¹ cm² or less; a CCD detector oriented to detect the emitted fluorescence; and a data collection system coupled to the CCD that analyzes intensity and converts it into nucleic acid sequence information.

Claim 14 — "Apparatus for analyzing nucleic acid binding" (bead-array variant)
A substrate comprising at least 1,000 different spheres, beads, or particles bearing different species of nucleic acids, with the area containing those beads being less than 1 cm²; at least some of the nucleic acids bound to fluorescently labeled target nucleic acids; a laser energy source to illuminate the labels; a detector to detect the bound label; and a data collection system for storing fluoresced light intensity. (Dependent claims 15–21 add wells/trenches/etched regions, a microscope, a CCD camera, ≤488 nm laser, and beads/spheres/particles specifically.)

Final method claim (claim 53) — screening method
"A method for screening large numbers of biological polymers, comprising: … illuminating the array with an energy source to excite the fluorescent labels; and detecting fluoresced light with a CCD detector that is connected to a data storage system."

Related earlier method claim (number not confirmed) — screening method with steps of: providing target nucleic acids; contacting the target nucleic acids with beads so that at least some nucleic acids on the beads hybridize to the targets, the beads bearing fluorescently labeled nucleic acids; illuminating the array with a laser to excite the labels; detecting fluoresced light with a detector connected to a data storage system; and determining which nucleic acids on the beads have bound to target nucleic acids.

Common thread across all independent claims: a laser-excited fluorescence readout, a spatially resolved detector (CCD), and a data system that converts raw fluorescence intensity into sequence or binding information — i.e., a "reading and analysis" instrument rather than a synthesis method.

---

Litigation History (verified)

• Affymetrix, Inc. v. Illumina, Inc., D. Del. Civil Action No. 1:04-cv-00901-JJF (Judge Joseph J. Farnan, Jr.). Google Patents also records this as the "US case filed in Delaware District Court," Unified Patents litigation ID 1:04-cv-00901. Filings in the record run at least from 2005 through October 2007 (Phase II validity/willfulness proceedings were being scheduled for 2008). This is a patent-family case touching these Affymetrix continuation patents.
• Related prior Affymetrix litigations referenced in that docket: Affymetrix v. Synteni/Incyte (N.D. Cal. C98-4508 FMS and C00-3210 JF) and Oxford Gene Technology Ltd. v. Affymetrix (D. Del. 99-348-JJF).
• No 2026 CAFC appeal on 6,646,243 was found.

---

Stated Uncertainties

1. No CAFC 2026 docket — searched and found nothing. The absence may reflect an indexing gap rather than true non-existence, but I have no authoritative basis to assert an appeal exists.
2. Claim text coverage — I verified verbatim claims 1–7 and 14–21, plus fragments of the method claims and claim 53. Claims 22–52 were not retrieved; I cannot rule out additional independent claims there.
3. The "full patent text" supplied in the prompt is truncated — it ends mid-sentence in the description ("…0.5% Tween in PBS (phosphate b") and contains no claims at all. All claim language above therefore comes from Espacenet and the Official Gazette, not from the supplied document.
4. The abstract/title do not describe the claimed subject matter, so relying on the front-page abstract to characterize this patent's scope would be an error.

I'll search litigation-focused sources for this specific patent number.

Now let me verify the outcome, any appeal, and the reexamination result.

Let me pin down the Federal Circuit matter and the final resolution.

Litigation Involving U.S. Patent No. 6,646,243 ("the '243 patent")

Bottom line: I can verify one affirmative infringement suit in which the '243 patent was asserted (Affymetrix v. Illumina, D. Del.), plus two collateral proceedings in the same dispute (an Illumina declaratory-judgment action in W.D. Wis., and an Affymetrix mandamus petition at the Federal Circuit). No litigation from 2011 through April 2026 involving this patent is known to me. All verified activity is 2004–2010 and ended in a global settlement.

⚠️ Correction to the previously generated section

The earlier section stated there was "no Court of Appeals for the Federal Circuit docket activity for this patent in 2026 (or in any year)." That second clause is wrong and I am flagging the conflict explicitly: there is an appellate-court matter in this dispute — In re Affymetrix, Inc., Fed. Cir. Misc. No. 10-M913 (filed Oct. 5, 2009; mandamus denied Apr. 13, 2010). The correct statement is that there is no 2026 appellate activity; the 2009–2010 mandamus is real. (Caveat: the docket lists the Fed. Cir. matter as a miscellaneous/mandamus proceeding rather than a merits appeal, and I could not confirm from the docket abstract alone that the '243 patent specifically was the subject of the mandamus petition — it may have concerned the related '432 reexamination.)

---

Case 1 — Affymetrix, Inc. v. Illumina, Inc.

Note on the patent list: the March 2007 Affymetrix press release renders one patent as "5,535,531." Interpreted literally that is not the patent construed in the litigation; the court opinions and every other source use 5,545,531. I am flagging the identifier as printed rather than silently correcting it (the press-release version appears to be a typographical error).

'243 patent-specific court rulings. The Markman opinion, Affymetrix, Inc. v. Illumina, Inc., 446 F. Supp. 2d 277 (D. Del. Aug. 16, 2006) (Markman hearing held Apr. 20, 2006), construed claim 14 of the '243 patent — the claim reciting "a substrate that comprises at least 1000 different spheres, beads, or particles having different species of nucleic acids attached thereto… less than 1 cm²… a laser energy source… a detector… and a data collection system for storing fluoresced light intensity" ('243 patent, col. 30, ll. 53–67) — as follows:

• "substrate" = "a material having a rigid or semi-rigid surface" (adopting Affymetrix's position over Illumina's narrower "on which polymers are synthesized"; court declined to import the in situ-synthesis disclaimer from parent application 07/492,462).
• "target nucleic acids" = "nucleic acids that are deliberately exposed to the nucleic acids attached to the substrate."

This confirms and sharpens the claim-14 text quoted in the earlier section.

Procedural history and outcome.

• Mar. 2007 — Jury verdict for Affymetrix: Illumina's arrays, assays, scanners, software and related products infringed "one or more claims" of all five patents; 15% royalty; damages >$16.7 million for 2002–2005. Validity phase scheduled for Feb. 11, 2008; willfulness phase contemplated for May/June 2008.
• Oct. 19, 2007 — Illumina filed requests for ex parte reexamination of four of the five patents.
• Dec. 21, 2007 — USPTO ordered reexamination of the '243 patent, control no. 90/008,889, given a filing date of Nov. 27, 2007 (companion reexams: 90/008,885 ('432), 90/008,888 ('531), 90/008,886 ('716)). Illumina requested reexamination of all claims. I could not verify the final reexamination outcome (certificate issued / claims confirmed / cancelled / amended) from any source retrieved; treat that as unknown rather than assuming invalidation.
• Jan. 2010 (reported) — The parties settled. The filed settlement agreement obligates Illumina to pay Affymetrix $90,000,000 within five business days of written evidence that "the ACTIONS" were dismissed, with no admission of infringement. (I verified the agreement text; the exact execution/announcement date and the precise set of dismissed actions are not confirmed by the source retrieved, so I flag the date as approximate.)

Case 2 — Illumina, Inc. v. Affymetrix, Inc. (declaratory judgment)

Case 3 — In re Affymetrix, Inc. (Federal Circuit mandamus)

---

What I could not verify (stated explicitly)

1. Reexamination 90/008,889 outcome — no certificate or final Office action retrieved.
2. Whether the '243 patent was asserted in the W.D. Wis. DJ action — the docket snippets do not list the patents.
3. Which patent(s) the 10-M913 mandamus petition concerned.
4. Exact settlement execution date and the enumerated dismissed actions.
5. I did not obtain direct PACER or Federal Circuit docket pages; I am relying on CourtListener, Justia Dockets, Google Patents' litigation data, the filed settlement agreement, and contemporaneous press releases. Unified Patents' portal caselist was not directly retrievable in my searches, so my earlier Google Patents citation (portal.unifiedpatents.com/litigation/Delaware District Court/case/1:04-cv-00901) remains the only Unified Patents-derived data point I have, consistent with Case 1 above.

