US 6355432

I'll research US patent 6,355,432 across the specified sources.

Let me verify the claim set and check for any Federal Circuit / 2026 activity.

Research Summary — US 6,355,432 (literal ID: "6355432")

Confidence note up front: I found no evidence of any 2026 activity for this patent, in the CAFC docket or elsewhere. Every litigation hit I could retrieve is from 2004–2011. This is consistent with the patent's recorded legal status. If your instruction to check "CAFC 2026 dockets" was meant to surface a live appeal, I could not confirm one — see the Litigation/status section below for what I could and could not verify.

1. Bibliographic data

Continuation chain (per the printed Official Gazette entry, Internet Archive/USPTO):
09/585,659 (Jun 2, 2000) ← cont. of 09/362,089 (Jul 28, 1999) ← div. of 09/056,927 (Apr 8, 1998, now 6,197,506) ← cont. of 08/670,118 (Jun 25, 1996, now 5,800,992) ← div. of 08/168,904 (Dec 15, 1993, abandoned) ← cont. of 07/624,114 (Dec 6, 1990, abandoned) ← CIP of 07/492,462 (Mar 7, 1990, now 5,143,854) ← CIP of 07/362,901 (Jun 7, 1989, abandoned).

Two data discrepancies worth flagging literally, as found:

• Google Patents' timeline shows the "Priority claimed from US07/492,462" event dated 1990‑03‑07, while separately listing the prior-art/priority date as 1989‑06‑07. These are the two different dates in the chain above, not an error I corrected.
• A Unified Patents portal snippet renders the priority date as "1989-06-06," one day off from the 1989‑06‑07 shown by Google Patents and by the OG/Indiana University record for application 07/362,901. I did not reconcile these; treat 06‑07 as the better-supported value.

2. Abstract (verbatim, from the front page)

"The present invention provides methods and apparatus for sequencing, fingerprinting and mapping biological macromolecules, typically biological polymers. The methods make use of a plurality of sequence specific recognition reagents which can also be used for classification of biological samples, and to characterize their sources."

Note the mismatch that is common in this family: the title and claims are bead/nucleic-acid-specific, while the abstract is the broad, umbrella abstract shared across the Affymetrix "VLSIPS" continuation family. The European sibling EP 1 231 282 A3 (same inventors, same family, filed 06.12.1991, priority US 624114) carries a more claim-matched abstract: targets are identified by exposing a sample to a collection of substrate-bound probes where the substrates are beads, each bead bearing a different probe and additionally tagged (coded) so the bound probe's identity is revealed when the tag is read.

3. Plain-language overview of the independent claim

Based on the Espacenet claim text and the D. Del. Markman opinion (which quotes col. 82), claim 1 is the sole independent claim; claims 2–22 all depend, directly or indirectly, from it.

Claim 1 (independent) — a collection of beads:
A collection of beads where
(a) the beads have binding polymers of different target‑specific sequence attached to them; and
(b) the beads are coded with an encoding system so that you can tell which target‑specific polymer sequence is on which bead.

How the court read it (Affymetrix v. Illumina, 446 F. Supp. 2d 277 (D. Del. Aug. 16, 2006), Judge Farnan):

• "Said beads being coded with an encoding system" = "said beads having a property associated with each bead (separate from the binding polymer) that can be used to distinguish one bead from another." The court expressly excluded using the binding polymer itself as the code, reasoning that this would be "circular identification" (using a binding polymer to identify itself).
• "Target specific sequence" = "a known polymer sequence that has affinity for another sequence" — i.e., it need not have affinity for the sequence ultimately being determined (Illumina's narrower reading was rejected).

Selected dependent claims (plain language), from the Espacenet claim listing:

• 2–11: The bead-bound binding polymer is an oligonucleotide of a given length drawn from all possible sequences of that length; coverage thresholds of at least ~20% (claim 3), ~70% (claim 4); lengths of ≥5 (5), 8 (6), 9 (7), ≥10 (8) nucleotides; and absolute numbers of distinct bead-bound sequences of at least 10,000 (9), 100,000 (10), 1,000,000 (11).
• 12–13: The polymer is a polynucleotide or polypeptide (12); or the binding polymer is a protein with an enzyme or antibody binding site (13).
• 14–16, 19: Surface/linker chemistry — glass surface with attachment through amines (14) or hydroxyl groups (15); oligodeoxyribonucleotides attached via phosphate immediately to hydroxyls on an organic hydrophilic coating (16); a TEFLON copolymer graft surface with a hydrophilic organic layer terminating in hydroxyl sites (19).
• 17: The encoding system is magnetic, shape-based, color-based, or a combination.
• 18, 22: The bead-bound polymer is an oligonucleotide probe of a given length (18); claim 22 likewise recites a given oligonucleotide length but its text was truncated in the source I retrieved ("…wherein the given length of the oligonucleotide attached to a plurality of…") — I cannot reliably state its full scope.
• 20–21: At least one nucleotide analogue (20); the beads are reusable (21).

4. Litigation and post-grant history (what the record shows)

• Affymetrix, Inc. v. Illumina, Inc., D. Del. C.A. No. 04‑901‑JJF (filed July 26, 2004). The '432 was one of six asserted patents (also 5,545,531; 5,795,716; 6,399,365; 6,607,887; 6,646,243).
• Markman ruling: August 16, 2006, 446 F. Supp. 2d 277 — constructions quoted above.
• Invalidity: Illumina moved for summary judgment of invalidity of the asserted claims of the '432 (D.I. 279, filed July 14, 2006). The proposed order in the docket material states claims 2, 5, 8, and 9 of the '432 are "adjudged to be invalid." Affymetrix also moved for JMOL of infringement of claims 2 and 9 (D.I. 418, March 11, 2007), seeking $24,237,000 lost profits plus a ≥12% royalty.
• Reexamination: The USPTO ordered reexamination of the '432 and of 6,646,243 ("Nucleic acid reading and analysis system"), as reported by GenomeWeb.
• Federal Circuit: The only CAFC docket item I found tied to this dispute is In re Affymetrix, No. 10‑M913 (petition filed Oct. 5, 2009, with opinions Oct. 13, 2009 and April 13, 2010) — a mandamus petition, with Illumina as respondent. Separately, Illumina, Inc. v. Affymetrix, Inc., Nos. 2011‑1145/1146 (Fed. Cir., nonprecedential order Aug. 18, 2011) affirmed the W.D. Wis. consolidated cases 09‑CV‑0277/0665. I could not verify from these searches a merits CAFC opinion squarely deciding the validity of the '432 claims.

5. Explicit uncertainties / limits of this search

1. No 2026 CAFC docket activity found for 6,355,432. Searches combining the number with "CAFC 2026" returned zero results. Given the recorded anticipated expiration of June 7, 2009 and "Expired – Fee Related" status, a 2026 appeal on this patent would be highly anomalous. If you have a specific 2026 docket number, I'd need it to verify — I will not infer one.
2. I did not run a direct PAIR/PatentCenter fetch; the bibliographic data above comes from Google Patents, the printed OG entry (Indiana University virtual library mirror), Espacenet, and the D. Del. record. Claim 22's full text is unverified, and I cannot rule out a second independent claim from the sources retrieved (the Espacenet list and the court opinions both treat claim 1 as the base claim).
3. The "1989‑06‑06" vs "1989‑06‑07" priority-date discrepancy is reported as found, not resolved.

