US 5974164

I'll search for the specific patent number and any related litigation dockets.

Let me dig for the claim text and any litigation specifically involving this patent.

Let me look for the claim language and any specific litigation/docket references.

Executive summary

I ran open-web searches targeting the exact string 5,974,164 (plus CAFC-2026 framing). No results referencing a Court of Appeals for the Federal Circuit 2026 docket for this patent surfaced. What did surface was (a) the patent's own Google Patents/OSTI records, (b) an Affymetrix–Hyseq interference naming this patent, and (c) many later patents citing it as prior art. I did not directly query USPTO PatentCenter/Public Search or the CAFC docket system in this session, so treat the absence of a CAFC hit as "not found," not as "verified none."

---

1. Bibliographic data (per the authoritative Google Patents record; treated literally)

Inventorship caveat: the Google Patents bibliographic block names only Mark S. Chee as inventor for 08/531,137, but the recorded 1996-03-11 assignment lists assignors Jevons, Luis C.; Wang, Chunwei; Bernhart, Derek H.; Chee, Mark S.; and Lipshutz, Robert J. The same five inventors appear on the closely related US 5,795,716 ("Computer-aided visualization and analysis system for sequence evaluation"), filed 1994-10-21 as application 08/327,525 (OSTI ID 871785). US 6,733,964 states: "Application 09/049805 is a continuation of application No. 08/327,525, filed on Oct. 21, 1994, granted, now 5795716." So the '164 (filed 1995-10-16, claiming 1994-10-21 priority) sits in the same family, but I could not confirm from the retrieved text the exact §120 relationship between 08/531,137 and 08/327,525.

2. Abstract (verbatim, per OSTI ID 872618 / Google Patents)

"A computer system (1) for analyzing nucleic acid sequences is provided. The computer system is used to perform multiple methods for determining unknown bases by analyzing the fluorescence intensities of hybridized nucleic acid probes. The results of individual experiments may be improved by processing nucleic acid sequences together. Comparative analysis of multiple experiments is also provided by displaying reference sequences in one area (814) and sample sequences in another area (816) on a display device (3)."

3. Independent claims — plain-language overview

⚠️ Important uncertainty flag: the full text supplied for US 5,974,164 is truncated before the claims. I therefore do not have verbatim claim language for the '164 patent, and the issued claim count is not in the retrieved text. The overview below is drawn from the patent's own summary-of-the-invention paragraphs in the specification (which are authoritative for what the claims cover) — not from claim text.

The specification discloses seven separately summarized inventions, each consistent with at least one independent claim:

1. Intensity-ratio base calling. A computer system identifies an unknown base by inputting multiple probe intensities, each associated with a nucleic acid probe, comparing the intensities to each other, and calling the unknown base from the comparison. The summary expressly recites that "a higher probe intensity is compared to a lower probe intensity" — the ratio-comparison core (FIGS. 9, 19).
2. Reference method. Probe intensities of a sample sequence are compared to probe intensities of a reference sequence (two implementations: per-base ratios of ratios, FIGS. 10A–10B; and neighboring-position normalized-difference analysis, FIGS. 11A–11D).
3. Statistical method. Probe intensities of a sample sequence are compared to statistics (mean/standard deviation) about probe intensities of a reference sequence across multiple experiments (FIG. 12).
4. Pooling processing. A method of processing reference and sample sequences together to reduce experiment-to-experiment variation: label each with distinguishable fluorescent (or otherwise distinguishable) markers, combine, fragment, hybridize to one chip, and scan both channels (FIG. 13).
5. Mutation identification using hybridization data from more than one base position. Compare a first set of probe intensities (reference) to a second set (sample), select hybridization regions where the two differ, and identify mutations from characteristics of those selected "dark regions"/bubble regions — including bubble-width-vs-probe-length logic and interrogation-position localization (FIGS. 14A–15).
6. Comparative analysis and visualization of multiple sequences. Display reference sequences in one area and sample sequences in another, with linking, comparison, and intensity-graph windows (FIGS. 16–18).
7. Base-calling/mutation-calling outputs including ambiguity codes (IUPAC-style), confidence codes, and false-positive suppression.

Cross-check (not a substitute): the closely related continuation US 6,733,964 (same title, Chee/Wang/Jevons/Bernhart/Lipshutz, Affymetrix, 43 claims) has claim 1: "A computer implemented method of identifying an unknown base in a sample nucleic acid sequence … inputting a plurality of probe intensities for an array of nucleic acid probes that differ from each other by a single base at an internal interrogation position … comparing said plurality of probe intensities to each other, wherein said comparing step includes … calculating a ratio of a higher probe intensity to a lower probe intensity; and … generating a base call …" This is consistent with invention #1 above, but it is the '964 claim, not verified '164 claim text.

4. Litigation / post-grant proceedings found

• Interference No. 104,552 — declared between US 5,795,716 and US 5,974,164 (Affymetrix) and Hyseq Application No. 09/358,875. Resolved by an Interference Settlement Agreement effective October 24, 2001, under which Hyseq abandoned the contest (to be filed on or before November 15, 2001) and the parties exchanged covenants regarding future interferences. Source: Law Insider copy of the Settlement of Interference Agreement (Hyseq, Inc. and Affymetrix, Inc., Oct. 24, 2001). This confirms the '164 patent was the subject of a priority contest, not merely cited art.
• No 2026 CAFC appeal, IPR, or district-court docket naming 5,974,164 was returned in my searches. This is unsurprising: the patent expired 2014-10-21 (anticipated expiration per Google Patents), so a live 2026 CAFC appeal on this number would be unusual. I cannot confirm or rule out a 2026 docket; I did not query PACER, the CAFC docket, or USPTO PatentCenter directly.
• For context, the large Affymetrix–Illumina and Affymetrix–Oxford Gene Technology litigations of 2000–2009 involved different Affymetrix patents (e.g., the six-patent Delaware suit and the OGT Southern patent issued Dec. 23, 1997). I found no source placing 5,974,164 among those asserted patents-in-suit; do not assume it was.

5. Family / citing-art context (literal identifiers as retrieved)

• Family members sharing the exact title, per the Google Patents priority links: US 6,600,996 B2 (from 08/828,952, 1997-03-28), US 6,242,180 B1 (from 09/158,765, 1998-09-23), US 6,607,887 B2 (from 09/796,071, 2001-02-27), and US 20030220748A1 / US 20040175718A1.
• The '164 patent is widely cited as prior art in later array-analysis patents (e.g., EP 2,851,091 A1 [0125]; EP 2,593,125 B1 [0324]; EP 2,566,492 B1 [0321]; EP 2,563,383 B1 [0319]; EP 2,327,796 A1 [0118][0129]) and in Affymetrix product literature patent lists (e.g., "5,733,729; 5,795,716; 5,974,164; 6,066,454; 6,090,555…").
• EU/Foreign equivalents: the record shows no EP/WO equivalent for this specific filing in the retrieved data; the family link on Google Patents relates to litigation family ID 23276902.