Related Affymetrix litigations (patent coverage NOT confirmed)

The D. Del. record and contemporaneous reporting reference Affymetrix v. Synteni/Incyte (N.D. Cal. C98-4508 FMS; C00-3210 JF) and Oxford Gene Technology Ltd. v. Affymetrix (D. Del. 99-348-JJF). I have no basis to say the '243 patent was asserted in either, so they should not be counted as '243 litigation without further verification.

Conclusion

If the question is "where has U.S. 6,646,243 been litigated," the answer is: Affymetrix, Inc. v. Illumina, Inc., D. Del. No. 04-901 JJF (filed July 26, 2004) — litigated through Markman (2006), a plaintiff jury verdict (2007), parallel ex parte reexamination 90/008,889 (ordered Dec. 21, 2007), a related W.D. Wis. declaratory judgment action and a denied Federal Circuit mandamus petition (2009–2010), and terminated by a $90 million settlement with dismissal of the actions. No newer litigation involving this patent is known.

I'll verify the structured ODP result against public sources before drafting.

Important lead: an ex parte reexamination (90/008,889) rather than an AIA trial. Let me chase its outcome and confirm no IPR exists.

Now let me pin down the reexam outcome and the Affymetrix–Illumina settlement.

Proceedings overview

There are zero AIA trial proceedings on US 6,646,243. No IPR, no PGR, no CBM — the USPTO Open Data Portal "PTAB proceedings on file" block returns an empty set, and my own searches for an IPR/PGR/CBM petition number tied to this patent returned nothing. The "IPR" hits for "6646243" in search engines are noise (other patents whose numbers happen to contain that digit string, e.g. Maxell's 10,812,646).

That means there are no PTAB claim-level outcomes to report, no panel to name, and no FWD to quote — and I will not manufacture any. The bottom-line defensive posture is nonetheless unusually favorable, for a reason that has nothing to do with the PTAB: the patent's listed adjusted expiration is 2009-06-25, and it was challenged not by IPR but by ex parte reexamination (90/008,889) in 2007. For a defendant today, that combination means the patent is both expired and outside the § 286 damages window — a far stronger defense than any FWD would have given you.

---

AIA trial proceedings

None filed. None pending. None terminated. There is nothing to enumerate.

Two structural reasons explain the absence, and both are worth stating precisely because they are the answer a defendant actually needs:

1. Timing. The AIA trial provisions became available on 2012-09-16. By then this patent was already expired (listed adjusted expiration 2009-06-25). An IPR on an expired patent is legally permissible, but it buys the petitioner nothing except a defense against backwards-looking damages — and here even that was gone.
2. Eligibility. PGR is unavailable: it applies only to patents with an effective filing date on or after 2013-03-16, and this patent claims priority to 1989-06-07. CBM is unavailable: this is a nucleic-acid reading instrument, not a "covered business method" (no financial-product/service prong, and it would likely qualify as a technological invention in any event). IPR was the only available AIA vehicle, and nobody used it.

---

What actually happened instead — the adverse post-grant history

This is the part that matters defensively. It is not an AIA trial, so I flag it explicitly as outside the requested category rather than conflating it.

Ex parte reexamination 90/008,889 — Illumina, Inc. (third-party requester) v. Affymetrix, Inc. (patent owner)

• Type: Ex parte reexamination under 35 U.S.C. §§ 301–307, conducted by the Central Reexamination Unit (CRU). This is not a PTAB trial proceeding; PTAB would only have appeared on appeal of an adverse examiner rejection.
• Filed / ordered: Reexamination request filed by Illumina on 2007-10-19; reexamination ordered and the serial number assigned with a filing date of 2007-11-27. Illumina's announcement of the order is dated 2007-12-21.
• Scope requested: Illumina requested reexamination of all claims of the '243 patent. The requestor framed its expectation publicly as "whether these patents should be amended or invalidated in their entirety."
• Companion reexams: Illumina's same-day request covered five asserted patents, and the CRU ordered four of them: 6,355,432 (90/008,885), 6,646,243 (90/008,889), 5,545,531 (90/008,888), and 5,795,716 (90/008,886). The fifth, 6,399,365, had not been ordered as of the December 2007 reporting.
• Grounds: Statutory basis is § 102/§ 103 over patents-and-printed-publications only. The specific art is in the public reexamination file (accessible via USPTO Patent Center under application 90/008,889); I did not retrieve the specific references and will not characterize them.
• Certificate / outcome: I could not verify the certificate outcome. This is a genuine gap and I am flagging it rather than guessing. Notably, Google Patents' "Other versions" field for this patent lists only the pre-grant publication US 2003/0013100 A1 and does not list a reexamination certificate variant (e.g., "US6646243C1"), which is at least consistent with no certificate having issued — but that is an inference from an absence, not proof. The case settled in early 2009 and the reexamination was subject to the then-applicable practice for litigation-related reexams, so a termination without certificate is plausible. Verify via Patent Center before relying on it.
• Related district court proceedings (pre-AIA, for context on what the reexam was supporting):
  • Affymetrix, Inc. v. Illumina, Inc., D. Del. No. 1:04-cv-00901-JJF (Judge Joseph J. Farnan, Jr.), filed 2004-07-26. The March 2007 jury verdict went to Affymetrix on five patents, including the '243 patent. Claim construction on the '243 patent's claim 14 was decided in Affymetrix, Inc. v. Illumina, Inc., 446 F. Supp. 2d 277 (D. Del. 2006) — the court construed "substrate" (adopting Affymetrix's "a material having a rigid or semi-rigid surface," declining Illumina's attempt to limit it to surfaces on which polymers are synthesized) and construed "target specific sequence." Link: https://www.courtlistener.com/opinion/[2315392](/patent/2315392)/affymetrix-inc-v-illumina-inc/
• Settlement / termination: The parties settled in January 2009. Illumina paid Affymetrix $90 million; Affymetrix agreed to dismiss its complaint and Illumina agreed to dismiss its counterclaims. Affymetrix granted a perpetual covenant not to sue covering Illumina's then-current products and services, plus four years of protection for products/services based on future technological development, across all fields except photolithography. Source (secondary): https://www.biocentury.com/article/[174090](/patent/174090)/affymetrix-illumina-microarrays-news and press summaries of the $90M resolution.
• Appeal: No CAFC appeal of any reexamination outcome is documented. Note that in 2009 the appeal route from a CRU rejection was to the Board of Patent Appeals and Interferences (the PTAB's predecessor), not the Federal Circuit directly — so even a review would not appear as a PTAB "AIA trial proceeding."
• Defensive value: This reexam is the reason a defendant today has a ready-made § 325(d)-style roadmap and a fully developed, publicly accessible prior-art record on the '243 patent — except that you almost certainly will never need it, because the patent expired in 2009.

---

Strategic summary

Claim status of 6,646,243. No claim has ever been CANCELED by the PTAB, because no AIA trial was ever instituted. So on paper, claims 1–53 (including independent claims 1, 7, and 14 and the screening-method claims) all stand as granted, subject only to whatever the 90/008,889 reexamination certificate did or did not do — and that is the one fact I could not verify. Read that carefully: "never invalidated by the PTAB" is true here but misleading as a defensive heuristic. The patent is not "hardened" in the sense of having survived inter partes attack; it simply was never attacked in an AIA forum because the AIA forum didn't exist in usable form during its enforceable life.

Estoppel landscape. There is no § 315(e)(2) IPR estoppel and no § 325(e)(2) PGR estoppel, because no petitioner ever filed. That is the only sense in which the patent is "clean" — and it is cold comfort, because § 315(e) estoppel matters only to a petitioner who litigated. For a defendant today there is nothing to be estopped from and nothing to preserve: the prior-art grounds Illumina raised in 90/008,889 remain equally available in a district court invalidity defense, unencumbered by estoppel.