I'll search the litigation databases for this specific patent number, then cross-check against court dockets.

Let me verify against court records and check for any additional assertions of this specific patent.

Litigation Involving US 6,355,432 B1 — Case List

Bottom line: exactly one (1) patent-specific infringement case is documented for US 6,355,432. It is Affymetrix, Inc. v. Illumina, Inc., D. Del. (Judge Farnan), filed 2004‑07‑26, settled January 2008. No other district court case, ITC investigation, or appeal asserting the '432 surfaced in this search. The one litigation link on the patent's own Google Patents page points to that same Delaware case.

A hard warning about noise, because it matters here. My searches returned a large volume of hits for other patents that happen to end in the digits "432" — e.g., USAA's remote‑check‑deposit '432 (asserted against Wells Fargo, PNC and Mitek, IPR2021‑01071/01074), ZS Pharma's '432 asserted against an ANDA filer (Oct. 5, 2022 complaint), and various unrelated dockets. None of those are US 6,355,432 and I have excluded all of them. I also deliberately did not carry in the WO search‑report citation to "US6355431 B1," which is a different patent number. Per the operating rule, I have not merged similar numbers into this list.

1. The one case on point

Procedural milestones (all confirmed this session):

• 2006‑08‑16 — Markman order. Memorandum Opinion and Order, 446 F. Supp. 2d 277 (D. Del. 2006) (CourtListener opinion 2315392; Casetext mirror). The court construed, inter alia:
  • "said beads being coded with an encoding system" (claim 1, col. 82, ll. 53–55) = "said beads having a property associated with each bead (separate from the binding polymer) that can be used to distinguish one bead from another." The court reasoned: "Using the binding polymers as the encoding system then, would amount to using a binding polymer to identify itself. This circular identification cannot be what the inventors intended."
  • "target specific sequence" = "a known polymer sequence that has affinity for another sequence" — rejecting Illumina's narrower proposal.
  • The opinion confirms the '432's sole independent claim is claim 1 and quotes it verbatim.
  • Affymetrix publicly announced the order on 2006‑08‑17 (BioSpace).
• 2007‑03‑05 — Phase 1 trial commenced (rescheduled from 2006‑10‑16 at the parties' request; multi-phase structure).
• 2007‑03‑13 — Jury verdict for Affymetrix. Infringement of "one or more claims" of all five patents; damages of more than $16.7 million for 2002–2005; 15% royalty (GenomeWeb).
• Validity phase was scheduled for 2008‑02‑11, with willfulness to follow in May/June 2008. Those phases never ran.
• 2008‑01 — global settlement. Dismissed with prejudice in the U.S., German and UK actions; Affymetrix granted a perpetual covenant not to sue. (Per the earlier sections in this file: Illumina's one‑time payment was $90 million; the covenant ran to Illumina, its affiliates and its customers.)

Case‑tracker corroboration: the Google Patents page for US6355432B1 carries exactly one litigation entry — a U.S. case filed in Delaware District Court, 1:04‑cv‑00901, sourced to the Unified Patents litigation dataset: https://portal.unifiedpatents.com/litigation/Delaware%20District%20Court/case/1%3A04-cv-00901. There is no second litigation entry on that page.

2. Adjacent proceedings — flagged so they are not mistaken for '432 litigation

These are real, but they are not district court cases asserting the '432. I list them only so the record is complete and so no one conflates them with the list in §1.

1. Ex parte reexamination 90/008,885 — administrative, NOT litigation. Illumina's requests filed 2007‑10‑19; USPTO ordering date reported by Illumina as 2007‑11‑27, announced 2007‑12‑21. Companion reexams: 90/008,886 ('716), 90/008,888 ('531), 90/008,889 ('243). Sources: patentdocs.org; BioSpace/I‑BO; GenomeWeb. Outcome still unverified — I could not confirm whether the certificate confirmed, amended, or cancelled claims.

   • Note a discrepancy I am reporting, not resolving: patentdocs.org says the four requests were filed 2007‑10‑19; Illumina's own release states the reexams "were assigned the same filing date of November 27, 2007." Treat the exact reexam filing date as unconfirmed.

2. Illumina, Inc. v. Affymetrix, Inc., W.D. Wis. 3:09‑cv‑00277 (consolidated with 09‑cv‑0665) — asserted Illumina's patents (7,510,841 / 7,612,020 per the earlier sections), not the '432. Fed. Cir. Nos. 2011‑1145/1146 affirmed dismissal, Aug. 18, 2011.

3. In re Affymetrix, Fed. Cir. No. 10‑M913 — a mandamus petition filed Oct. 5, 2009, with Illumina as respondent. I could not verify that the '432 was the subject patent; treat this as related‑dispute context only.

4. The 2007 U.S./German/UK filings referenced in the earlier litigation summary — I could not locate a separate U.S. case number for any of them asserting the '432. Treat as unverified.

3. What I could NOT find (stated as a negative, not as proof of absence)

• No second U.S. district court case naming the '432 as an asserted patent.
• No ITC Section 337 investigation involving the '432.
• No PTAB proceeding of any kind (0 IPRs, 0 PGRs, 0 CBMs) — consistent with the "PTAB challenges" section already in this file. Note the noise trap: the searches surfaced USAA's unrelated IPR2021‑01071 and IPR2021‑01074 against USAA's '432, and a ZS Pharma/AstraZeneca '432 asserted in 2022 — neither concerns US 6,355,432.
• No attorney-fee/exceptional-case, antitrust, or DJ action naming the '432.
• No 2026 activity of any kind. The patent expired 2009‑06‑07 per Google Patents ("Expired – Fee Related"), which is consistent with the absence of post‑2011 docket activity — though it does not prove there has been none.

Sourcing limits: I worked from web search results plus the Markman opinion, not from a fresh PACER or CourtListener docket pull. PACER, CourtListener's RECAP docket for D. Del. 1:04‑cv‑00901, and the Darts‑ip family link on the Google Patents page are the sources that would confirm the absence of any additional case with procedural certainty. I did not retrieve the German or UK case numbers.

4. Net assessment for a demand-letter file

There is one assertedly infringed case on this patent, brought by Affymetrix (an operating company) against Illumina (a direct competitor) in its own name, and it settled in January 2008 before the validity phase. The only surviving judicial gloss on the claims is the 2006‑08‑16 Delaware construction of claim 1 — which is a narrowing that requires the bead code be a property separate from the binding polymer. If a current demand cites the '432, the operative defenses are the 2009‑06‑07 expiry, the § 286 six‑year bar, the 2006 construction, and possibly the January 2008 covenant not to sue — not any litigation history showing the patent has been vindicated.

I'll verify the structured "no PTAB activity" signal with targeted searches before drafting.

Now let me check for any AIA trials filed later, and pin down the reexamination and appeal history.

Proceedings overview

Total AIA trial proceedings on file: 0. Zero IPRs, zero PGRs, zero CBM reviews — no institution decisions, no Final Written Decisions, no settlements, no Federal Circuit appeals from the PTAB, and therefore no § 315(e)(2) estoppel running against anyone. The breakdown is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials. That absence is not a sign the patent is untouchable; it is a sign of timing and cash: the PTAB only began accepting IPRs on 2012-09-16, three years after this patent's anticipated expiration of 2009-06-07, and the one real validity war over the '432 (Affymetrix v. Illumina, D. Del. 1:04-cv-00901-JJF) ended in a 2008-01 global settlement with a broad covenant not to sue — before any AIA regime existed. The defensive posture is therefore not "the patent survived two IPRs and is hardened," and also not "claims are canceled." It is: an expired patent whose only litigated claim was narrowed by a Delaware Markman ruling, whose validity was separately put into ex parte reexamination by Illumina in 2007, and which has never been tested at the PTAB. If a demand letter asserts this patent today, the lead defense is the calendar (§ 286 six-year recovery bar plus expiry), not IPR.