6. Explicit limitations on this answer

• I did not retrieve the granted claims of 5,974,164; §3 is an inference from the specification's own summary paragraphs and is not verified claim text. Anyone needing claim scope must pull the granted claims from the USPTO full-text record.
• I could not verify a CAFC 2026 docket for this number; the search tools available returned no such record.
• Inventor lists differ between the Google Patents bibliographic field (Chee only) and the recorded assignment (five assignors); I report both literally rather than reconciling them.

Primary sources used: https://patents.google.com/patent/US5974164/en · https://www.osti.gov/biblio/872618 · https://www.osti.gov/biblio/871785 · https://webapp1.dlib.indiana.edu/virtual_disk_library/index.cgi/[5628977](/patent/5628977)/FID3294/OG/html/1282-2/us06733964-20040511.html · https://www.lawinsider.com/contracts/14u1Z4ttG4z

I'll search for litigation involving US patent 5974164.

Let me search more specifically for litigation databases and Affymetrix enforcement campaigns.

Bottom line

Based on the sources I could reach (Google Patents, Justia, CourtListener/RECAP, CourtListener dockets, and general litigation-database searches), I could not identify any U.S. district court or Federal Circuit litigation in which US Patent 5,974,164 was asserted as a patent‑in‑suit. I want to be explicit about that rather than fill the gap with inferences from Affymetrix's other, well‑documented campaigns.

One caveat you should weigh: the Google Patents bibliographic page for US 5,974,164 carries the flag "Family has litigation — First worldwide family litigation filed," with a link into Darts‑ip (https://patents.darts-ip.com/?family=23276902...). That flag is family‑level, not patent‑level, and I was unable to retrieve the underlying case record (Darts‑ip is a licensed database). So it is evidence that some member of the '164 family has been litigated somewhere, but it does not establish that US 5,974,164 itself was asserted, and I could not obtain the case caption, court, or number. I am not able to confirm the family's litigation membership.

The one proceeding I did find that directly names the '164 patent

This is an administrative interference proceeding, not litigation — I list it because it is the only contested proceeding I could verify that names US 5,974,164 specifically:

• Proceeding: PTO Interference No. 104,552 (Interference No. 104,552)
• Parties: Affymetrix, Inc. (U.S. Patent Nos. 5,795,716 and 5,974,164) v. Hyseq, Inc. (Application No. 09/358,875)
• Jurisdiction: United States Patent and Trademark Office (not a court)
• Case number: Interference No. 104,552
• Filing/declaration date: Not established by the sources I retrieved
• Outcome: Settled. The parties executed an "Interference Settlement Agreement" effective October 24, 2001, resolving the interference and related litigation, with Affymetrix obligated to file the agreement with the PTO under 35 U.S.C. § 135.
• Source: the settlement agreement text as indexed at lawinsider.com ("Settlement of Interference Agreement… between Hyseq, Inc. … and Affymetrix, Inc.").

Note the agreement recites that the parties had also "resolved and settled the Litigation" — but the indexed excerpt does not identify the caption, court, or case number of that litigation, and I could not confirm that the '164 patent was asserted in it (the interference involved both the '716 and '164 patents, so the companion litigation may have concerned either or both).

Affymetrix enforcement campaigns I checked and could NOT tie to the '164 patent

I examined these because they are the plausible vehicles, but none of the sources I retrieved identify US 5,974,164 among the patents‑in‑suit:

• Affymetrix, Inc. v. Illumina, Inc., No. 1:04‑cv‑01231 (D. Del.), filed July 26, 2004 — six Affymetrix microarray patents asserted (per Illumina's 10‑K and Kirkland & Ellis coverage); the sources describe the patents generally and do not list the '164 patent. Outcome: jury verdict of infringement March 13, 2007; case later resolved by settlement/consolidated appeals.
• Affymetrix, Inc. v. PE Corp. (NY), 219 F. Supp. 2d 390 (S.D.N.Y. 2002) — patents‑in‑suit were the Carruthers patents (4,458,066; 4,500,707; 5,132,418; 5,158,319; 4,978,679), not the '164 patent. Affymetrix was the declaratory‑judgment plaintiff.
• E8 Pharmaceuticals LLC v. Affymetrix, Inc., No. 1:08‑cv‑11132 (D. Mass.), filed July 1, 2008 — MIT patent asserted against Affymetrix; not the '164 patent.
• Affymetrix, Inc. v. Illumina, Inc., No. 3:11‑cv‑00184 (W.D. Wis.), filed March 14, 2011 (docket listing located) — I could not confirm the patents asserted.

What I could not do, and why

• I did not have direct PACER or CourtListener full‑text search access to run the patent number against docket text; my searches returned only secondary hits (patent citations and patent‑family citations to US 5,974,164, which are not litigation).
• The Unified Patents litigation caselist and Darts‑ip entries could not be queried directly for this number.
• US 5,974,164 expired on its anticipated expiration date of October 21, 2014 (per the Google Patents legal‑status record: "Expired – Lifetime"), which is consistent with any enforcement having occurred, if at all, before that date — but expiry is not evidence of litigation.

Recommendation: to definitively confirm or rule out litigation, run 5,974,164 and 5974164 as a full‑text query in PACER's Case Locator / Docket Alarm / Lex Machina and check Darts‑ip family 23276902 directly. If you tell me the family‑level Darts‑ip hit is what matters, I can help you reverse‑engineer which family member was litigated — but I will not assert that the '164 patent itself was asserted without a caption to point to.

I'll verify the structured-data finding with targeted searches before answering.

The searches are returning a lot of different patents that share the "'164" shorthand. Let me isolate the exact 5,974,164 number.

Two confirming searches to close the loop — one on PTAB docket databases, one on any Federal Circuit activity.

PTAB / AIA Trial Analysis — US 5,974,164

Proceedings overview

There are zero AIA trial proceedings on file for US 5,974,164 — no IPR, no PGR, no CBM — so the breakdown is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials, and the bottom-line defensive posture for a defendant is not that "the patent has survived and is hardened" but that the patent is a 1994-priority filing that expired 2014-10-21, meaning any demand letter citing 5,974,164 is citing a dead patent with no enforceable post-expiration damages window.

This aligns with (and does not contradict) the canonical structured block supplied in the prompt: "The USPTO ODP API returns no AIA trial proceedings for this patent as of the most recent ingest." My independent web searches surfaced no proceeding that the ODP has missed. That said, I did not query PTAB E2E, the PTAB Decisions database, or CourtListener's docket API directly in this session — my searches were web-level. Treat this as "not found," which here is a strong signal, not a verified-zero certification.