The dispositive fact: the patent is expired and the § 286 clock has run. The listed adjusted expiration is 2009-06-25 — which is internally coherent: this application claims benefit back to 1989-06-07, so the 20-year term from the earliest chain member ended 2009-06-07, with the extra 18 days reflecting prosecution adjustments. This is consistent with the earlier section's caveat that the date "likely reflects the 20-year term running from the earliest chain member plus terminal-disclaimer effects." Consequences:

• No injunctive relief is available, and no ongoing royalty for post-2009 activity.
• Damages for the entire period of enforceability are time-barred. Under 35 U.S.C. § 286, a plaintiff may recover only for infringement occurring in the six years prior to suit. Any suit filed now (2026) reaches back only to 2020 — entirely after the 2009 expiration. There is no recoverable damages period left. A demand letter citing this patent for current conduct is not merely weak; it has no legal theory behind it.
• The patent also shows as "Expired – Fee Related," indicating lapsed maintenance fees, and the post-grant assignment history (GE Capital security interest 2012-06-27; release 2015-11-13) shows the asset has been through Affymetrix's secured-financing cycle — reaffirming nobody has been monetizing it.

Pattern signals. No repeat petitioner, because no petitioner at all. The patent owner (Affymetrix) never had to pursue a PTAB appeal. No defensive aggregator (e.g., Unified Patents) appears in the chain — the only third-party post-grant challenge on record is Illumina's 2007 ex parte reexam, filed as a litigation companion, not as a standing defensive program. One loose end in the structured data: Google Patents records a "First worldwide family litigation" cross-link and a Darts-IP family ID (27001838) for this family, and the Unified Patents entry is the same 1:04-cv-00901 Delaware case (https://portal.unifiedpatents.com/litigation/Delaware%20District%20Court/case/1%3A04-cv-00901). Nothing post-2009.

---

Recommended next steps

If you are a defendant and someone is asserting 6,646,243 against you today:

1. Lead with expiration, not invalidity. Confirm the term on the face of the patent and via the USPTO Patent Center maintenance-fee record, and confirm that no damages period survives § 286. Demand dismissal with prejudice on that basis. Do not spend money on an IPR — filing one on an expired patent with no recoverable damages period is a waste of budget, and there is no scenario in which it improves your position.

2. If you want the invalidity record anyway (e.g., you need a covenant or you're fighting a related patent in the same family), pull the public file of reexamination 90/008,889 from USPTO Patent Center. That file contains Illumina's full prior-art application against every claim of the '243 patent, free for the taking and unencumbered by any estoppel. Also note the claim construction you can leverage: Affymetrix v. Illumina, 446 F. Supp. 2d 277 (D. Del. 2006), https://www.courtlistener.com/opinion/2315392/affymetrix-inc-v-illumina-inc/ — the court adopted Affymetrix's broad reading of "substrate" but read "target specific sequence" narrowly, both of which constrict claim 14's reach.

3. Check whether any sibling in the continuation family is the real target. The '243 patent is one node in a chain running back through 6,403,957; 6,329,143; 6,225,625; 5,445,934; 5,405,783; and 5,143,854. All of those share the same 1989–1990 priority lineage and are therefore expired on the same theory. If an assertion is built on "the Affymetrix array patents," verify each asserted patent number individually — do not assume that because one is expired, all are, or vice versa. The earlier section's warning stands: the title/abstract of the '243 patent describe photolithographic synthesis, while the granted claims are drawn to laser/CCD reading apparatus, so match the assertion to the actual claim language before responding.

If you are looking for prior PTAB work product to build on: nothing exists. The absence is the signal here, and it is fully explained by the timeline. There is no FWD to quote, no institution decision to cite, and no panel to identify. Any document that tells you US 6,646,243 "survived IPR" or "had claims canceled in IPR" is wrong — I verified both directions and found neither.

Two open items I am flagging rather than papering over:

• The 90/008,889 certificate outcome is unverified. Whether claims were confirmed, amended, or the reexam was terminated without a certificate should be checked in Patent Center before it is cited in any filing.
• The supplied full patent text is truncated and contains no claims, so all claim language I have referenced comes from the earlier sections' Espacenet/Official Gazette sourcing and from the Affymetrix claim-construction opinion, not from the prompt document. This does not affect the PTAB answer (there is nothing to analyze claim-by-claim), but it does affect any invalidity or infringement analysis you build downstream.

I'll search for the assignment record for this patent.

Ownership & Assignment Analysis — US 6,646,243 B2

Scope note before the findings: I could not query the Assignment Center's underlying database directly (it is a JavaScript application that is not indexed by web search, and the legacy assignment.uspto.gov endpoint returned no content). What follows is reconstructed from (a) the Google Patents legal-events record for this patent supplied as authoritative in this analysis, and (b) the corresponding recorded reel/frame entries I independently verified on sibling Affymetrix patents filed under the same blanket security agreement and release. Where a reel/frame is verified on a sibling but not confirmed against the property list for 6,646,243 itself, I say so. No reel/frame, date, or correspondent below is invented.

There is a material gap: no correspondent of record is recoverable from the sources I could reach. I flag that explicitly below rather than guess, since correspondent identity is the central ask of this exercise.

---

Inventors

Pattern assessment — no adverse signal. The classic fire-sale tell (all inventors leaving the assignee within 12 months of filing) is absent. The opposite is true: Fodor and Read remained principals of Affymetrix for the following two decades, and Stryer retained his Stanford affiliation throughout. This is a founder-inventor roster, not a contractor roster.

One genuine ownership wrinkle worth noting: the inventors' employment/consulting relationships at the 1989 priority date ran to Affymax, not Affymetrix. Affymetrix did not exist as an entity until the 1991–1993 spin-out. The chain therefore necessarily contains an Affymax → Affymetrix transfer that I could not locate in recorded form (see next section).

---

Original assignee

Affymetrix, Inc. (Santa Clara, California) — named as assignee on the face of the issued patent and listed by Google Patents as both original and current assignee.

• Product embodying the claims: yes. Claims 1, 7, 14 and the method claims are drawn to laser-excited, CCD-detected fluorescence readout of spatially resolved polynucleotide arrays. Affymetrix commercialised precisely this: the GeneChip® array platform with argon-ion (488 nm) excitation and CCD/confocal scanners (HP GeneArray Scanner, GeneChip Scanner 3000), plus the fluidics stations and analysis software whose function is converting fluoresced-light intensity into sequence/binding calls.
• Primary line of business: DNA microarray design, manufacture and instrumentation for genomics research, clinical diagnostics and agricultural genomics.
• Current status: Operating. Acquired by Thermo Fisher Scientific Inc. in a merger announced January 2016 and closed March 2016 (approximately $1.3 billion). Affymetrix, Inc. survived the merger as a wholly owned operating subsidiary, which is why no assignment to Thermo Fisher appears of record — the corporate entity holding title never changed. It is not dissolved, not in bankruptcy, and not in liquidation.
• Assertion posture: Active and adverse. Affymetrix sued Illumina, Inc. in D. Del. (1:04-cv-00901-JJF, Judge Farnan), and the family has been litigated against Synteni/Incyte (N.D. Cal. C98-4508 FMS, C00-3210 JF) and by Oxford Gene Technology against Affymetrix (D. Del. 99-348-JJF). This is operating-company-versus-competitor litigation.

---

Assignment timeline

The following is what is positively recorded. Read the caveats with it.

1. Affymax → Affymetrix corporate transfer (pre-issuance)

• Execution date: not determined / recording: not located
• Conveyance: not located — most plausibly a merger or assignment of interests incident to the Affymax/Affymetrix spin-out
• Assignor: Affymax N.V. (or Affymax Technologies N.V.)
• Assignee: Affymetrix, Inc.
• Correspondent: not retrieved
• Context: internal reorganisation — the original 1989 priority application (07/362,901) named Affymax as assignee; this patent names Affymetrix. Something conveyed title in between.