No AIA trial proceedings to list

The structured "PTAB proceedings on file" block (USPTO ODP) returns no AIA trials, and independent searching surfaced none: no IPR/PGR/CBM docket number for 6,355,432 appears in any PTAB or litigation-data source reviewed. Per the operating rule to prefer the structured ODP list and not invent proceeding numbers, I am not fabricating a proceeding to fill this section. I also found no evidence that a defensive aggregator such as Unified Patents ever filed against it — the only Unified-sourced item in the record is its litigation tracker entry for the district court case.

Because the AIA-trial section is empty, the following two non-AIA events are the substance a defendant actually needs. They are labeled as outside the scope of the requested proceeding type, not as AIA trials.

90/008,885 — ex parte reexamination (Illumina, Inc. as third-party requester) — NOT an AIA trial

• Type: Ex parte reexamination under 35 U.S.C. §§ 301–307 (pre-AIA reexam practice; no PTAB involvement, no APJ panel, no FWD)
• Filed: requester's filing date reported inconsistently in the public record — 2007-10-19 (patentdocs.org, reporting the four Illumina requests) vs. a reexam filing date of 2007-11-27 (Illumina's press release). Treat the exact date as unverified until the file wrapper is pulled.
• Status: Reexamination ordered by the USPTO; a substantial new question of patentability was found. Illumina asked for all claims of the '432 to be reexamined.
• Judge panel: N/A — no APJs; handled by a Central Reexamination Unit examiner.
• Petition grounds: Prior-art patents/printed publications only (§ 301). Illumina's request covered every claim; the specific art set is in the file wrapper, which I have not retrieved and will not guess at.
• Institution decision: Ordered per Illumina's 2007-12-21 announcement (serial 90/008,885 for the '432; companion reexams 90/008,886 ('716), 90/008,888 ('531), 90/008,889 ('243)).
• Final Written Decision: N/A.
• Outcome / certificate: Not confirmed. I could not verify from available sources whether the reexamination certificate confirmed all claims, confirmed them as amended, or cancelled any claim. Any claim-level statement about the '432's post-reexam scope would be fabrication at this point — pull the reexam certificate and the file history from USPTO Patent Center before relying on claim scope.
• Settlement / termination: The reexamination was overtaken by the 2008 settlement of the parallel litigation (below), but an ex parte reexam is not "settled" — it runs to a certificate on the merits regardless of the parties' deal.
• Appeal: No PTAB appeal identified. (Reexam appeals would have gone to the BPAI, then to the Federal Circuit; none surfaced.)
• Defensive value: This is a ready-made invalidity toolbox. Illumina's 2007 request and the examiner's SNQ determination are public prior-art/obviousness work product you can repackage in a district-court § 282 defense or a DJ action — and, unlike an IPR petitioner, you carry no § 315(e)(2) estoppel from it. The reexam file also gives you a prosecution-history-disclaimer record to mine for claim-narrowing arguments.

Affymetrix, Inc. v. Illumina, Inc., No. 1:04-cv-00901-JJF (D. Del.) — district court, not a PTAB proceeding

• Type: Patent infringement action (asserting five Affymetrix patents, including the '432)
• Filed: 2004-07-26
• Status: Terminated — dismissed by settlement (Illumina filed an 8-K on 2008-01-09 reporting a material definitive agreement entered the same day).
• Panel: Judge Joseph J. Farnan, Jr. (D. Del.)
• Grounds / claims at issue: The '432 was one of five patents asserted ('432, 6,646,243, 5,545,531, 6,399,365, 5,795,716). For the '432, the only claim the court construed was claim 1, the bead-collection claim.
• Claim construction (the key event): Memorandum Opinion and Order, 2006-08-16, 446 F. Supp. 2d 277 (D. Del. 2006) — CourtListener link. The court quoted claim 1: "A collection of beads comprising a plurality of beads which have binding polymers of different target specific sequence attached thereto; said beads being coded with an encoding system whereby the target specific sequence of the polymer attached to the beads can be identified." ('432 patent, col. 82, ll. 51–54.) It then rejected Affymetrix's broader reading and construed "said beads being coded with an encoding system" as "said beads having a property associated with each bead (separate from the binding polymer) that can be used to distinguish one bead from another." The panel reasoning: "Using the binding polymers as the encoding system then, would amount to using a binding polymer to identify itself. This circular identification cannot be what the inventors intended." The court also construed "target specific sequence" as "a known polymer sequence that has affinity for another sequence" — rejecting Illumina's narrower proposal on that term.
• Jury verdict: March 2007 — infringement of "one or more claims" of all five patents-in-suit; 15% royalty; >$16.7 million damages for 2002–2005. The validity and willfulness phases (scheduled for 2008-02-11 and mid-2008) never ran to verdict.
• Settlement: 2008-01 — Illumina paid Affymetrix a one-time $90 million without admitting liability; the parties dismissed U.S., German, and UK actions with prejudice; Affymetrix granted Illumina, its affiliates and Illumina's customers a perpetual covenant not to sue on then-current products and evolutions, plus a four-year covenant for future-technology products, excluding the field of photolithography. Terms beyond that summary are not public in detail.
• Appeal: The '432 case itself was not appealed — it settled.
• Defensive value: Two things. First, the claim-1 construction is binding precedent on the intrinsic record and is the single best non-invalidity attack on the '432: if your accused product relies on tagged binding polymers themselves as the identifier (e.g., sequencing-by-hybridization bead decoding without a separate encoding moiety), it falls outside claim 1 as construed. Second, the covenant not to sue is product-scoped and party-scoped — if you are a customer or privy in the covenant chain, check it before doing anything else.

Strategic summary

Claim status: all claims UNTESTED at the PTAB; none CANCELED. No AIA tribunal has ever invalidated or sustained any claim of the '432. The only claim-level legal event in the record is the 2006-08-16 Delaware construction narrowing claim 1 — the bead-collection claim — so that the encoding system must be a property of the bead separate from the binding polymer. That is a narrowing of scope, not a cancellation. Every other claim of the '432 is, so far as the public record shows, unadjudicated. I do not have a verified claim count or claim-by-claim text for the whole patent from the fetched materials and will not state one; the '432 is a VLSIPS-family continuation claiming priority to 1989-06-07 and it expired 2009-06-07 (Google Patents lists "2009-06-07 Anticipated expiration" and status "Expired – Fee Related"). Because the patent is expired, "which claims survive" is largely academic: surviving claims still cannot support prospective injunctive relief, and damages are bounded by the § 286 six-year lookback (see below).

Estoppel landscape: clean. § 315(e)(2) estoppel applies only to petitioners, real parties in interest, and privies in an IPR/PGR that was instituted and terminated by FWD. There were no institutions here, so no ground is estoppel-barred. Illumina's ex parte reexamination request (90/008,885) creates no statutory estoppel either. Practically, this means a defendant today may freely raise, in district court or a DJ action, any § 102/§ 103 ground — including all art Illumina put before the examiner in 2007 and all art it did not, plus the § 112 and § 101 theories no one has ever litigated. Conversely, remember § 315(a)(1)/(b): if you have already served or filed a civil action challenging validity, or you were served with a complaint more than a year ago, an IPR is not available to you anyway — and given the 2009 expiry it would be a poor use of budget.