---

⚠️ Mandatory false-positive warning: the "'164" shorthand collision

Before anything else, this is the single most important practical point. Nearly every search hit containing "'164 patent" and "IPR" refers to a different, unrelated patent. Practitioners and even PTAB petitions routinely abbreviate patent numbers to the trailing three digits, and there are at least three heavily-litigated U.S. patents ending in 164 in the biotech/electronics space alone:

None of these is US 5,974,164. If a demand letter or a litigation hold notice describes "the '164 patent," verify the full seven-digit number in the caption before you plan any PTAB strategy. If the asserted patent is 8,681,164 or 9,207,164, there is a live PTAB track to work with; if it is truly 5,974,164, there is not.

---

Proceeding-by-proceeding

None to report. There are no IPR, PGR, or CBM proceeding numbers to list, and I will not invent any.

---

What did happen to this patent — and why it matters

Because there is no AIA trial record, the practitioner's job is to explain the near-miss and the adjacent history, so a defendant isn't blindsided.

1. The patent was never eligible for most AIA trial types — structurally, not accidentally.

2. The closest thing to a "PTAB" proceeding here was a pre-AIA interference — before the BPAI, not the PTAB. Per the previously generated section of this analysis: Interference No. 104,552, declared between US 5,795,716 and US 5,974,164 (Affymetrix) and Hyseq Application No. 09/358,875, terminated by an Interference Settlement Agreement effective 2001-10-24 under which Hyseq abandoned the contest (filing due on or before 2001-11-15) and the parties exchanged covenants regarding future interferences. Note the date: 2001 — eleven years before the AIA created IPR. This was an Interference before the Board of Patent Appeals and Interferences, an Article III-adjacent administrative priority contest, not an AIA trial. Do not describe it as an "IPR" in any filing.

3. No Federal Circuit appeal on this patent number. The major Affymetrix appellate activity of the era — the Illumina litigation (D. Del., filed 2004-07-26, jury verdict of infringement 2007-03-13, later Fed. Cir. proceedings on US 7,510,841 and US 7,612,020), the OGT dispute (settled, ~$62.5M reported), the Incyte/Synteni litigation, and the Hyseq litigation asserting US 5,744,305 / US 5,800,992 / US 5,795,716 — involves other Affymetrix patents. I found no source placing 5,974,164 among the patents-in-suit in those cases, and no Fed. Cir. docket bearing this number. My earlier section flagged this; it holds.

4. A separate false lead to avoid: an ex parte reexamination certificate confirming claims 1-17 exists for US 5,807,755 (Affymetrix/Multilyte dispute), not for 5,974,164. I found no reexamination certificate, request, or file history event for 5,974,164 in the retrieved record. Unverified either way, but nothing affirmatively suggests one exists.

---

Strategic summary

Claim status. There is no PTAB-ordered cancellation of any claim of 5,974,164, because there is no PTAB proceeding. Every claim that issued on 1999-10-26 stands as issued — which is a distinction that sounds like a win for the patent owner until you note the operative fact: the patent's term ran out on 2014-10-21 (Google Patents: "Expired - Lifetime"; anticipated expiration 2014-10-21). Google Patents notes that its legal-status field "is an assumption and is not a legal conclusion," but the arithmetic is not seriously contestable: a 1994-10-21 priority filing with the then-applicable 20-year term from earliest U.S. filing expires in October 2014. So there are no "surviving claims" to list in any meaningful enforcement sense. The claims are unadjudicated, un-canceled, and unenforceable against post-expiration conduct — a patent's exclusive rights run only during the term (35 U.S.C. § 154(a)(1)), and there is no post-expiration royalty absent a pre-expiration license or a Hatch-Waxman-style regulatory carve-out, neither of which is suggested here.

Estoppel landscape. § 315(e)(2) estoppel is inapplicable — estoppel attaches only upon issuance of a Final Written Decision in an IPR, and none issued. There is therefore no § 315(e)(2) bar on any petitioner or privy, and equally, no petitioner obtained the benefit of a claim-cancellation judgment. For a defendant being asserted against today, the relevant question isn't which prior-art grounds remain unestopped; it's whether any enforceable claim remains. It does not, for ongoing conduct.

Pattern signals. No repeat petitioner (no petitioner at all). No defensive aggregator — I found no evidence that Unified Patents or any similar entity ever challenged this patent; the '164's expiry predates the modern unified-aggregator IPR playbook's relevance to it. The patent owner never had to defend at the PTAB, so there is no signal about how aggressively Affymetrix (now part of Thermo Fisher) would litigate this particular patent family. The family's stated "Family has litigation" flag on Google Patents (Darts-IP family 23276902) points to district-court and interference activity, not AIA trials.

---

Recommended next steps

If you are a defendant and the demand letter cites US 5,974,164:

• Verbatim, the operative disposition is: no AIA trial was ever filed; no Final Written Decision exists; the patent expired 2014-10-21. There is no FWD to link to, and I will not manufacture a URL for a nonexistent decision. The defensible citation is the Google Patents legal-status record (https://patents.google.com/patent/US5974164/en, "Expired - Lifetime," anticipated expiration 2014-10-21) plus the absence of any ODP/PTAB record.
• Attack the demand letter on term, not on validity. Pre-2014 conduct may still have supported a damages claim if suit was timely filed (subject to § 286's six-year bar), so preserve that issue; post-2014 conduct cannot infringe.
• Run a full-number capture check of the pleadings. Given the collision documented above, there is a real chance the asserted patent is 8,681,164 or 9,207,164 — in which case you have live IPR options (including the Parthenon IPR2015-01503 line and the Ancestry/DNA Genotek IPR2024-01419 line respectively), and an entirely different analysis applies.

If you are defending against a live patent in this family: the continuation family members are where the PTAB and district-court action actually is — US 6,600,996 B2 (from 08/828,952), US 6,242,180 B1 (from 09/158,765), US 6,607,887 B2 (from 09/796,071), and publications US 2003/0220748 A1 and US 2004/0175718 A1 — plus related US 6,733,964. Pull the PTAB docket for each of those before forming a strategy; do not assume 5,974,164's clean PTAB record tells you anything about them, and do not assume it tells you anything about US 5,795,716 either.

If you need certainty beyond this analysis: query PTAB E2E / the PTAB Decisions database directly for both the full number 5,974,164 and application 08/531,137, and separately confirm the expiration computation against the face of the printed patent (which will state the term and any terminal disclaimer). My searches here were web-level and could not substitute for a docket-level sweep.

If no PTAB activity exists — and here it does not: say so plainly in your own analysis. The absence is itself the signal. A patent that was never IPR'd, in a family that was actively litigated across four adversaries (Hyseq, Incyte/Synteni, OGT, Illumina) and that expired in 2014, was never a worthwhile PTAB target. That is a fact about the patent's commercial and legal relevance, and it belongs in your risk memo.

I'll research the assignment record for US 5,974,164, including the USPTO Assignment Center records and any NPE-related context.

Let me pull the INPADOC legal events for the '164 patent specifically and check for reel/frame numbers.

Let me confirm the Affymetrix corporate chain (Thermo Fisher acquisition) and check for any assertion history.