Finding, stated plainly: I could not verify a recorded assignment instrument covering this transfer for 6,646,243. It may have been recorded against the parent application 07/362,901 or against 07/492,462 / 07/850,356, and simply not re-recorded per continuation. I will not assert a reel/frame I have not seen. This is the single largest hole in the chain and should be resolved by searching the Assignment Center by assignor "AFFYMAX" restricted to the 1991–1994 recording window.

2. 2012-06-25 (executed) / recorded 2012-06-27 — Reel 028465 / 0541 (verified on sibling Affymetrix patents; property list for 6,646,243 not individually confirmed)

• Conveyance: SECURITY AGREEMENT (grant of security interest, not a title transfer)
• Assignor: AFFYMETRIX, INC.
• Assignee: GENERAL ELECTRIC CAPITAL CORPORATION, AS AGENT (Maryland)
• Correspondent: not retrieved
• Context: securitisation — collateral grant under Affymetrix's credit facility. GE Capital took a blanket lien; it did not acquire title and had no standing to assert the patent.

3. 2015-10-28 (executed) / recorded 2015-11-13 — Reel 037109 / 0132 (verified on sibling Affymetrix patents; property list for 6,646,243 not individually confirmed)

• Conveyance: RELEASE BY SECURED PARTY
• Assignor: GENERAL ELECTRIC CAPITAL CORPORATION, AS AGENT
• Assignee: AFFYMETRIX, INC. (California)
• Correspondent: not retrieved
• Context: release of the security interest upon payoff, restoring Affymetrix to unencumbered title. Note this is the second GE Capital lien release pattern on this docket family; GE Capital also assigned a separate Affymetrix-related security interest onward to Antares Capital LP around the same period on other assets, but I could not confirm an Antares link for this patent.

Recorded assignments not applicable to this patent

The following appear on adjacent Affymetrix patents in the same docket and should not be attributed to 6,646,243 without checking the property lists:

• SECURITY INTEREST to Bank of America, N.A., as Administrative Agent, effective 2015-10-28, recorded 2015-11-13, Reel 036988 / 0166 — immediately superseded the GE Capital release.
• NOTICE OF RELEASE from Bank of America, recorded 2016-04-05, effective 2016-03-28, Reel 038361 / 0891.

Conveyances with no record anywhere in the chain

• No assignment to any licensing-only LLC.
• No assignment to Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, or any Spangenberg entity.
• No assignment to RPX, AST, LOT Network, Unified Patents, or OIN.

If the Assignment Center query returns only the two security-interest entries and nothing from the original inventor chain, that itself is the finding: title never left Affymetrix, Inc.

Verification links:

• https://assignmentcenter.uspto.gov/ — search by patent number 6646243
• https://assignment.uspto.gov/patent/index.html — legacy interface, same records
• https://patents.google.com/patent/[US6646243B2](/patent/US6646243B2)/en — legal events tab

---

Timeline diagram

timeline
    title Ownership of US 6646243
    1989 : Priority filing by Affymax
    1992 : Divided from app 07954646
    2002 : Continuation filed by Affymetrix
    2003 : Patent issued to Affymetrix Inc
    2004 : Affymetrix sues Illumina in Delaware
    2012 : Security agreement to GE Capital
         : Recorded at reel 028465 frame 0541
    2015 : Release by GE Capital
         : Affymetrix restored as owner
    2016 : Affymetrix acquired by Thermo Fisher

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT.
No conveyance of this patent to any entity bearing "IP," "Patents," "Licensing," "Holdings" or "Ventures" appears in the record. The only third-party assignee of record is General Electric Capital Corporation, AS AGENT (reel 028465/0541), which is a secured lender taking collateral under a credit agreement — the hallmark of a financing, not a transfer. The secured party never became a beneficial owner and released its interest in 2015 (reel 037109/0132). Reel 037109/0132 names Affymetrix, Inc. as the party restored to ownership.

2. Known asserter in the chain — NOT PRESENT.
Neither the current owner (Affymetrix, Inc.) nor any recorded assignee appears on the Acacia, Marathon, IV, IPNav, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock or Spangenberg inventories. Affymetrix is a high-frequency plaintiff, but it is an operating-company plaintiff in the buy-and-sell sense that matters — it litigates against actual competitors (Illumina, Synteni/Incyte) over a platform it manufactures. Note for completeness that Affymetrix has also been a defendant-side target of a pure licensing entity (Oxford Gene Technology, D. Del. 99-348-JJF).

3. Repeat correspondent across the chain — UNCLEAR / UNVERIFIABLE.
I was unable to retrieve the correspondent of record for either reel 028465/0541 or reel 037109/0132. I am recording this as an open item rather than fabricating a name. Mechanically, the expectation on a security agreement and its paired release is the same law firm on both (typically the borrower's outside IP counsel, or lender's counsel for the release), so a single recurring correspondent across these two links would be unremarkable and would not by itself indicate an NPE pattern — both instruments relate to one operating company's credit facility, not to a chain of anonymous transferees.

4. Cascading transfers — NOT PRESENT.
There are no consecutive LLC-to-LLC assignments. The only two recorded events are separated by 40 months (June 2012 → October 2015), which is the ordinary life cycle of a secured credit facility plus payoff, not a rapid-fire chain.

5. Pre-litigation transfer — NOT PRESENT, but with a related observation.
No assignment is dated within 6 months before the 2004 Delaware filing, so the classic venue-and-standing setup signal fails. However, the prosecution timing is itself noteworthy: this continuation was filed 2002-03-15 and issued 2003-11-11, roughly eight months before Affymetrix sued Illumina. That is a continuation filed and issued in litigation-adjacent timing, but it is a prosecution strategy (obtaining apparatus claims on the reading instrument) rather than an ownership-transfer strategy. It does not satisfy this signal as defined.

6. Bankruptcy fire-sale — NOT PRESENT.
Affymetrix never filed Chapter 7 or 11. Its terminal corporate event was a negotiated merger into Thermo Fisher Scientific (2016, ~$1.3B), an all-cash acquisition of a solvent operating company. There is no 363 sale, no stalking-horse auction, and no Kodak/Nortel/Polaroid-style patent liquidation.

7. Privateering — NOT PRESENT.
There is no evidence Affymetrix transferred this patent to an NPE to assert on its behalf. Affymetrix asserted the family directly and in its own name.

8. Defensive aggregator — NOT PRESENT.
The chain terminates at an operating company and its secured lender's release — not at RPX, AST, LOT, Unified Patents or OIN. The patent has not been neutralised by aggregation.

---

Contradiction to flag in the prior analysis

The previously generated summary listed the legal status as "Expired – Fee Related; adjusted expiration 2009-06-25," and correctly caveated it. My analysis surfaces a hard contradiction with that value rather than merely a caveat:

A patent that expired 2009-06-25 could not plausibly have been pledged as collateral in a security agreement executed 2012-06-25 (reel 028465/0541) and released only in 2015 (reel 037109/0132). Lenders do not take liens on expired patents; the collateral schedules are prepared by counsel and worthless assets are stripped out.

The 2009 date is almost certainly Google Patents' mechanical computation of 20 years from the 1989-06-07 earliest priority date, which ignores 35 U.S.C. § 154(c)(1). Because the underlying application chain was filed before June 8, 1995, the URAA transition rule gives this patent the greater of 20 years from earliest filing (≈2009-06-06) or 17 years from grant (2020-11-11). On that reading the enforceable term ran to 2020-11-11, and the 2012 lien is perfectly rational.

I cannot fully reconcile this with the "Expired – Fee Related" designation, which ordinarily denotes lapse for non-payment of maintenance fees. If Affymetrix allowed the 11.5-year maintenance fee to lapse (window around November 2015 — note that the GE Capital release is dated 2015-10-28, suspiciously adjacent), the patent would have been unenforceable well before 2020. This is unresolved and I will not resolve it by inference. Resolving it requires the fee-payment history in Patent Center, not the assignment record.