Pattern signals. No repeat petitioner — there is no petitioner at all. No patent-owner PTAB appeal strategy exists to read, because the patent owner never had a PTAB case (its aggressive posture was in district court: the 2004 D. Del. suit, a 2007 wave of U.S./German/UK filings, all resolved by the $90M January 2008 deal). No defensive aggregator in the chain. The reactive filer of record is Illumina, and its chosen tool in 2007 was ex parte reexamination — chosen precisely because IPRs did not yet exist and reexam is cheaper and creates no estoppel against the requester. The contemporaneous post-trial press coverage makes the strategy explicit: Illumina told investors it expected the PTO "to consider closely the applicability of the prior art in their evaluation of whether these patents should be amended or invalidated in their entirety."

Recommended next steps

1. If you have received a demand or complaint asserting the '432 today, lead with expiration and the statute of limitations. Anticipated expiration 2009-06-07 means no accused act after that date can infringe, and 35 U.S.C. § 286 bars recovery for any infringement more than six years before filing — the last actionable window closed in 2015. If the demand cites claim 1, also cite the 2006-08-16 construction: "said beads having a property associated with each bead (separate from the binding polymer) that can be used to distinguish one bead from another" (FWD-stage opinion: 446 F. Supp. 2d 277, D. Del. 2006-08-16). There is no FWD to quote because no FWD exists — do not let opposing counsel imply otherwise.
2. Pull the 90/008,885 file wrapper and the reexamination certificate from USPTO Patent Center before you take any position on claim scope. I could not confirm the reexam outcome from the sources available; it is the single largest open factual gap in this report, and it may show claims confirmed as amended (creating prosecution-history arguments and possible intervening rights) or claims cancelled outright.
3. Check whether you are inside the January 2008 covenant. Affymetrix's perpetual covenant ran to Illumina, its affiliates, and Illumina's customers for then-current products and evolutions (plus four years for future-technology products, excluding photolithography). A demand against a downstream user in that chain may be barred by contract even before the patent-law defenses.
4. Confirm the litigation/family record yourself. The Delaware case is tracked at Unified Patents litigation data, 1:04-cv-00901 and the Darts-ip family link on the Google Patents page. The patent's own record is at patents.google.com/patent/US6355432B1/en. If you need a court opinion beyond the Markman order, CourtListener and PACER are the sources — I found no Federal Circuit opinion on the '432 itself, only the separate Affymetrix/Illumina appeal over Illumina's 7,510,841 and 7,612,020 (affirmed dismissal, August 2011), which does not involve the '432.
5. Watch for newly indexed activity, but do not expect it. The ODP ingest shows nothing, and I found nothing contradicting it. The realistic residual risk is not a PTAB filing on an expired 1989-priority patent; it is a demand letter over pre-2009 conduct, litigation over a covenant's scope, or a reissue/continuation-family theory — none of which an IPR would have addressed in the first place.

I'll research the assignment record, litigation history, and entity status for US 6,355,432 before writing.

Let me dig into the recordation details, the litigation outcome, and Affymetrix's corporate fate.

Let me check the litigation outcome and try to pin down reel/frame data.

US 6,355,432 B1 — "Products for detecting nucleic acids" — Ownership Chain Reconstruction

Scope note (read first): I was able to reconstruct the chain from the patent record itself (Google Patents legal-events tab, which mirrors the post-issuance recordations), the D. Del. litigation docket/opinions, and press/SEC coverage. I was not able to retrieve reel/frame numbers or correspondent-of-record names — the USPTO Assignment Center (assignmentcenter.uspto.gov) is a JavaScript application that my fetches could not render, and the secondary mirrors do not expose those fields. Reel/frame and correspondent fields below are therefore marked not retrieved rather than estimated. You should pull those four records directly in the Assignment Center UI (search by patent number 6355432) before relying on this for standing or due-diligence work.

Inventors

Pattern check: No fire-sale precursor. The "all inventors departing within 12 months of filing" pattern is not present — the 2002-01-02 recordation shows all three inventors assigning to the Affymax/Affymetrix side, and the patent issued 2002-03-12 with the same corporate family as assignee. This is a textbook in-house, founder-inventor portfolio, not an orphaned inventor-owned asset.

Original assignee

Affymetrix, Inc. (Delaware corporation; NASDAQ: AFFX at the time of issuance; listed as current assignee on both Google Patents and Unified Patents' patent page for US-6355432-B1).

• Products embodying the claims: Yes — Affymetrix was the pioneering commercial microarray ("GeneChip") company built on the VLSIPS technology described in this specification. Thermo Fisher's 2016 closing release states Affymetrix had shipped more than 2,300 microarray systems and had ~1,100 employees with manufacturing in Cleveland, San Diego, Vienna and Singapore.
  • Caveat on claim scope: the asserted independent claim of the '432 patent (claim 1, col. 82) is a collection of encoded beads, not a planar array. The bead-claim/product fit is arguable; Affymetrix's own commercial embodiment was the planar GeneChip. That is a claim-scope nuance, not an ownership finding.
• Primary line of business: Life-science tools / genetic analysis — consumables (arrays, reagents, antibodies), instrumentation and software for genomics research, clinical and applied markets.
• Current status: Acquired. Thermo Fisher Scientific completed its acquisition of Affymetrix on 2016-03-31 for $14.00/share cash, ~$1.3 billion, following a competing bid by Origin Technologies (a vehicle formed by former Affymetrix executives) that was rejected and then withdrawn. Affymetrix stock was delisted from NASDAQ. No bankruptcy, no Chapter 7/11, no assignment for the benefit of creditors.

Assignment timeline

Four post-issuance records appear on the patent's legal-events record. All four are visible on Google Patents at https://patents.google.com/patent/US6355432/en (Legal Events / "Assigned to" entries).

• Executed date not shown / recorded 2002-01-02 — Reel not retrieved/Frame not retrieved

  • Conveyance: Assignment of assignors' interest ("ASSIGNMENT OF ASSIGNORS' INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: Stephen P. A. Fodor; Dennis W. Solas; William J. Dower
  • Assignee: AFFYMAX TECHNOLOGIES N.V.
  • Correspondent: not retrieved
  • Context: Inventor-to-employer perfection of title, recorded on the eve of grant — housekeeping, not a market transfer.

• Executed date not shown / recorded 2002-01-02 — Reel not retrieved/Frame not retrieved

  • Conveyance: Change of Name
  • Assignor: AFFYMAX TECHNOLOGIES N.V.
  • Assignee: AFFYMETRIX, INC.
  • Correspondent: not retrieved
  • Context: Internal reorg — change of name only; no change in beneficial ownership.

• Executed date not shown / recorded 2012-06-27 — Reel not retrieved/Frame not retrieved

  • Conveyance: Security Agreement
  • Assignor: AFFYMETRIX, INC.
  • Assignee: GENERAL ELECTRIC CAPITAL CORPORATION, AS AGENT
  • Correspondent: not retrieved
  • Context: Securitization — collateral pledge supporting the $100,000,000 Credit Agreement dated as of June 25, 2012 among Affymetrix, Inc. as Borrower, GE Capital as Agent, Silicon Valley Bank as Syndication Agent, and GE Capital Markets/SVB as joint lead arrangers (agreement text indexed at contracts.justia.com). A lien, not a sale.