Ownership Chain Reconstruction — US 5,974,164

Important sourcing caveat up front. I could not open the USPTO Assignment Center detail page for 5,974,164 in this session, so I do not have the recorded correspondent-of-record (the filing attorney/agent) for any link in this chain. The reel/frame numbers below are reported where they are corroborated by the identical-date events on sibling Affymetrix patents (see notes), and are marked as such. Nothing below is inferred from naming alone.

---

Inventors

Pattern note (important discrepancy, carried over from the prior section): the Google Patents bibliographic block names only Mark S. Chee for application 08/531,137, while the recorded 1996-03-11 assignment lists five assignors. Two readings are possible: (a) the assignment covers a batch of Affymax applications rather than 08/531,137 alone (common for 1990s-era employment assignments recorded under one reel/frame), or (b) the Google bibliographic inventor field is incomplete. I report both literally rather than reconciling them. This is a documentation oddity, not an ownership risk.

Departure/fire-sale signal: not present / not determinable. I found no evidence of any inventor leaving Affymax/Affymetrix within 12 months of the 1995-10-16 filing, and no record of a portfolio sale tied to inventor departures. I could not verify departures either way — treat as not found, not as "verified none."

---

Original assignee

• Entity on the issued patent (Google Patents "Original Assignee"): Affymetrix, Inc.
• First recorded assignee of the inventors' rights: Affymax Technologies N.V. (a Netherlands Antilles corporation) — the venturing parent from which Affymetrix was spun out. This is why the "original assignee" field and the first assignment link differ.
• Primary line of business: DNA microarray ("GeneChip") instrumentation, arrays, and analysis software. The '164 patent covers the computer-aided visualization and analysis ("ViewSeq"-style) side of that stack, i.e., the software used to call bases from scanned array images.
• Did it ship a product embodying the claims? Yes at the corporate level. Affymetrix commercialized the GeneChip platform and its accompanying analysis software, and this patent is listed in Affymetrix product-literature patent lists (e.g., alongside 5,733,729; 5,795,716; 6,066,454; 6,090,555). I did not independently confirm a specific shipped software release maps to claim 1 of the '164, so treat "product embodiment" as strongly indicated but not claim-charted.
• Current status: Operating, as a wholly-owned subsidiary of Thermo Fisher Scientific. Thermo Fisher completed its ~$1.3 billion acquisition of Affymetrix on 2016-03-31 (after a bidding contest with Origin Technologies). Affymetrix was not dissolved and was not in bankruptcy; its stock ceased trading on NASDAQ. The '164 patent itself had already expired (anticipated 2014-10-21) before that acquisition.

---

Assignment timeline

The Google Patents legal-events feed shows six ownership/security events. The 2012 and 2015 reel/frames below are corroborated across the Affymetrix portfolio; the 1996 and 2001 reel/frames are not retrieved for this patent and are left blank rather than guessed.

• 1995-10-16 (executed) / filed 1995-10-16 — Application 08/531,137 filed; initial applicant recorded as Affymetrix, Inc. (no separate reel on the filing event)

  • Conveyance: (initial filing — not an assignment)
  • Assignor: n/a
  • Assignee: Affymetrix, Inc.
  • Correspondent: not retrieved.
  • Context: initial filing of the application; priority claim back to 1994-10-21.

• ~1995 (executed) / recorded 1996-03-11 — Reel not retrieved

  • Conveyance: Assignment (inventor → company employment assignment)
  • Assignor: Jevons, Luis C.; Wang, Chunwei; Bernhart, Derek H.; Chee, Mark S.; Lipshutz, Robert J.
  • Assignee: Affymax Technologies N.V. (Netherlands Antilles)
  • Correspondent: not retrieved. I cannot flag recurrence without this field.
  • Context: original employment assignment of inventor rights to the Affymax/Affymetrix parent entity.

• ~1996 (executed) / recorded 1996-08-26 — Reel not retrieved (a sibling Affymetrix patent, US 5,510,270, records the equivalent Affymax→Affymetrix transfer at 008955/0530; not confirmed for the '164)

  • Conveyance: Assignment
  • Assignor: Affymax Technologies N.V.
  • Assignee: Affymetrix, Inc.
  • Correspondent: not retrieved.
  • Context: internal reorganization — the spun-out operating subsidiary (Affymetrix) takes the portfolio from the parent.

• ~2001 (executed) / recorded 2001-05-04 — Reel not retrieved (the analogous merger on US 5,510,270 is at 009662/0497; not confirmed for the '164)

  • Conveyance: Merger
  • Assignor: Affymetrix, Inc., a California corporation
  • Assignee: Affymetrix, Inc., a Delaware corporation
  • Correspondent: not retrieved.
  • Context: change of domicile only — California→Delaware reincorporation; no change in beneficial ownership.

• 2012-06-25 (executed) / recorded 2012-06-27 — Reel 028465/0541 (reel/frame reported from identical-date blanket recordings on sibling Affymetrix patents; not read off the '164-specific page)

  • Conveyance: Security Agreement
  • Assignor: Affymetrix, Inc.
  • Assignee: General Electric Capital Corporation, as Agent (Maryland)
  • Correspondent: not retrieved.
  • Context: securitization / collateral lien — a portfolio-wide security interest securing Affymetrix debt, not a transfer of ownership.

• 2014-10-21 — Anticipated expiration (Google Patents legal status: "Expired – Lifetime")

  • Context: patent term ends; no assignment event.

• 2015-10-28 (executed) / recorded 2015-11-13 — Reel 037109/0132 (same reporting caveat as above)

  • Conveyance: Release by Secured Party
  • Assignor: General Electric Capital Corporation, as Agent
  • Assignee: Affymetrix, Inc.
  • Correspondent: not retrieved.
  • Context: release of the 2012 security interest — lien discharged; ownership had remained with Affymetrix throughout.

Not found: I did not locate any recorded assignment naming a Thermo Fisher entity, any NPE/licensing LLC, or any post-2015 assignee. The 2016 Thermo Fisher merger does not appear as a recorded assignment on the Google Patents legal-events feed for this patent, which is consistent with a stock merger effected at the corporate level (and with the patent having already expired in 2014).

---

Timeline diagram

timeline
    title Ownership of US 5974164
    1994 : Priority date 1994-10-21
    1995 : Application 08 531 137 filed
    1996 : Inventors assign to Affymax Technologies NV
         : Affymax assigns to Affymetrix Inc
    2001 : Merger into Affymetrix Delaware
    2012 : Security agreement to GE Capital
    2014 : Anticipated patent expiration
    2015 : GE Capital releases security interest
    2016 : Affymetrix acquired by Thermo Fisher

(Note: Mermaid timeline labels omit commas/slashes/parentheses as required; the dates above are rendered as plain hyphen-free tokens intentionally.)

---

NPE / troll-pattern signals

---

Verdict

Operating-company assertion.