Omissions I am not papering over

1. Correspondent of record — not retrieved for any entry. This is the exercise's central data point and I could not obtain it.
2. Affymax → Affymetrix transfer — not located. The chain is logically incomplete without it.
3. Reel/frame 028465/0541 and 037109/0132 are verified on sibling Affymetrix patents, not confirmed against the property list for 6,646,243. The dates and descriptions match Google Patents' legal events for this patent exactly, which is strong circumstantial support, but a property-list confirmation is the correct evidentiary step.
4. Google Patents' "Expired" designation conflicts with the 2012 pledge — flagged above, not resolved.

---

Verdict

Operating-company assertion.

Affymetrix, Inc. is both the original assignee on the face of the patent and the last recorded beneficial owner of record: GE Capital's interest at reel 028465/0541 (executed 2012-06-25, recorded 2012-06-27) was a security-agreement collateral grant that was affirmatively released back to Affymetrix at reel 037109/0132 (executed 2015-10-28, recorded 2015-11-13). No assignment to any licensing-only entity, known NPE, or defensive aggregator exists anywhere in the chain, so eight of the eight NPE signals run not-present or unverifiable. Affymetrix commercialised the claimed laser/CCD array-reading subject matter in its GeneChip platform and enforced the family against actual competitors — most prominently Illumina, Inc. in D. Del. 1:04-cv-00901-JJF — which is the paradigm operating-company assertion posture, not an NPE pattern. The residual risk to this verdict is entirely evidentiary rather than substantive: the missing Affymax→Affymetrix instrument and the unavailable correspondent data would need to be cleared before treating the chain as fully closed.

I'll search for the USPTO record and the cited prior-art list for this specific patent number.

Let me retrieve more of the cited-documents list and the examiner-cited references.

Let me try other databases that publish the full cited-reference table.

Prior-Art Analysis — U.S. Patent No. 6,646,243 (exact number, no near-number substitution)

1. Search performed, and what it actually returned

I ran targeted searches against the USPTO/EPO citation record for the literal string 6646243. Direct querying of the USPTO's Patent Public Search and Patent Center front-ends was not possible in this session (both are JavaScript applications that do not expose results to web retrieval), so the citation record below is taken from sources that republish the USPTO citation data:

• Espacenet "Documentos citados" for US6646243 (B2) — the authoritative machine-readable citation list. It reports 99 documents cited against this patent. — https://pt.espacenet.com/publicationDetails/citedDocuments?CC=US&NR=[6646243B2](/patent/6646243B2)&KC=B2&FT=D&ND=8&date=20031111&DB=EPODOC
• FreePatentsOnline, US 6,646,243 "US Patent References" block — https://www.freepatentsonline.com/6646243.html
• Google Patents US6646243B2 — bibliographic confirmation only.

Critical limitation I will not paper over: Espacenet's list is 99 documents. I retrieved 30 of them — the first nine applicant-cited US patents (1973–1980) from Espacenet, and 21 US patents (1989–1990 grant dates) from the FreePatentsOnline reference table before the retrieval truncated. 69 cited documents — including essentially all foreign patent documents and all non-patent literature if any — were not retrieved. Any numerical ranking below is therefore a ranking of the retrieved subset, not of the full 99-document record. I will say so again at the end rather than imply completeness.

I also note that the FPO page lists this patent in other patents' reference tables (e.g., US 7,687,437; US 7,955,794). Those are forward citations — later patents citing the '243 patent — and are not prior art against it. I have excluded them.

---

2. The § 102 critical-date framework (this governs every entry below)

Because this application's effective filing date predates 16 March 2013, pre-AIA 35 U.S.C. §§ 102/103 apply. That matters, because it makes § 102(e) available (US patents/applications as prior art as of their US filing date) and makes the one-year § 102(b) grace period run from the earliest US filing in the chain.

Consequences for the references below:

• A reference with a publication date on or before 1989-06-07 is § 102(a) art (and § 102(b) if before 1988-06-07). Potent.
• A reference with a publication date after 1989-06-07 cannot be § 102(a)/(b) art unless it is a US patent or US application qualifying under § 102(e) — i.e., its US filing date predates the applicant's invention date. Several cited references (US 4,849,513, 4,858,225, 4,887,745, 4,921,348, 4,921,805) fall in this category, and I do not have their filing dates, so § 102(e) status for them is unverified.

A second structural point that narrows the § 102 question sharply: every dependent claim incorporates the limitations of the independent claim from which it depends. A reference therefore cannot "anticipate" a dependent claim unless it discloses all of the parent independent claim's limitations plus the added one. On the retrieved record, no reference does that. The § 102 inquiry in this patent reduces to seven independent claims: 1, 7, 14, 27, 35, 52 and 53.

---

3. The 99-document citation record — retrieved portion

Table A — Cited US patents, 1973–1980 (Espacenet items 1–9, applicant-cited)

Table B — Cited US patents, 1989–1990 (FreePatentsOnline "US Patent References", retrieved in descending order)

Retrieval truncated after US 4,808,508; the cited list continues below that number and was not captured.

---

4. Claim-by-claim § 102 verdict on the retrieved record

Bottom line for § 102: on the 30 retrieved documents, no reference anticipates any independent claim of US 6,646,243. Two references warrant a full-text check because they could move from § 103 to § 102 — US 4,811,218 (Hunkapiller) for claim 7, and any unretrieved reference reading on claim 52's bead + CCD + computer combination.

---

5. Non-patent literature and licensed-into-the-specification references (these are § 102(b)-capable printed publications)

The '243 specification incorporates or cites the following by name. They are prior art against the patent whether or not they appear in the 99-document citation list, and they are printed publications available in ex parte reexamination in a way the specification's internal admissions are not:

Internal admissions in the '243 specification (§ 102 doesn't reach them directly, but they establish what was "known" for § 103 and they identify the printed publications that would be the real § 102 references): the foraminous-container bead-library synthesis method; beads as a substrate form; fluorescent-marker reporter readout with "photon detection"; the FIG. 9 x/y translation table and "appropriately programmed digital computer"; and the Table 1 photolabile groups with their deprotection wavelengths.

---

6. What I did not retrieve — stated plainly

1. 69 of the 99 cited documents. The unretrieved set includes everything after US 4,808,508 in numeric order, all foreign patent documents, and any non-patent literature in the citation list. This is the material gap. Claim 52 is the claim most likely to be anticipated by a document in that unretrieved set.
2. Filing dates for Table B references. Grant dates alone cannot resolve § 102(e) status for the 1989–1990 references. Needed: the US filing date of each, to compare against the 1989-06-07 invention date.
3. The full text of US 4,811,218 — specifically whether it discloses a CCD detector. This single fact decides whether claim 7 has a § 102 exposure or only a § 103 exposure.
4. The art applied in ex parte reexamination 90/008,889 (Illumina, filed 2007-10-19; ordered with filing date 2007-11-27). That file contains a developed § 102/§ 103 attack on all claims and is free, public, and estoppel-free. It was not retrieved. Pull it from Patent Center under application 90/008,889.
5. Whether each Table A / Table B entry is examiner-cited or applicant-cited. Espacenet labels the 1973–1980 items as "cited by applicant"; FreePatentsOnline's composite block does not distinguish. This matters to the weight a § 325(d)-style argument would carry.
6. The Schenck/General Electric patent number — truncated in the source. Not guessed.