• Executed date not shown / recorded 2015-11-13 — Reel not retrieved/Frame not retrieved

  • Conveyance: Release by Secured Party
  • Assignor: GENERAL ELECTRIC CAPITAL CORPORATION, AS AGENT
  • Assignee: AFFYMETRIX, INC.
  • Correspondent: not retrieved
  • Context: Lien release — the 2012 encumbrance is extinguished; Affymetrix (later Thermo Fisher) held clean title.

One transfer that does not appear as a per-patent recordation: the 2016-03-31 Thermo Fisher merger. Mergers effected by operation of law frequently are not recorded patent-by-patent, and Google Patents still lists the current assignee as Affymetrix Inc. Treat Thermo Fisher's ownership of this asset as unclear at the recordation level — verify by searching the Assignment Center for Thermo Fisher/Affymetrix merger recordations, or by confirming Thermo Fisher's chain in its 10-K IP disclosures.

Correspondent recurrence: Unclear / not assessable. I could not retrieve correspondents; no recurrence finding can be made either way.

Timeline diagram

timeline
    title Ownership of US 6355432
    1989 : Priority application filed
    2000 : Continuation application filed
    2002 : Patent issued 12 March
         : Inventors assign to Affymax Technologies NV
         : Name change to Affymetrix Inc
    2004 : Affymetrix sues Illumina in Delaware
    2006 : Markman claim construction ruling
    2007 : Jury finds Illumina infringes
    2008 : Validity phase and USPTO reexamination
    2009 : Anticipated patent term ends 7 June
    2012 : GE Capital security interest recorded
    2015 : Security interest released
    2016 : Thermo Fisher acquires Affymetrix

NPE / troll-pattern signals

1. Shell-entity transfer — not present. No assignment to any entity bearing an "IP / Patents / Licensing / Holdings / Ventures" suffix appears in the record. The only non-operating transferee on the chain is General Electric Capital Corporation, as Agent (recorded 2012-06-27), which is a secured lender under a $100M credit facility, not a licensing vehicle — and its interest was released 2015-11-13, restoring title to Affymetrix.

2. Known asserter in the chain — not present. Current assignee per Google Patents and Unified Patents is Affymetrix Inc. No match against Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or any Spangenberg entity. The named plaintiff in the only patent-specific suit is Affymetrix, Inc. in its own name (D. Del. 1:04-cv-00901-JJF).

3. Repeat correspondent across the chain — unclear. Correspondent-of-record data not retrieved (see scope note). No finding possible.

4. Cascading transfers — not present. Four recordations spread over ~13 years (2002, 2002, 2012, 2015), none involving chained LLCs, and none within 24 months of another substantive transfer. This is the opposite of the cascading pattern.

5. Pre-litigation transfer — not present. The 2002-01-02 recordations predate the July 2004 Delaware complaint (1:04-cv-00901) by ~2.5 years and consist of a change of name plus an inventor assignment. The one pre-suit-adjacent transaction is the 2012-06-27 GE Capital security agreement, which post-dates the 2004 suit by eight years and is a financing lien, not a standing-arranging transfer.

6. Bankruptcy fire-sale — not present. Affymetrix never filed Chapter 7 or 11. It was acquired in a solvent, all-cash $1.3B merger (2016-03-31). No Kodak/Nortel/Polaroid-style asset sale.

7. Privateering — not present. Affymetrix asserted the '432 against Illumina, a direct competitor, in its own name, at its own expense, with its own in-house counsel of record (Michael J. Malecek, Daniel R. Reed, Andrea L. Gross of Affymetrix, Emeryville, CA) alongside Morris, Nichols, Arsht & Tunnell. The relationship was adversarial in both directions — Illumina later sued Affymetrix in the Western District of Wisconsin (Illumina, Inc. v. Affymetrix, Inc., 3:09-cv-00277), where Affymetrix won summary judgment of noninfringement.

8. Defensive aggregator (anti-NPE) — not present. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN.

Additional record context worth carrying into the file: the patent's anticipated expiration is 2009-06-07, and the legal status is Expired – Fee Related. It was reexamined during the Affymetrix–Illumina dispute (USPTO ordered reexamination of the '432 and '6,646,243; reported by GenomeWeb, 2008), and Illumina moved for summary judgment of invalidity of asserted claims 2, 5, 8 and 9 (D. Del. Docs. 279, 284, filed July 2006). A March 2007 jury found Illumina infringed five Affymetrix patents with $16.7M in past damages and a 15% royalty; the validity phase was set for February 2008. I did not confirm the final disposition of the invalidity motion or the reexamination certificate in this session — treat that as an open item.

Verdict

Operating-company assertion.

The chain contains no licensing entity at all: the only post-issuance transfers are an inventor assignment and a corporate change of name (both recorded 2002-01-02 to Affymetrix), plus a secured-lien pledge and release to/from GE Capital as Agent (2012-06-27 / 2015-11-13) under Affymetrix's $100M credit facility. The sole assertion was brought by Affymetrix, Inc. in its own name against competitor Illumina, Inc. in D. Del. 1:04-cv-00901-JJF, by Affymetrix in-house and outside counsel — and Affymetrix is an operating company that shipped over 2,300 microarray systems before being acquired by Thermo Fisher on 2016-03-31 for ~$1.3B. With the patent expired since 2009-06-07 and title resting with Affymetrix/Thermo Fisher, there is no NPE exposure here.

Verify at:

• https://assignmentcenter.uspto.gov/ (search patent number 6355432 — pull reel/frame and correspondent for all four records)
• https://patents.google.com/patent/US6355432/en (Legal Events)
• Litigation: D. Del. 1:04-cv-00901-JJF; 446 F. Supp. 2d 277 (D. Del. Aug. 16, 2006); Illumina v. Affymetrix, W.D. Wis. 3:09-cv-00277

I'll research the prior-art citations on the face of US 6,355,432, then assess anticipation per claim.

Let me pull the actual face-page citations from an independent full-text source.

Let me try to surface the actual face-page citation lists from Espacenet and Google Patents.

Prior Art of US 6,355,432 B1 — Cited-Reference Screen and §102 Mapping

Pat. No. 6,355,432 B1, "Products for detecting nucleic acids," Fodor, Solas & Dower; app. 09/585,659 filed 2000‑06‑02; issued 2002‑03‑12; assignee Affymetrix Inc.; expired 2009‑06‑07. Pre‑AIA statute controls (application filed before 2013‑03‑16), so §§102(a), (b), (e), (g) all apply.

This section builds on — and does not repeat — the earlier summary, litigation, PTAB, assignment, and strategic sections. Two of those carried open gaps that this pass closes (flagged in §6 below).

1. Sourcing caveat — read before using the tables

I did not successfully retrieve the literal, verbatim front‑page "References Cited" block of the '432 in this session. Google Patents' patent page for US6355432B1 renders that block behind the anchor fragment, and Espacenet's "Documentos citados" / "Documents cited" tab did not return its contents to my queries. I will not reconstruct that block from memory.

What I could verify, and what the tables below are built from:

Explicit non‑finding: I cannot tell you the total number of references on the '432's face, and any list presented as "all citations for 6355432" would be fabricated. The "231" figure belongs to the Unified Patents aggregate, not to the patent's printed citation count.