Justification (2–3 sentences). Every recorded link in this chain stays inside a single operating-company family — Affymax Technologies N.V. → Affymetrix, Inc. (1996-08-26) → Affymetrix, Inc., Delaware (merger, recorded 2001-05-04) — and the only non-operating entity ever appearing is General Electric Capital Corporation as collateral agent (reel 028465/0541, recorded 2012-06-27), whose interest was fully released at reel 037109/0132, recorded 2015-11-13. There is no shell-entity transfer, known NPE assignee, cascading LLC chain, bankruptcy sale, privateering, or defensive-aggregator termination; the correspondents needed to test signal #3 were not retrieved, so that single signal remains unclear rather than negative. Note the qualification: the '164 itself was not located as an asserted patent-in-suit, and it expired 2014-10-21, so this is an operating-company-held patent, not a demonstrated operating-company assertion of this specific number.

Verify at: https://assignmentcenter.uspto.gov/ (search patent number 5974164), with the Google Patents legal-events mirror at https://patents.google.com/patent/US5974164/en.

---

Explicit limitations

1. I did not read reel/frame or correspondent fields off the '164-specific Assignment Center page. Reels 028465/0541 and 037109/0132 are corroborated by identical-date blanket recordings on sibling Affymetrix patents (Dimensions/Espacenet records for US 6,715,500, US 8,628,918, US 6,410,229, US 5,510,270) — they are highly likely but not confirmed for '164.
2. The 1996 and 2001 reel/frames are blank; the values shown in the table for those years (008955/0530 and 009662/0497) belong to US 5,510,270 and must not be attributed to the '164.
3. No 2016 Thermo Fisher assignment record was located for this patent; the merger is evidenced by Thermo Fisher's 2016-03-31 closing press release, not by an assignment reel.
4. Inventor-list discrepancy (Chee alone vs. five assignors) is reported literally and unresolved.

Sources: https://patents.google.com/patent/US5974164/en · https://app.dimensions.ai/details/patent/US-[6715500](/patent/6715500)-B2 · https://pt.espacenet.com/publicationDetails/inpadoc?CC=US&NR=[5510270A](/patent/5510270A) · https://pt.espacenet.com/publicationDetails/inpadoc?locale=pt_PT&CC=US&NR=[6410229B1](/patent/6410229B1) · https://thermofisher.mediaroom.com/2016-03-31-Thermo-Fisher-Scientific-Completes-Acquisition-of-Affymetrix

I'll search for the specific patent's citation record and claim text. Let me run several targeted queries in parallel.

Let me pull the full Espacenet cited-document list and attempt to retrieve the granted claims.

Method note (read first)

What I actually queried: open web search hitting the USPTO-facing databases from the outside — Google Patents (patents.google.com/patent/US5974164), Espacenet's "cited documents" view for US5974164A, EPO Global Patent Index, and PlainSite's USPTO record for application 08/531,137. I could not open USPTO PatentCenter / USPTO Full-Text (PatFT) directly, and my search budget was exhausted mid-task. Everything below is sourced from those mirrors of the USPTO/EPO record, stated literally.

Two hard constraints you need to know before relying on the §102 column:

1. The granted claims of US 5,974,164 were not retrieved. The supplied full text is truncated before the claims; the Google Patents page content returned by search is likewise truncated. So I cannot truthfully say "reference X anticipates claim 7." Any claim mapping below is conditional and keyed to the invention families the patent's own specification summarizes (per the previously-generated section), not to verified claim language.
2. The cited-reference list is only partially retrieved. Espacenet reports 20 documents cited in the search report; the returned text cut off at item 9. Items 10–20 were not retrieved. I will not guess them.

---

1. Confirmed identity (patent number used literally: 5,974,164)

✅ This resolves a caveat flagged in the prior section. The earlier section said the §120/priority relationship between 08/531,137 and 08/327,525 "could not be confirmed." Espacenet's bibliographic record for US 5,974,164 lists the priority document as US 08/327,525 (1994-10-21) — i.e., the '164 claims benefit of the very application that issued as US 5,795,716. Commit that.

⚠️ Contradiction to flag against the prior section's §5. That section stated: "the record shows no EP/WO equivalent for this specific filing." That is now contradicted. EPO Global Patent Index shows EP 0 717 113 A3 (application 95307476, filed 1995-10-20, priority US 32752594 / 1994-10-21), whose abstract is tagged "[origin: US5974164A]" and whose DOCDB simple family expressly lists US 53113795 A 19951016 — the '164 application — alongside US 32752594 A 19941021, US 15876598, US 4980598, US 79137304, US 79607101. EP 0717113 is a family member of the '164. This matters a lot for the prior-art question, because its A3 search report publishes the examiner's own X/Y relevance ratings for this family.

---

2. The prior-art citations — what the record actually shows

(A) Search-report citations retrieved for US5974164 (Espacenet, "patentes citadas no relatório de pesquisa" — 20 total, items 1–9 only)

On item 9: the title matches US 5,202,231 (Drmanac & Drmanac), and I hold that association with moderate-to-high confidence, but I did not confirm the number, inventor or date in this session — verify before citing it.

(B) The examiner's relevance ratings for this family — sibling EP 0717113 A3 search report (published 1996-07-17)

This is the single most probative document I found, because it is the same family and shows how the examiner weighted art, not just what was listed:

⚠️ I could not retrieve WO 95/00530's title or international filing date in this session. My recollection is that it is the PCT publication of the Affymetrix/Affymax "arrays of nucleic acid probes on biological chips" work (priority 1993-06-25), but I am flagging that as unverified rather than asserting it.

(C) Also-pertinent art expressly incorporated into the '164 specification

WO 90/15070 and WO 92/10092 are incorporated by reference in the '164 specification, as is US 5,143,854. Note the legal nuance: material incorporated by reference becomes part of that patent's own disclosure, so it is not automatically "prior art to" the claims it was incorporated to support — but those documents independently qualify as §102(a)/(b) art against other claims.

---

3. §102 analysis — pre-AIA applies

The application was filed 1995-10-16 with 1994-10-21 priority, so pre-AIA 35 U.S.C. §§ 102/103 govern. Critical date for §102(a) is the 1994-10-21 priority date; for §102(b), one year before the 1995-10-16 filing = 1994-10-16.

Two doctrinal points worth stating for the record:

• "Applicant's own" earlier Affymax work still counts. Pre-AIA §102(a) and §102(b) as applied to printed publications and patents contain no "by others" requirement. US 5,143,854, WO 92/10092 and WO 92/10588 are Affymax/Affymetrix-family documents and are still available as prior art — this is the classic corporate double-patenting/self-collision problem, and it is plainly why the specification incorporates them.
• Claim-by-claim §102 anticipation is not available to me. I am therefore giving you, below, the conditional mapping.