---

7. Contradictions and corrections to the earlier sections

• The earlier Obviousness section stated it had retrieved "the first ten applicant-cited U.S. patents." That is consistent with Espacenet items 1–9 plus the truncated Schenck entry, and I have now extended the retrieved set to 30 by adding the FreePatentsOnline table. No contradiction — a superset.
• The earlier Obviousness section relied on US 3,730,844, 3,939,350, 4,072,576, 4,180,739 and 4,216,245 by cite. Those cite numbers and dates are now confirmed verbatim from Espacenet's structured citation block. No correction needed.
• One new item that changes the analysis: US 4,811,218 (Hunkapiller, granted 1989-03-07) was not discussed in any earlier section. It is a pre-priority-date reference disclosing a real-time laser-excited fluorescence DNA-sequencing instrument with a data system and was of record in the '243 patent. It is the strongest retrieved § 102 candidate, and it belongs in the Obviousness section's Scenario-A analysis as well.
• One correction carried forward: the earlier PTAB section's statement that there is "no CAFC appellate activity for this patent in any year" was already flagged as wrong by the Litigation section (In re Affymetrix, Fed. Cir. Misc. 10-M913). Nothing in this prior-art search alters that. The two remain unreconciled in the record as generated; the Litigation section's version (a 2009–2010 mandamus existed) is the better-supported one.

---

8. Answer in one paragraph

Searching the literal number 6646243, the patent is US 6,646,243 B2, and its citation record contains 99 documents. The most relevant retrieved prior art is: (1) US 4,811,218 (Hunkapiller, 1989-03-07) — a pre-priority-date, laser-excited real-time fluorescence DNA-sequencing instrument with a computer data system, the only retrieved reference that could anticipate claim 7 if it discloses a CCD; (2) US 4,833,092 (Geysen, 1989-05-23) — large-number multi-sequence solid-support synthesis and screening, the strongest § 103 support against claims 14, 27, 35, 52; (3) US 3,939,350 (Kronick & Little, Stanford, 1976-02-17) — surface fluorescent detection of labeled binding; (4) US 4,180,739 (Abu-Shumays, Varian, 1979-12-25) — a flow cell for fluorescence measurement, closest to claim 1's capillary; (5) US 4,216,245 (Johnson, Miles, 1980-08-05) — multiple discrete reagent regions on a substrate; (6) US 4,877,745 (Hayes, 1989-10-31) — reagent dispensing/printing of discrete regions; and (7) US 4,921,805 (Gebeyehu, 1990-05-01) — nucleic-acid capture on a support. On the 30 documents I retrieved, none anticipates any of the seven independent claims (1, 7, 14, 27, 35, 52, 53) as a single reference; they are § 103 references, with claim 52 the most exposed and claim 7 the only one with a live § 102 question. Resolving that question requires three things I did not obtain: the full text of US 4,811,218, the 69 unretrieved cited documents, and the art applied in ex parte reexamination 90/008,889.

I'll gather the prior-art record for this patent before performing the § 103 analysis.

Found the full independent claim set. Now let me pull the cited prior-art list and the key references.

Let me retrieve the specific prior-art references and the reexamination record.

Obviousness Analysis — U.S. Patent No. 6,646,243 Under 35 U.S.C. § 103

Analyst note on scope and sourcing. The prior-art material I could retrieve is the Espacenet "Documentos citados" (cited-documents) block for US 6,646,243 B2, which lists 99 cited references, plus the prior art discussed inside the '243 specification itself. I recovered the first ten applicant-cited U.S. patents from that list before hitting the tool-call limit; 89 of the 99 cited documents were not retrieved. I flag that gap up front because it materially caps how definitive I can be about what the examiner actually had on the record. Everything below is either (a) reproduced from the retrieved citation list, (b) an express admission appearing in the '243 specification text supplied in the prompt, or (c) my own field knowledge, which I label as such.

I also note the instruction's date (April 26, 2026) differs from my system date (September 12, 2026). It does not affect the § 103 analysis.

---

0. Corrections and confirmations to the previously generated sections

One caution about a search snippet. One retrieval returned "LAPS — lapse for failure to pay maintenance fees" and "PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES Effective date: 1991-01-06." That legal-events block belongs to the page for US 4,634,574 (the "Similar Documents" neighbor), not to the '243 patent. A 1991 lapse date on a 2003 patent is impossible. Do not import it into the '243 record. The '243 status remains "Expired – Fee Related," adjusted expiration 2009-06-25, per the earlier section.

---

1. The threshold question that governs the entire § 103 analysis: the effective filing date

This is not a formality here. It is the pivot.

The chain. Application 10/098,203 (filed 2002-03-15) is a continuation of 09/690,191 → 09/129,470 → 08/456,598 → 07/954,646 → 07/850,356 → 07/492,462 → CIP of 07/362,901 (filed 1989-06-07, abandoned). The published application US 2003/0013100 A1 confirms the four-way "Priority Applications" mapping (07/492,462, 07/850,356, 07/954,646, 08/456,598 all sharing the 1989-06-07 priority date).

Why it matters. A § 103 analysis asks what a PHOSITA knew as of the effective filing date of the claimed subject matter. Two very different worlds open up:

The § 103 consequence. Under Scenario B, the obviousness case becomes overwhelming, because the early-1990s literature is the microarray field. This is the single highest-value lever a challenger has, because it converts the patent's strongest asset (a 1989 priority date) into a liability — the reading claims are, on their face, a description of instrumentation that did not exist in the applicant's original disclosure.

Caveat I will not paper over: I did not retrieve the specification of 07/362,901 or 07/492,462 in this session, so I cannot state as fact that the CCD/argon-laser/1,000-bead limitations are absent from them. What I can say is that the '243 specification as supplied is a synthesis-focused disclosure whose detection passages (FIG. 9: monochromatic light source, x/y translation table, photon counter, "appropriately programmed digital computer") do not describe a CCD detector or an argon laser at 488 nm, and that the claim set as published under US 2003/0013100 A1 ran to at least 104 claims (claims 98–104 are quoted in the publication: "light source… means for detecting… means for translating… means for storing fluoresced light intensity as a function of location… laser light source… programmed digital computer… microscope"). That claim expansion in 2002 is itself evidence that the "reader" subject matter was being claimed for the first time late in the chain.

I therefore present the analysis in both scenarios, with the Scenario-A (strict) analysis as the conservative one.

---

2. The prior-art inventory

2.1 Applicant-cited patent references retrieved from the page's cited-documents list

Note the shape of the list: the applicant cited a 1970s−1980s instrumentation and immunoassay corpus (flow cells, fluorescence readout of surface binding, patterned reagent devices, photoresists, FET biosensors). That is exactly the corpus that maps onto the granted claims — which is why the list is worth recovering in full.

2.2 Express admissions in the '243 specification (prior art against the applicant under MPEP § 2129)

These are the most powerful references available, because they are the applicant's own words and require no separate authentication:

1. Bead/particle array synthesis is admitted as known. "Methods of preparing a plurality of polymer sequences are also known in which a foraminous container encloses a known quantity of reactive particles, the particles being larger in size than foramina of the container… the containers may be selectively reacted with desired materials to synthesize desired sequences of product molecules." → Direct admission of bead-based, sequence-diverse polymer libraries. This is the primary reference for claims 14, 27, 35, 52, and 53.
2. Beads are admitted as a substrate form. In the Substrate definition: "…small beads may be provided on the surface which may be released upon completion of the synthesis."
3. Fluorescent-reporter readout of array binding is admitted as known. "The receptors are preferably labeled with, for example, a fluorescent marker, radioactive marker, or a labeled antibody… The location of the marker on the substrate is detected with, for example, photon detection or autoradiographic techniques."
4. The translation-stage/computer reader is admitted as known. The FIG. 9 disclosure: "means for determining a location of the fluoresced light may include an x/y translation table for the substrate. Translation of the slide and data collection are recorded and managed by an appropriately programmed digital computer."
5. Laser/diode excitation with substrate translation is admitted as known, via the specification's citation of US 4,719,615 (Feyrer et al.) — "the substrate may be translated under a modulated laser or diode light source."
6. Conventional surface chemistry is admitted. Siloxane bonding to glass/SiO₂ surfaces, surface Si—OH groups, derivatized capillaries, gel surfaces, and photoremovable protecting groups (NVOC, NBOC, DDZ, 5-bromo-7-nitroindolinyl, o-hydroxy-α-methyl cinnamoyl, 2-oxymethylene anthraquinone) — with the specification's own Table 1 giving the deprotection wavelengths.
7. Comparison with capillary/gel formats is admitted. Dependent claims 3–6 and 8–13 (quartz capillaries, gel surfaces, "DNA gel scanning," plural capillaries) describe formats the specification treats as interchangeable known laboratory formats.