Number disambiguation (no auto‑correction applied):

• US 6,451,536 B1 carries the identical title "Products for detecting nucleic acids" with a different priority date (1990‑12‑06) and appears in the '432's Cited By list. A title‑only search will conflate the two. They are distinct patents.
• The Google Patents sibling for the '432 is US 6,440,667 B1, "Analysis of target molecules using an encoding system," from Ser. No. 09/362,089, filed 1999‑07‑28 — the immediate parent of the '432's application.
• One search result renders the assignee as "Affymetrix Lnc." — an OCR artifact on the DE 3 505 287 C2 page, not a distinct entity.
• Sitemap hits for US6355432 reproduce claim 1's text as the title line; that is Google's sitemap rendering, not a second title.

2. The §102 target and the critical‑date problem (this is the whole ballgame)

Claims to be tested: 1 (independent) and 2–22 (dependent). Claim 1, per the Espacenet claim set and the court's quotation at 446 F. Supp. 2d 277:

"A collection of beads comprising a plurality of beads which have binding polymers of different target specific sequence attached thereto; said beads being coded with an encoding system whereby the target specific sequence of the polymer attached to the beads can be identified."

Three elements must appear in a single reference for anticipation: (i) a collection of beads; (ii) a plurality of beads bearing binding polymers of different target‑specific sequence; (iii) an encoding system that identifies which sequence sits on which bead.

And the construction compounds it: per the 2006‑08‑16 D. Del. opinion, the encoding system must be "a property associated with each bead (separate from the binding polymer)." So a reference that "identifies" bead‑bound probes by sequencing or hybridizing the bound probe itself does not anticipate claim 1 as construed. That construction knocks out a large slice of the bead‑decoding art.

Critical dates. The '432 claims §120 benefit back through: 09/362,089 (1999‑07‑28) ← 09/056,927 (1998‑04‑08, now 6,197,506) ← 08/670,118 (1996‑06‑25, now 5,800,992) ← 08/168,904 (1993‑12‑15, aband.) ← 07/624,114 (1990‑12‑06, aband.) ← 07/492,462 (1990‑03‑07, now 5,143,854) ← 07/362,901 (1989‑06‑07, aband.).

This is the single highest‑value prior‑art question on the patent, and it is unresolved in the public record I retrieved. Note the irony: the '432's own reexamination requester, Illumina, holds the 1994 encoded/tagged‑oligonucleotide art (US 5,846,719) that is closest to claim 1's encoding element — but that art only bites if the claim loses its pre‑1990 priority.

3. Table A — References cited in the '432 specification [SPEC]

These are verifiable from the patent's own text. Most are the applicant's own family work or enabling technology, not third‑party §102 art.

Non‑patent literature cited in the specification [SPEC] — the ones with genuine §102 weight:

• Sanger et al. (1977) PNAS 74:5463‑5467; Maxam & Gilbert (1980) Methods Enzymol. 65:499‑559 — the two sequencing baselines. Neither discloses bead‑bound encoded probe collections; §103 background only.
• Lysov, Yu. et al. (1988) Dokl. Akad. Nauk SSSR — the sequencing‑by‑hybridization theory. This is the most substantively relevant NPL item and the direct conceptual ancestor of the spec's §4 worked example. §103 anchor against claims 2–11 and 18/22; not a §102 reference for claim 1, because it is a theoretical/algorithmic disclosure of array hybridization, not encoded beads.
• Evans et al. (1989) PNAS — cosmid multiplex mapping by overlap. Mapping theory; §103 vs. dependent claims only.
• Kanehisa (1984) NAR 12:203‑213 — hybridization stability/length trade‑offs; relevant to §103 on claims 5–8 (probe‑length ranges).
• Patchornik (1970) JACS 92:6333‑6335 and Amit et al. (1974) J. Org. Chem. 39:192‑196 — photoremovable protecting groups. Pre‑1988, so they are clean §102(b) art — but only against photochemistry limitations, and the '432's claims contain none. §103 background.
• Chatterjee et al. (1990) JACS 112:6397‑6399 — crosslinking reagents; post‑1988, mapping to claims 14–16/19.
• Harlow & Lane (1988); Goding (1986) — antibody production. Against claim 13 (antibody binding sites); §103 only.
• Innis et al. (1990) PCR Protocols; Gait (1984) Oligonucleotide Synthesis — method references.
• Tuerk & Gold (1990) Science 249:505‑510 — random/defined‑sequence nucleic acid libraries. Against the "all possible sequences" selection of claims 2–11; post‑1989, so priority‑dependent.
• Sheldon et al. US 4,582,789 (above); Carrano et al.; Wood et al.; Stryer (1988) — background.

Bottom line on Table A: not one [SPEC] item discloses a collection of beads bearing different target‑specific polymers plus a separate bead‑encoding system. The applicant's own family patents are not §102 art at all. The specification's citation list is a scientific bibliography, not an anticipation threat.

4. Table B — Third‑party patents in the retrieved art list [UP]

Ranked by proximity to claim 1. Dates are priority dates as reported by the Unified Patents portal; I have not independently verified each. I flag every mapping as an assessment, not a holding.

B‑1. Closest art, and it is uncomfortably close

B‑2. Pre‑1988 art — clean §102(b) status, but no encoding element

B‑3. Post‑1989 art — §102 only if priority is broken

5. Claim‑by‑claim §102 assessment

Stated as a screening opinion. "Strongest available reference" means the best candidate on the material I retrieved — not a conclusion that any of them in fact anticipates.

6. Gaps closed, and one non‑contradiction flagged

1. Claim 22 is now complete. The earlier summary recorded claim 22 as truncated ("…wherein the given length of the oligonucleotide attached to a plurality of…"). The Espacenet claim set retrieved in this pass supplies the full text: "wherein the given length of the oligonucleotide attached to a plurality of beads is selected from the group consisting of oligonucleotide sequences having a variable number of nucleotides." This is a completion of an explicit open item, not a contradiction. The earlier document's caution was correct and is now retired.
2. Claim count confirmed at 22, claim 1 confirmed as the sole independent claim — consistent with the earlier sections and the court's treatment.
3. No contradiction found with any prior section. The prior sections' most consequential open item — the disposition of reexamination 90/008,885 — remains open, and it matters more here than anywhere: the operative §102 art set for this patent is whatever Illumina put in that 2007 request, and I still could not retrieve the file wrapper or the reexam certificate.
4. One new literal discrepancy to log: the Unified Patents portal's art list gives US‑6025136‑A a "1994‑12‑08" priority and US‑5846719‑A a "1994‑10‑12" priority. Both post‑date the '432's 1989 priority by years, which is the opposite of what one expects in a "cited‑in" art list for a 1989‑priority patent — consistent with the '432's actual application having been filed 2000‑06‑02. It is a reminder that this family's §102 universe is defined by the 2000 filing date unless and until §120 benefit is affirmatively established for the bead/encoding limitations.