Conditional claim mapping (⚠️ keyed to specification-summarized inventions, not verified claims)

Based on the seven invention families set out in the prior section:

Ranking — "most relevant prior art"

1. WO 92/10588 (Affymax, 1992-06-25) — closest to the core base-calling/hybridization-sequence-determination subject matter.
2. Item 9 / US 5,202,231 (SBH genomes) — same axis, pending verification of the number.
3. US 5,002,867 (Macevicz, 1991-03-26) — multiple-probe sequence determination.
4. WO 89/10977 (Southern, 1989-11-16) — array-hybridization sequence analysis (examiner treated it as [Y], not [X]).
5. US 4,980,827 (Hara, 1990-12-25) — signal-processing-to-base-call.
6. Pease 1994 / Southern 1992 — §103 fuel, and §102(b) art in their own right.

---

4. Explicit limitations on this answer

• I did not retrieve 11 of the 20 search-report citations (items 10–20). This list is incomplete by construction.
• I did not verify item 9's patent number. Treat "US 5,202,231" as a lead, not a citation.
• I did not retrieve the granted claims of US 5,974,164. The "which claims" column is therefore conditional and mapped to spec-summarized inventions. Anyone needing a real §102 chart must pull the granted claims from USPTO PatentCenter / PatFT.
• I did not open USPTO PatentCenter or PatFT. The authoritative front-page "(56) References Cited" list for the '164 remains unverified against these mirrors.
• WO 95/00530's title and filing date are unverified in this session.
• Tangent from search results that is not this patent: I hit a large volume of later Affymetrix product-literature patent lists (e.g. 5,733,729; 5,795,716; 5,974,164; 6,066,454; 6,090,555… in the GeneChip Expression Analysis Technical Manual and numerous later patents). These are forward citations of the '164, not prior art against it — do not confuse the two directions.

Primary sources used: https://pt.espacenet.com/publicationDetails/citedDocuments?CC=US&NR=[5974164A](/patent/5974164A) · https://pt.espacenet.com/publicationDetails/biblio?CC=US&NR=5974164A · https://patents.google.com/patent/US5974164/en · http://data.epo.org/gpi/EP0717113A3-UNTITLED.html · https://www.plainsite.org/courts/united-states-patent-and-trademark-office/computer-aided-visualization-and-analysis-system-for-sequence-evaluation/1l36dezp4/

I'll first try to pull the prior-art/citations section for this specific patent, since the fetched page text was truncated before the claims and the citation list.

Let me try to retrieve the actual cited-references ("Prior Art") list and the granted claims for this exact number.

Important scope flag before the analysis

The Google Patents page text supplied for US 5,974,164 is truncated before the claims and does not contain the "Citations / Prior Art" tables (the (56) References Cited block and the NPL list). I therefore do not have (i) the verbatim granted claims of the '164 patent, or (ii) the examiner's own citation list for the '164.

To avoid fabricating either, I did the following and flag it as a proxy, not a substitute:

• Retrieved the citation lists that are on the family record — PubChem's citation list for sibling US 6,242,180 B1 (same priority 1994-10-21, same specification) and the Justia "Referenced Cited" list for sibling US 6,607,887, plus the related-literature list on OSTI ID 872618 (the DOE record for US 5,974,164 itself).
• Used sibling claim text (US 6,242,180 claim 1; US 6,733,964 claim 1; US 6,600,996 claims 13–19) only to illustrate the disclosure's inventions. These are not the '164 claims.

Any conclusion below should be re-run against the actual (56) block on the '164 front page and the granted claims before it is relied on.

---

1. Framework and the date problem that drives everything

Governing law: the '164 has an effective filing date no later than 1995-10-16 (priority 1994-10-21), so pre-AIA 35 U.S.C. §103 applies, as construed by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). Obviousness must be proven by clear and convincing evidence (Amgen v. Hospira), with the Graham v. John Deere factors. Affymetrix is a sophisticated, well-funded, well-counseled patentee. There is no IPR/PGR on record for this number (and none is now available — the patent expired 2014-10-21, Google Patents "Expired – Lifetime"), so any §103 attack is a district-court invalidity defense or DJ claim.

Critical date bifurcation (CIP risk). Espacenet lists two priorities for the '164: US19940327525 (1994-10-21) and US19950531137 (1995-10-16). Application 08/327,525 issued as US 5,795,716 on 1998-08-18, i.e., the parent was still pending when 08/531,137 was filed on 1995-10-16 — so the '164 is a continuation or continuation-in-part, and the record I retrieved does not say which.

This matters enormously for §103:

• Claims supported only by matter new in the 1995 filing get the 1995-10-16 date. Everything published between 1994-10-21 and 1995-10-16 then becomes §102(a)/(b)/(e) art against them.
• That window contains a lot: Guo et al., Nucleic Acids Res. 22(24):5456–5465 (published Dec. 11, 1994); WO 95/11995 (published May 4, 1995); Fodor et al., US 5,424,186 (issued July 11, 1995); Lipshutz et al., BioTechniques 19:442 (1995); Chee et al., ASHG abstract (Oct. 18–22, 1994); Chee et al., Genome Sequencing & Analysis Conf. V abstract (Sept. 1994).
• Conversely, Schena et al., Science 270:467 (published Oct. 20, 1995) is four days after the filing date and is not §102(a) art to the '164 — a common error worth avoiding.

A second, decisive date trap — §103(c). Pre-AIA §103(c) disqualifies, for §103 purposes only, art that qualifies solely under §102(e)/(f)/(g) and was commonly owned at the time the invention was made. Affymetrix's own earlier applications — US 5,837,832 (Chee et al.), US 5,861,242 (Chee et al.), the WO 95/11995 filing (priority US 08/082,937, 1993-06-25), and U.S. application Ser. No. 08/249,188 (expressly incorporated by reference in the '164) — are commonly owned with the '164 and are therefore excluded from any §103 combination. A §103 case against the '164 must be built on third-party art with §102(a)/(b) status (patents, printed publications, public use). That is the single most important constraint on the analysis below, and it eliminates the most tempting "combine Affymetrix with Affymetrix" theories a casual reader would reach for.

---

2. The prior-art set that is actually usable (third-party, §102(a)/(b) status)

---

3. Combination-by-combination obviousness analysis

I organize by the seven invention groupings identified in the earlier section (which came from the '164's own summary paragraphs), because I cannot tie limitations to verified claim numbers.

Combination 1 — Intensity-ratio base calling (FIGS. 9, 19)

Primary: US 5,143,854 (or WO 92/10092 for the readout) + US 5,202,231 and US 5,002,867 (SBH algorithms) + Guo et al. 1994 (four-probe interrogation-position set, signal-ranked base call).