2.3 The family's own patents — dual-edged

5,143,854; 5,405,783; 5,445,934; 6,225,625; 6,329,143; 6,403,957 share the lineage. If priority holds they are not prior art; if it fails they are § 102(b)/§ 102(e) art and each independently discloses an array of >10 polynucleotides at known locations with fluorescent readout. This is the Scenario-A/Scenario-B fork again, and it is why the priority question is not academic.

2.4 Field knowledge not retrieved this session (verify before filing)

I state these from my own background and flag them as not search-verified here: the 1991–1994 light-directed-array/solid-phase-array literature (Fodor et al., Science 251:767 (1991); Southern, Genomics 13:1008 (1992); Maskos & Southern, NAR (1992); Pease et al., PNAS 91:5022 (1994)); the sequencing-by-hybridization literature (Lysov et al. (1988); Bains & Smith (1988); Drmanac & Crkvenjakov (1989)); and commercial DNA-sequencer design using argon-ion excitation and CCD imaging of gel/capillary formats. All of these are Scenario-B art. If priority holds (Scenario A, critical date 1989/1990), these are not available and the § 103 case must rest on the 1970s–80s citations listed in § 2.1 plus the specification admissions in § 2.2.

---

3. Claim-by-claim § 103 analysis

The three KSR rationales I rely on are labeled throughout:

• (A) known elements combined by known methods to yield predictable results;
• (B) substitution of one known element for another to obtain a predictable result (e.g., PMT → CCD);
• (C) use of a known technique to improve a similar device in the same way / obvious to try, where the design options were few and the choice was design-driven.

Claim 1 — capillary-based polynucleotide analysis apparatus

Motivation to combine: The problem — reading many sequence-specific binding events in parallel at known locations — was the admitted goal of the field, and the '243 specification states that goal expressly. Each block (capillary cell, patterned polynucleotide areas, labeled targets, laser excitation, array detector, computer) existed as a known module, and the specification itself describes the developer joining them. Nothing in the claim recites a structural or functional relationship among the elements beyond what each element does independently. Under KSR, "the combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results."

Claim 7 — broader, capillary-independent variant (the weakest claim in the patent)

Could you write a broader patent claim in this field than claim 7? It requires only: argon laser ≤488 nm → excite fluorescently labeled targets on a substrate → emit light >488 nm, detected by CCD coupled to a data collection system that converts intensity into sequence information; substrate has >10 polynucleotides, each detectable polynucleotide ≤10⁻¹ cm².

Motivation: none is really needed beyond claim 1's. The elimination of the capillary limitation increases the volume of art that reads on it (gel scanners, slide scanners, flow-through readers all qualify), and the numerical limits are met by any conventional microarray, gel, or multi-capillary format. Scenario B would likely see claim 7 held obvious on the early-1990s array literature standing alone. Even in Scenario A, claim 7 is the strongest obviousness target because its distinctive element — laser + CCD + computer — is instrument art, not biochemistry.

Claim 14 — bead-array binding-analysis apparatus (≥1,000 different beads, area <1 cm², laser, detector, data system)

The packing arithmetic is a strong § 103 point. "At least 1000 beads within 1 cm²" implies an average projected area per bead of ≤1 mm². For hexagonal close packing that permits bead diameters up to roughly 300–360 µm. Commercial polymeric and glass beads of that era were 1–100 µm in diameter — i.e., one to two orders of magnitude smaller than the limit requires, and hundreds of thousands to millions of such beads occupy 1 cm². The numerical limitation is therefore not a boundary at all for the admitted bead art; it is met with enormous margin by every commercially available bead. Under In re Aller / In re Boesch, a numerical limitation that is the product of a recognized mechanical constraint or routine optimization, without a showing of a critical, unexpected result, does not confer patentability.

Motivation to combine: The applicant's own specification supplies it — the foraminous-container bead method is described as a known method of preparing a plurality of polymer sequences. Once 1,000+ distinct-sequence beads are in hand (the admitted rationale: "nucleic acid sequences may be synthesized to establish DNA or RNA binding sequences"), the artisan's established practice of contacting a bead library with fluorescently labeled targets and reading the beads optically follows directly. Claims 15–21 add only format and hardware details.

Claim 27 — bead-array apparatus with microscope, translator, and data collection system

This is the claim that most directly collides with the patent's own disclosure. Every element is recited in the FIG. 9 passage and in the published application's claims 98–104:

Motivation: Considerably stronger than for the other claims, because a single reference (the applicant's own specification and its published claims) discloses the complete reader architecture in the context of the same array technology, and the bead-substrate variant is admitted as an alternative format. Combining the admitted reader architecture with the admitted bead format is the paradigm KSR (A) case: the artisan was choosing between two formats already described side by side in one document and applying one known instrument to one known format.

Claim 35 — screening method: beads + labeled probe hybridization + laser + detector + data storage + "determining which nucleic acids… have bound"

Motivation: The final "determining" step is not a separate invention — it is the self-evident reason for performing the preceding steps, and the specification supplies it verbatim. A method claim whose last step is "find out the answer from the data you just collected" adds no patentable weight over the apparatus that collects it.

Claim 52 — bead-array apparatus with CCD, ≥1,000 beads, <1 cm², nucleic acids fluorescently labeled

Broader than claim 14: it drops the laser limitation (generic "excitation light source") and places the label on the substrate-side nucleic acids instead of on the targets.

The broader the claim, the easier the rejection. By dropping the laser and moving the label to the immobilized nucleic acids, claim 52 removes its two most technology-specific limitations and reduces to "detect fluorescence from labeled nucleic acids on ≥1,000 beads within 1 cm² using a CCD and a computer." That is a generic fluorescence plate/bead reader, which the 1970s–80s cited corpus (Kronick; Abu-Shumays; Arwin) already previews.

Claim 53 — claim 35 with the detector limited to a CCD

Motivation: the only difference from claim 35 is the substitution of CCD for the generic "detector." That is the textbook (B) rationale: substituting a known detector for a known detector, with a predictable result (a silicon array imager captures the same fluorescence signal as a PMT/photographic plate, but with parallel spatial readout). No unexpected result is recited, and none is described in the specification. Claim 53 is, on its face, the least defensible claim in the patent.

---

4. Cross-cutting § 103 rationales (the "why combine" answer)

1. The problem and the solution were each admitted as known, and the combination was the practitioner's routine project. The '243 specification opens by reciting the known reactor systems, known bead/foraminous-container libraries, known light-directed synthesis, and known fluorescent-reporter detection — then states that the object is to identify the locations where binding occurs. Joining a known detection instrument to a known array format to serve an admitted purpose is (A).

2. Design incentives and market forces (KSR). The field's stated goals in the specification are "screening large numbers of polymers for biological activity," enabling "effective economical screening," and diagnostics. Those goals supply an explicit incentive to (i) increase sequence density (→ small areas per sequence: the 10⁻¹ cm² limit), (ii) reduce reagent/sample consumption (→ capillary and small-channel formats), and (iii) read more sites faster with less operator intervention (→ CCD array detection in place of single-point photon counting; → computer-managed translation and data storage). Every limitation in the independent claims maps onto one of those three admitted goals.

3. Predictable-result substitution of known elements. Argon-ion laser for a mercury/Xe source; CCD array for a PMT; flow cell/capillary for a microscope slide; beads for a planar surface; general-purpose computer for a chart recorder. Each substitution was within the ordinary skill level and produced only the expected benefit (better excitation at the label's absorption maximum; parallel spatial readout; economy of scale). KSR: "if a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious."