7. What to pull next (in priority order)

1. The '432's literal front‑page "References Cited" block — from the USPTO Patent Center / Global Dossier PDF of US6355432B1 (the 55‑page original document fetched via Espacenet originalDocument is the reliable vehicle; page 1 carries it). Do not rely on aggregator art lists. My Tables B and C are a screen, not the record.
2. Reexamination 90/008,885 file wrapper — the request's art set plus any examiner SNQ determination. This is the actual, curated §102 record created by a skilled adversary with full discovery, and it is free of §315(e)(2) estoppel.
3. The §120 support question. For each of claims 1–22, identify the earliest application in the chain (07/362,901 / 07/492,462 / 07/624,114 / 08/168,904) whose specification provides §112 ¶1 written‑description support for the encoded bead limitation. If none does, the effective date collapses toward 1993‑12‑15 or 2000‑06‑02 — at which point US 5,846,719, US 6,025,136, US 5,871,697 and the 1992 Brenner‑Lerner encoded‑combinatorial‑chemistry literature all become live §102 art, and claim 1's validity posture changes materially. This is the pivot point for the entire prior‑art analysis.
4. Then, and only then, screen §102(g)/§102(a) "prior invention by another" for bead‑based encoded libraries (split‑pool synthesis with separate tags) in the 1985–1989 window — the category the retrieved art list does not appear to represent well.

Standing caveat, per the operating rules: I did not retrieve the verbatim statutory citation list, and several issue/filing dates above are stated as reported by secondary sources rather than verified from primary documents. Where I marked "not verified," treat the value as unconfirmed rather than as a finding, and prefer the front‑page PDF when the two differ.

I'll ground this analysis in the reference record for the '432, then build the § 103 combinations.

Key finding: Needels et al. 1993 — the seminal encoded-bead paper — lists W. J. Dower and R. W. Barrett of Affymax Research Institute as authors, i.e., the same inventive/assignee family as the '432. Let me pin down the remaining reference dates and the reexam outcome.

Obviousness Analysis — US 6,355,432 B1 under 35 U.S.C. § 103

0. Scope, and a consistency check against the earlier sections

This builds on the four sections already generated. Two points of cross-reference first, because they control everything below:

• No contradiction found with the earlier sections. I independently re-confirmed the reexamination serials and dates: Illumina filed ex parte reexamination requests on 2007-10-19, and the USPTO ordered reexamination of the '432 as control no. 90/008,885 (announced 2007-12-21), alongside 90/008,886 ('716), 90/008,888 ('531) and 90/008,889 ('243). Source: https://patentdocs.org/2008/01/02/uspto-to-reexam-2/
• One materially new fact, not in the earlier sections, that changes the § 103 picture: W. J. Dower and R. W. Barrett — Dower being a named inventor on the '432 — are co-authors of the Needels et al. 1993 PNAS encoded-bead paper, out of Affymax Research Institute, Palo Alto. The single most on-point piece of encoded-bead prior art is therefore the patentee's own assignee's work. That is a § 102(a)/(b) versus § 103(c) problem, not just a technical-identity problem.

Critical caveat on claim text: I do not have verbatim claim language for this patent. Claim 1 comes from the Markman opinion's quotation of col. 82, ll. 51–54, and claims 2–22 are summarized from the Espacenet listing in the earlier sections. Claim 22 remains truncated and unverified, and I cannot rule out a second independent claim. I have not invented claim text anywhere below; where I am working from the summary, I say so.

1. The gating issue: what is the effective filing date of the claim?

Under pre-AIA § 103(a) the question is what "would have been obvious at the time the invention was made." Priority chain (from the earlier section, confirmed against Google Patents' legal-events timeline):

09/585,659 (2000-06-02) ← 09/362,089 (1999-07-28) ← 09/056,927 (1998-04-08) ← 08/670,118 (1996-06-25) ← 08/168,904 (1993-12-15) ← 07/624,114 (1990-12-06) ← CIP of 07/492,462 (1990-03-07) ← CIP of 07/362,901 (1989-06-07).

Claim 1 is a bead-collection-plus-encoding-system claim. Whether it gets the 1989 date depends entirely on whether the 1989 and 1990 applications (both directed to light-directed planar synthesis on a substrate) provided pre-AIA § 112 ¶ 1 written-description support for a collection of beads bearing different binding polymers, each bead coded with a separate encoding system. That is the pivot. Rank the references accordingly:

Consequence. If the claim is entitled to 1989-06-07 or 1990-12-06, the obviousness case is weak-to-fair, resting on Lysov, Macevicz and the applicants' own VLSIPS disclosures. If the claim is entitled only to 1993-12-15 or later, the case is strong and, in my assessment, close to conclusive — because the two references that together use every element of claim 1 (Lam and Brenner & Lerner) both become § 102(b) statutory bars.

2. Level of ordinary skill

A POSITA here would hold a Ph.D. (or equivalent) in molecular biology, biochemistry or organic chemistry, with roughly 2–4 years of hands-on experience in solid-phase synthesis of biopolymers and library screening. By the 1990–1993 window, that person routinely used split-and-pool synthesis on resin beads, phosphoramidite oligonucleotide chemistry, FACS, Edman microsequencing, and probe-hybridization assay design, and was a regular reader of Nature, PNAS, Science and JACS. This is the art level both parties implicitly accepted in the Delaware case; nothing in the '432's specification suggests otherwise.

3. Claim 1 — element decomposition

4. Combination A (primary): Lam + Brenner & Lerner → anticipates the substance of claim 1

This is the combination that matters.

Lam et al. (1991-11-07) teaches 1.1 and 1.2 directly: a "one-bead, one-peptide" library in which each of millions of beads bears a single peptide sequence, screened against a labelled acceptor, with the binding bead physically isolated. What Lam does not teach is a bead-borne code: Lam identifies the hit by Edman microsequencing the isolated bead — destructive, low-sensitivity, and limited to sequenceable polymers.

Brenner & Lerner (1992-06-15) teaches 1.3 and 1.4 in exactly the form the Delaware construction requires. Their abstract states that "a process of alternating parallel combinatorial synthesis is used to encode individual members of a large library of chemicals with unique nucleotide sequences," that "each of these oligodeoxynucleotides had a signature (unique) sequence corresponding to the linked peptide," and that "the nature of the chemical structure bound to the receptor is decoded by sequencing the nucleotide tag." The tag is a distinct chemical entity from the library member — the very "separate from the binding polymer" requirement the court adopted, and the very "circular identification" the court rejected as outside the claim. Source: https://pubmed.ncbi.nlm.nih.gov/[1608946](/patent/1608946)/

Motivation to combine — express, not imputed. Brenner & Lerner state their own objective as remedying precisely Lam's deficiency: the tag "can be amplified by replication and utilized for enrichment of the bound molecules by serial hybridization," and the method is described as one that "can be used to identify a few members in a large chemical library." Both references operate on the same physical format (porous resin beads; split-and-pool synthesis), so combination requires no change in library architecture — only the addition of a coding moiety that the second reference exists to supply. This is the paradigm case of an explicit teaching, suggestion or motivation, satisfying even the pre-KSR TSM test that would have governed in the 2006–2008 litigation window.

Reasonable expectation of success is documented, not speculative. Needels et al. (1993-11-15) — from the patentee's own Affymax Research Institute — reduced the combination to practice on 10 µm beads at ~10⁶ sequences with FACS isolation and PCR decoding. And note the direction of that evidence: an ~18-month gap between Lam's stated problem and a working encoded-bead system, by the same corporate family that owns the '432, is evidence of obviousness in the KSR sense — a finite number of identified, predictable solutions (peptide tag, oligonucleotide tag, chemically inert tag, RF tag), all pursued successfully in the same window. It is not evidence of a nonobvious advance.

5. Combinations for the dependent claims

Claims 2–11 (oligonucleotide probes of a given length; coverage ≥ ~20%/≥ ~70%; absolute counts ≥10,000/100,000/1,000,000).