Why a PHOSITA would combine: Drmanac '231 and Macevicz '867 already claim the step of inferring a base from which probe in a set binds best. Guo et al. reduces that to practice on a planar glass array with four single-base-variant probes and a fluorescence reader. Pirrung '854/WO 92/10092 supply the array and the reader. Nothing in the ratio method (dividing a higher intensity by a lower one, comparing to a cutoff, emitting an IUPAC ambiguity code when the ratio is below cutoff) is more than automating a comparison that the SBH references perform by inspection, using a numerically explicit decision rule. KSR rationales (i) and (iii): combining known elements to yield predictable results; using a known technique to improve a similar device in the same way. The cutoff values themselves (1.2, background-difference cutoffs) are classic obvious-to-optimize subject matter — In re Aller, 220 F.2d 454, 458 (CCPA 1955); In re Boesch, 617 F.2d 272 (CCPA 1980).

Rebuttal to anticipate: patentee will argue the references disclose qualitative presence/absence of binding, not ratio comparisons of a higher to a lower intensity, and that the ratio method's advantage ("much less sensitive to cross-hybridization") is unexpected. Expect prosecution-style argument that the art "taught away" from quantitative intensity comparison because cross-hybridization was known to be unreliable. Counter: Southern/Maskos & Elder explicitly quantify and compare hybridization signals, so the unreliability argument cuts the other way.

Combination 2 — Reference-comparison method, implementation 1 and 2 (FIGS. 10A–11D)

Primary: Southern, Maskos & Elder, Genomics 13:1008 (1992) + Maskos & Southern, NAR 20:1679 (1992) + US 5,202,231.

Why a PHOSITA would combine: Southern/Maskos & Elder is literally titled "Analyzing and comparing nucleic acid sequences by hybridization to arrays of oligonucleotides" and does exactly the '164's reference method — hybridize a known reference and an unknown sample (or variants) to the same probe set and compare the signal patterns. Maskos & Southern establish that a single mismatch measurably depresses duplex yield. The additional '164 features of implementation 2 — normalizing each base intensity to the sum of the four probes at a position, subtracting the normalized reference from the normalized sample, and flagging a difference exceeding a threshold ("bubble" detection in FIG. 11C) — are routine data normalization and thresholding applied to the comparison Southern already makes. KSR rationale (v): where there is a design need (experiment-to-experiment intensity variation) and a finite number of predictable solutions (normalize by sum, by total, or by reference channel), the variation is obvious to try. This is also the Digitech/Perfect Web family of "compare data, apply threshold, report result" reasoning (Perfect Web Techs. v. InfoUSA, 587 F.3d 1324 (Fed. Cir. 2009)).

Combination 3 — Statistical method across multiple experiments (FIG. 12)

Primary: Southern/Maskos & Elder 1992 + Dear & Staden, NAR 19:3907 (1991) + the automated base-calling/confidence art (US 4,811,218; Hara/Shiraishi group).

Why a PHOSITA would combine: computing a mean and standard deviation of a ratio across replicate experiments and calling a base when a sample value falls outside a standard-deviation cutoff is textbook statistical quality control applied to a measurement the art already made. Dear & Staden and the automated-sequencing patents supply the motivation to attach confidence values to automated calls and to store/compare them — automated sequencers of the era (ABI-type fluorescence instruments, reflected in the Hara/Shiraishi/Hunkapiller citations) routinely produced quality/confidence estimates on called bases. KSR rationale (vi): market and design pressure to reduce miscalls drove practitioners to replication and statistics. The '996 sibling claims — computing a base call for each of several probe sets and then taking the call that occurs most often — are a plain majority-vote consensus, In re O'Farrell, 853 F.2d 894 (Fed. Cir. 1988) (predictability not required; obvious to try with a reasonable expectation of success).

Rebuttal to anticipate: patentee will argue that applying statistics across experiments to hybridization calls was non-routine in 1994 and that a "standard deviation cutoff per base" (localized cutoffs) is a specific, non-suggested refinement. Counter: per-base or per-category cutoffs are optimization of a statistical test and would be routine; the specification itself offers no criticality data for the per-base cutoff values.

Combination 4 — Pooling / two-color competitive hybridization on one chip (FIG. 13)

Primary: Kallioniemi et al., Science 258:818 (1992) (with US 5,665,549, Pinkel et al.) + US 5,143,854 / WO 92/10092 + US 5,202,231.

Why a PHOSITA would combine: CGH squarely teaches labeling test and reference nucleic acids with two spectrally distinguishable fluorophores, co-hybridizing them and ratioing the two channels to score differences — the exact logic of pooling. Array technology supplied by Pirrung/WO 92/10092 substitutes a high-density probe array for metaphase chromosomes, a substitution of one known hybridization platform for another with a predictable result. Both references come from the same community (cytogenetics/genome analysis) publishing in Science within months of each other, giving a strong "same field, same problem, recognized need" motivation under KSR rationale (iv) ("applying a known technique to a known device ready for improvement"). The '164's add-on (correcting for signal overlap between the two channels before analysis) is a calibration step an ordinary artisan would perform as a matter of course.

Rebuttal to anticipate: the assertion that CGH operated on a different scale/format and that dual-label correction on a miniaturized array was non-obvious. This is a §103 argument about the predictability of the substitution, and it is usually met with the point that the '164 itself describes the pooling step as an implementation convenience, not a surprising result.

Combination 5 — Mutation identification from more than one base position ("bubbles," false-positive suppression) (FIGS. 14A–15)

Primary: Southern/Maskos & Elder 1992 + Maskos & Southern 1992 (mismatch depresses yield; magnitude depends on mismatch type) + Guo et al. 1994 + automated base-calling/decoding art (Hara/Shiraishi group).

Why a PHOSITA would combine: the '164's insight is that a true point mutation produces a region of depressed signal roughly equal in width to the probe length, centered at the interrogation position, whereas a one-cell dropout (dust, artifact) does not. The underlying physics — every probe overlapping a mismatch hybridizes poorly; the number of overlapping mismatched probes equals the probe length — is inherent in the SBH/array art (Drmanac '231; Southern/Maskos & Elder) and was understood in 1994. Given that, measuring the width of a signal-depressed region and comparing it to the known probe length is routine data analysis of a known signal pattern. The false-positive-suppression feature (sibling US 6,242,180 claim 1 requires "changing the mutant base call to a nonmutant base call if the probe intensities of at least one position that is near the position of the suspected mutation are inconsistent with a mutation") is, in substance, the obvious corollary: a real mutation has neighboring corroboration; an artifact does not. KSR rationales (i) and (vii) — the reference itself implies the improvement; "a patent composed of several elements is not proved obvious merely by demonstrating that each of its elements was, independently, known," but here the pattern was known and the rule follows directly.

Rebuttal to anticipate: this is the strongest non-obviousness position in the family, because "bubble width ≈ probe length" is a specific empirical insight, and FIG. 14B's claim that bubble shape identifies the mutation (C-A vs. U-G mismatch) is a much thinner reading. Expect the patentee to argue secondary considerations here (see §4). Expect the challenger to argue that identifying a dark zone and applying a known base-calling algorithm within it is exactly the "efficient use of base calling methods" the specification describes and is a labor-saving arrangement of known steps.