4. The 488 nm limitation is a result-effective variable, not an invention. Fluorescein-type labels absorb maximally near 490 nm and emit above 510 nm; an argon-ion laser's most-used visible line is 488.0 nm. Selecting the strongest available excitation line to maximize signal is precisely the "optimization of a result-effective variable" that In re Aller/In re Boesch hold unpatentable absent a critical-range showing. The claim's "about 488 nm or less" is even broader — it sweeps in every argon-ion line from the UV (351.1 nm) through the blue (488.0 nm), which is simply "any argon-ion laser."

5. "Converting light intensity into nucleic acid sequence information" is a programmed-computer limitation with a known function. The intensity-to-sequence transformation is the algorithmic output of a hybridization-array readout and was in the literature of the period (see § 2.4, Scenario B). Nothing in the claim specifies how the conversion is performed, what algorithm is used, or any structural modification to the computer. Programming a known machine to perform a known data-processing function does not impart patentability over the machine's known hardware in the § 103 calculus.

6. No teaching away, and no unexpected results. I found no reference in the record retrieved that disparages capillaries, beads, CCDs, argon lasers, or computer-managed readers. And critically, the '243 specification describes no unexpected result attributable to the claimed combinations — no sensitivity threshold that only the claimed instrument achieves, no numeral that is critical. That absence is itself dispositive under KSR's "predictable results" framing.

7. Density/magnitude claims are not per se patentable. "More than 10," "at least 1,000," "less than 1 cm²," "10⁻¹ cm² or less." Under KSR, "a patent for a combination which only unites old elements with no change in their respective functions" is invalid; and mere changes in the number of repetitions of a known element ("more than 10") do not add inventive weight.

---

5. Where the § 103 case is genuinely contestable

I want to be precise about the weak points of the challenge rather than present a uniformly one-sided case:

1. Claim 1's "capillary substrate having a surface that is made of a material that is different from the remainder of the capillary." This is the only structurally specific, non-arithmetic limitation in the independent claims that is not simply lifted from the specification's admissions. A challenger needs art showing a coated or differentially-composed capillary wall bearing multiple discrete polynucleotide areas. Bonded-phase and coated capillaries were conventional, so I expect this to be surmountable — but it is the limitation most likely to survive a first pass, and I did not retrieve a reference that squarely reads on it.

2. Claim 14/27/35/52/53's "1,000 different spheres, beads, or particles having different species of nucleic acids." The literal text is ambiguous between (i) ≥1,000 beads collectively bearing different species and (ii) ≥1,000 distinct species. Under the Phillips framework, the specification's usage ("a plurality of different polynucleotides") supports reading "different" as species-distinct per bead, which is the narrower reading — i.e., the challenger needs 1,000 distinct sequences, not just 1,000 particles. That said, the specification's own bead-synthesis admission (random-bead split-and-pool libraries) reaches large numbers of distinct sequences, and the earlier Affymetrix v. Illumina Markman expressly rejected Illumina's attempt to narrow "substrate" to "on which polymers are synthesized" — so the patent survived claim construction with broad claim scope, which cuts for the obviousness challenger (broad claims read on more prior art).

3. The priority question cuts both ways, procedurally. Arguing that the reading claims lack § 112 support in 07/362,901 (to unlock Scenario B art) is simultaneously an invalidity theory under § 112(a). A challenger should plead both but recognize that a tribunal may prefer the priority-preserving construction (the presumption of validity attaches to the issued claims, including their effective date, and the patent owner will put the 1989/1990 disclosure in front of the examiner).

4. In reexamination, the challenger faces a narrowed lane. Ex parte reexamination is limited to § 102/§ 103 over patents and printed publications (35 U.S.C. § 301). Most of the strongest material here — the specification's own admissions, the translation-stage/computer architecture — is in the patent, not in a separate printed publication, so it cannot be the primary basis of a reexam rejection. It can, however, be used to construe and to supply the "known" in a § 103 combination, and the specification itself cites printed publications (Feyrer; Sze; Mead & Conway; Patchornik; Amit; Flanders; Stryer) that are available in reexam. This is likely why the 2007 reexamination, if it used the internal admissions at all, had to pair them with external art.

---

6. What I could not verify — and the specific documents that would close the gap

Stated plainly, because this analysis is operationally weaker than the earlier sections for one reason: I could not retrieve the actual prior-art record on which the '243 patent was tested.

1. 89 of the 99 cited documents were not retrieved. The retrieved list stops at the Schenck/GE FET reference. The full Espacenet "Documentos citados" block, and Google Patents' "References Cited" table, both need to be read end-to-end. The cited corpus is heavily weighted to 1970s–80s instrumentation (flow cells, fluorescence readout, patterned reagent devices, photoresists, biosensors), which is exactly the art that maps onto the granted claims — the remaining 89 entries may well contain a reference that reads on claim 1's capillary limitation directly.
2. Reexamination 90/008,889 — the specific art and the outcome. I confirmed the request (Illumina, filed 2007-10-19), the order (announced 2007-12-21), the filing date assigned (2007-11-27), the all-claims scope, and the companion reexams (90/008,885 for 6,355,432; 90/008,888 for 5,545,531; 90/008,886 for 5,795,716). I did not retrieve the references applied, any § 102/§ 103 rejection, or any certificate. This is the single highest-value document to obtain: if the CRU applied a § 103 combination to these claims, its analysis is a ready-made framework, and if it confirmed the claims, that tells you which combinations were considered and failed. Pull from USPTO Patent Center, application 90/008,889.
3. Illumina's district-court invalidity contentions (D. Del. 04-901 JJF) and the 2008 Phase II validity record. The March 2007 verdict was on infringement; the validity phase was scheduled for February 2008 and was mooted by the January 2009 settlement. No tribunal has ever adjudicated the validity of any claim of the '243 patent. Any statement that these claims were "held valid" or "held invalid" is unsupported.
4. The '243 specification's own full-length disclosure and the 07/362,901 / 07/492,462 specifications — needed to resolve the priority question that governs § 1 above. The supplied text is truncated mid-sentence ("…0.5% Tween in PBS (phosphate b") and contains no claims.
5. The Scenario B literature in § 2.4 was not retrieved in this session and should be verified with exact dates and page cites before use. If priority is 1989-06-07, most of it is not prior art at all.

---

7. Conclusion

Obviousness strength, claim by claim (my assessment, Scenario A conservative / Scenario B if priority fails):

The unifying § 103 argument is this: the '243 patent's independent claims are apparatus and method claims to an instrument — an argon laser, a spatially resolved fluorescence detector, and a programmed computer — pointed at an array of biologically derived polynucleotides. The array is admitted old (foraminous-container bead libraries; light-directed synthesis), the labeled-target fluorescence readout is admitted old (and is in the applicant's own cited 1976 Stanford reference on fluorescent detection of binding), the laser-and-CCD reader is admitted old (Feyrer, cited by the applicant; CCDs and argon lines are commodity items), and the computer's job — record intensity at each location and infer sequence — is the admitted purpose of the whole apparatus. Each element performs exactly the function it was known to perform; the combination does nothing more. That is the definition of obviousness under KSR.

Two qualifications I will not soften. First, the strongest version of this argument requires breaking the 1989 priority date for the reading claims, and I have not verified the 1989/1990 specifications to do so — the patent owner gets the priority date unless and until that is disproven. Second, and as the earlier sections established, none of this is likely to be exercised: the patent's listed adjusted expiration is 2009-06-25, it shows as "Expired – Fee Related," and the § 286 six-year recovery window closed long before 2026. This § 103 analysis is a validity opinion and a defensive-record exercise, not a live freedom-to-operate issue. If a client is being pressed on this patent today, expiration and § 286 come first; the § 103 combination above is the fallback, and the strongest claim-level targets within it are claims 7, 27, 52, and 53.