• Substitute the binding polymer. Claim 12 makes explicit what claim 1 already permits: a "polynucleotide or polypeptide" binding polymer. To reach an oligonucleotide probe set, combine Combination A with Macevicz, WO 90/04652 (SBH using short defined probes) and Lysov et al. (1988) — both of which establish that the operative recognition reagent is a short oligonucleotide of defined sequence. Macevicz is cited on the face of the '432 and is discussed in its specification.
• The N-mer counts. Fodor et al., Science (1991-02-15) and US 5,143,854 supply the "all possible sequences of length N" concept and the arithmetic (65,536 8-mers; 262,144 9-mers; 1,048,576 10-mers) — the same arithmetic the '432 specification recites. Populating a bead format with a complete N-mer set is a change of substrate, not of principle.
• The coverage percentages (≥20%, ≥70%). Routine optimization of a library-completeness parameter, and the '432's own specification sets the floor low: "a substantial fraction will preferably be at least about 70%… About 20% would be preferred, more preferably at least about 30% would be desired." A claim whose lower bound is the specification's own stated preference is a textbook KSR design choice.
• The absolute counts (10,000 / 100,000 / 1,000,000). Mere recitations of library size, squarely met by Lam ("millions"), Needels (~10⁶ sequences on 10 µm beads), and the '432's own statement that "one million regions would take about 11 centimeters square."
• This is why claims 2, 5, 8 and 9 are the ones Illumina attacked on summary judgment in July 2006 (D.I. 279) — they are the length/coverage/count claims, i.e., the trivial-design-choice tier. Caveat, as flagged earlier: I could not confirm from a primary source that the resulting order was actually entered, and its text conflicts with the March 2007 verdict of infringement of "one or more claims" of the '432.

Claim 13 (binding polymer = protein with an enzyme or antibody binding site). Combination A plus the '432's own cited Harlow & Lane (1988) and Goding (1986) antibody methodology, and its own discussion of "promoter proteins, restriction enzymes, and the like." Antibody-based recognition reagents were the standard alternative to nucleic acid probes in 1990–1993.

Claim 17 (encoding system = magnetic, shape-based, colour-based, or a combination). Every branch is met by the known alternative encoding modalities: Ohlmeyer et al. (1993) (inert haloaromatic tags read by electron-capture GC — a code wholly independent of the binding polymer), Kerr et al. (1993) (peptide-based coding), Nicolaou et al. (1995) (RF SMART tags). Claim 17's alternatives also appear verbatim in the '432's own specification as generic knowledge — "An encoding system may include a magnetic system, a shape encoding system, a color encoding system, or a combination of any of these, or any other encoding system." Selection among known encoding modalities to suit a known bead format is a predictable mechanical substitution. Confidence note: I have not independently verified a specific pre-1993 shape-encoding reference; the magnetic/colour/chemical/RF branches are better supported.

Claims 14–16, 19 (glass surface with amine or hydroxyl attachment; oligodeoxyribonucleotides via phosphate to hydroxyls on a hydrophilic organic coating; TEFLON copolymer graft with hydrophilic layer terminating in hydroxyls). These are the VLSIPS surface-chemistry limitations, met by US 5,143,854 (Pirrung et al.), US 5,252,743 (Barrett et al.), and the '432's own cited Dattagupta, US 4,542,102 / US 4,713,326 and Chatterjee et al. (1990) JACS 112:6397–6399. Motivation: attaching oligonucleotides to a hydrophilic, hydroxyl-terminated derivatized surface was the established practice; KSR treats the selection of a known support material presenting known reactive groups as within the ordinary skill level absent a showing of unexpected results. § 103(c) caution: '854, '743 and '908 are Affymax/Affymetrix-family documents that may be disqualified as § 103 art if they qualify only under § 102(e)/(f)/(g) and were commonly owned. This is why the statutory-bar references (Fodor Science 1991; Lam; Brenner & Lerner; Ohlmeyer) carry the load.

Claims 20–21 (nucleotide analogue; reusable beads).

• Claim 20: Macevicz, WO 90/04652, which the '432 specification itself identifies as describing "Unnatural nucleotides or nucleotides exhibiting modified specificity of complementary binding." That is the applicant's own acknowledgement that analogues were known.
• Claim 21: the '432's specification discusses disrupting hybrid duplexes and recycling the matrix at length; reagent reuse was routine in bead- and solid-phase formats.

6. Counterarguments a patentee would raise — and how they fare

1. Priority defeats the art (the strongest defence). If the 1989 or 1990 applications describe encoded beads, Lam, Brenner & Lerner, Needels, Kerr and Ohlmeyer all drop out and the § 103 case must be rebuilt from Lysov/Macevicz/VLSIPS alone. This is not a formality; it is the whole ballgame, and it is an enablement/written-description attack dressed as a § 103 argument. A defendant must obtain the 1989 and 1990 applications and show they disclosed only positionally addressed planar arrays — the patent's own specification frames bead coding as an alternative to positional addressing, which is a useful admission.
2. Teaching away. Brenner's alternating peptide/tag synthesis requires orthogonal protecting-group chemistry; one might argue Lam's beads cannot host a co-synthesised tag. Needels refutes this empirically and shows the solution (bifunctional linkers with differentiated amine/hydroxyl functionality) was within the art — and again, by the patentee's own laboratory.
3. Common ownership (§ 103(c)). Disqualifies Affymax/Affymetrix § 102(e)/(f)/(g) art from § 103 consideration — but does not touch § 102(b) art, including the inventors' own Fodor Science 1991 publication and the third-party Lam/Brenner/Ohlmeyer references.
4. Secondary considerations. The nexus problem is severe: the '432's claim 1 is a bead collection, whereas the commercial product the earlier sections identify as Affymetrix's embodiment — the planar GeneChip — does not read on it. Affymetrix's ~2,300 systems and ~$1.3B acquisition value therefore do not presumptively tie to the claimed subject matter. Conversely, that Illumina independently arrived at encoded beads (the BeadChip) is as consistent with obviousness as with copying.

7. What I could not verify

1. The reexamination outcome. I still cannot state whether 90/008,885 confirmed the claims, confirmed them as amended, or cancelled any claim. This is the largest open factual gap and the single best repository of the art Illumina actually used — the file wrapper should be pulled from USPTO Patent Center before any § 103 position is finalised.
2. Full claim text. Claim 22 is truncated; I cannot rule out a second independent claim.
3. The D.I. 279 summary-judgment order. As previously flagged, the "claims 2, 5, 8 and 9 are adjudged invalid" text appears to be a proposed order and conflicts with the March 2007 verdict.
4. Pre-1993 shape/magnetic encoding art. I did not verify a specific reference for those branches of claim 17.
5. Precise § 102(e) dates for US 5,573,905 (filed 1992-03-30 per a court filing in Illumina v. Scripps) and US 5,708,153 (priority asserted 1991-09-18) — I am resting these on secondary sources and would confirm in Patent Center.

Bottom line: The § 103 case against claim 1 is a clean two-reference combination — Lam et al., Nature 354:82–84 (1991) for the bead library, and Brenner & Lerner, PNAS 89:5381–5383 (1992) for the separate, decodable bead-borne code — joined by an express motivation stated in the second reference, and confirmed as workable by the patentee's own Affymax group in Needels et al. (1993). The entire question is whether those references are prior art, which turns on the CIP priority chain, not on the merits of the combination.