Combination 6 — Comparative analysis / two-pane visualization GUI (FIGS. 16–18)

Primary: Dear & Staden, NAR 19:3907 (1991) + US 4,811,218 (Hara/Shiraishi/Hunkapiller automated sequence-reading displays) + the IntelliGenetics Suite 5.4 Advanced Training Manual (Jan. 1993) cited in the family record.

Why a PHOSITA would combine: displaying a reference sequence in one pane and a sample/consensus sequence in another, side by side, with differences highlighted and bases editable, plus a per-base intensity plot, are standard features of 1991–1993 sequence-assembly software. Dear & Staden's program is titled an "assembly and editing program"; IntelliGenetics is a commercial sequence-analysis suite. Arranging the two panes vertically, color-coding mismatches, and pulling up a graph of the underlying probe intensities for a selected base is a predictable, user-driven presentation of the same data, supported equally for §103 by KSR rationale (iii) and by the Federal Circuit's treatment of user-interface/data-presentation claims (e.g., the reasoning of Digitech, cautioned by Trading Techs. v. IBG only for §101). This grouping is the weakest from a §103-defense standpoint.

---

4. Secondary considerations (Graham factor 4)

• Commercial success: strong in gross terms (Affymetrix GeneChip franchise) but the nexus problem is severe — the products embody array manufacturing and assay technology claimed in other Affymetrix patents (e.g., the six-patent group asserted against Illumina in 2004–2007, per the DukeSpace case study), not the '164's data-analysis methods. A nexus to the comparative visualization and base-calling claims would need to be shown specifically.
• Long-felt need / failure of others: the need for automated resequencing-by-hybridization base calling was recognized and partially met by Guo (1994) and by the SBH groups (Drunac, Southern). That weakens "failure of others."
• Unexpected results: the only plausible candidate is the empirical "bubble width ≈ probe length" rule and the reference method's robustness to cross-hybridization. Note the irony: the '164's own text admits the ratio method fails where the reference method succeeds (the FIG. 20 comparison), which the patentee may use to argue synergy, but a challenger will characterize as two known alternatives, one of which the artisan would obviously select for noisy data.
• Licensing/industry acquiescence: the 2001 Hyseq interference settlement (family 23276902) shows the claims were regarded as commercially significant, but settlement of an interference is evidence of priority, not of non-obviousness.

---

5. Adjacent doctrines that may be more potent than §103

1. Obviousness-type double patenting (ODP). Six patents issued from this single 1994 disclosure — 5,795,716; 5,974,164; 6,242,180; 6,600,996; 6,607,887; 6,733,964 — with overlapping claims drawn to base calling and probe-intensity analysis. Absent terminal disclaimers, ODP (and, alternatively, statutory double patenting where claims are identical) is a live challenge. I found no terminal disclaimer in the retrieved data; this should be pulled from PatentCenter.
2. §101 (Alice/Mayo). Much of the specification claims comparing numerical values to cutoffs and displaying the results; the "computer program" claims (the Google page shows two "computer program" claims); two-pane display claims. These are candidates for the "mental process + generic computer" analysis, and the family's eventual shift to physical array/assay claims (later Chee patents) is telling.
3. §112 written description / enablement for the bubble-shape-fingerprints-mutation functionality, which the specification presents as hypothesis ("may be utilized to identify the particular mutation") without a working library.
4. The inventorship discrepancy already flagged (Google lists Chee alone; the 1996-03-11 assignment lists five assignors) is not §103 art but is a §102(f)/§116 or inequitable-conduct thread worth noting.

---

6. Bottom line

On the art I could actually verify, at least three claim groupings are strongly exposed to §103 attack:

• the intensity-ratio and reference-comparison base-calling methods (Combinations 1–2), which are numeric decision rules applied to comparisons the SBH and array-hybridization literature already made;
• the dual-label pooling method (Combination 4), which is CGH's two-channel ratio logic ported to a known array platform;
• the two-pane comparative visualization GUI (Combination 6), which mirrors 1991–1993 sequence-assembly software.

The "bubble" / multi-position mutation-localization and false-positive-suppression grouping (Combination 5) is the most defensible and should be expected to be the contested ground.

Whether the combination can actually reach the claims turns on two things I could not verify and that must be pulled before any invalidity chart is final: (a) the granted claims of the '164, and (b) whether the '164 is a continuation or a continuation-in-part of 08/327,525 — because if any claim depends on matter first added in the 1995-10-16 filing, Guo et al. (Dec. 11, 1994), WO 95/11995, Fodor US 5,424,186 and the Oct. 1994 Chee posters all enter the prior-art set, materially strengthening every combination above. Conversely, pre-AIA §103(c) removes Affymetrix's own earlier filings (US 5,837,832, US 5,861,242, WO 95/11995, and application Ser. No. 08/249,188) from any §103 combination — a point many analyses of this family get wrong.

---

Sources

• https://patents.google.com/patent/US5974164/en (bibliography, legal status, priority links; claims and citation tables not present in the retrieved text)
• https://pt.espacenet.com/publicationDetails/biblio?DB=EPODOC&II=22&ND=8&adjacent=true&locale=pt_PT&FT=D&date=19991026&CC=US&NR=[5974164A](/patent/5974164A)&KC=A (priority numbers US19940327525 / US19950531137)
• https://www.osti.gov/biblio/872618 (DOE record for US 5,974,164; related literature: Southern/Maskos/Elder 1992; Strezoska 1991; Drmanac 1991; Dear & Staden 1991; Drmanac 1993)
• https://pubchem.ncbi.nlm.nih.gov/patent/US-6242180-B1 (citation list on sibling US 6,242,180, same 1994-10-21 priority)
• https://patents.justia.com/patent/6607887 (Referenced Cited list on sibling US 6,607,887)
• https://patents.justia.com/patent/6242180 (US 6,242,180 claim 1 — sibling claim text used illustratively)
• https://patents.justia.com/patent/6600996 (US 6,600,996 claims 13–19 — sibling claim text)
• https://dukespace.lib.duke.edu/server/api/core/bitstreams/3b01a5ec-e3ba-4848-aff4-cd93f3604a37/content (Affymetrix–Illumina litigation history; patents-in-suit were not the '164)
• KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007); Graham v. John Deere Co., 383 U.S. 1 (1966); In re Aller, 220 F.2d 454 (CCPA 1955); In re Boesch, 617 F.2d 272 (CCPA 1980); In re O'Farrell, 853 F.2d 894 (Fed. Cir. 1988); Perfect Web Techs. v. InfoUSA, 587 F.3d 1324 (Fed. Cir. 2009); pre-AIA 35 U.S.C. §§ 102, 103(c).

Confidence statement: I have high confidence in the legal framework, the reference identities/dates listed as §102(b) art, and the §103(c) analysis. I have low confidence in any statement tying a specific reference to a specific claim number of the '164, because I could not retrieve the '164's granted claims or its own (56) citation block. Those two documents should be obtained from USPTO PatentCenter before this analysis is relied upon.