US 4683195

I'll search for the specific patent number and its litigation history.

Let me run more targeted searches on litigation history and current dockets.

Let me pull the full claim set and confirm the reexamination/expiration status.

US 4,683,195 ("US4683195") — Analyst Summary

Scope note on the identifier: I searched the literal string 4683195 (and its punctuated form 4,683,195). Per the operating rule, I did not substitute any similar number (e.g., the closely related 4,683,202, which is a separate Cetus PCR patent, appears frequently alongside it and is not the subject of this report). One search hit — Florida entity FEI 81-4683195 (Amitabha Realty, LLC) — is unrelated noise and was disregarded.

---

1. Bibliographic data (per the Google Patents record and Espacenet/INPADOC)

Priority chain from the specification itself: the application is a continuation-in-part of Ser. No. 824,044 filed Jan. 30, 1986 (now abandoned), which was a division of Ser. No. 791,308 filed Oct. 25, 1985, which was a CIP of Ser. No. 716,975 filed Mar. 28, 1985 (now abandoned). The earliest disclosed date in this chain is therefore March 28, 1985.

Classification: C12Q 1/68 and subclasses (C12Q 1/6844, 1/686 [PCR], 1/6876, 1/6858, 1/6881), Y10T 436/143333.

Assignment chain (INPADOC):

• 1986-03-03 — inventors → Cetus Corporation (Reel/Frame 004526/0934)
• 1992-01-17 — Cetus → Hoffmann-La Roche, Inc.
• 1997-01-27 — Hoffmann-La Roche → Roche Molecular Systems, Inc.
• 2005/2009 — later reassignments involving Bioniche Life Sciences / Laurus Master Fund appear in the same INPADOC feed; these look like portfolio/security filings rather than transfers of the PCR patent family, and I have not verified them.

---

2. Abstract (verbatim)

"The present invention is directed to a process for amplifying and detecting any target nucleic acid sequence contained in a nucleic acid or mixture thereof. The process comprises treating separate complementary strands of the nucleic acid with a molar excess of two oligonucleotide primers, extending the primers to form complementary primer extension products which act as templates for synthesizing the desired nucleic acid sequence, and detecting the sequence so amplified. The steps of the reaction may be carried out stepwise or simultaneously and can be repeated as often as desired. In addition, a specific nucleic acid sequence may be cloned into a vector by using primers to amplify the sequence, which contain restriction sites on their non-complementary ends, and a nucleic acid fragment may be prepared from an existing shorter fragment using the amplification process."

---

3. Independent claims — plain language

Claim 1 — the core "amplify-and-detect" (PCR + probe) process. A method of detecting whether a specific nucleic acid sequence is present in a sample (or of distinguishing two forms of a sequence), comprising:

• (a) treating the sample with one oligonucleotide primer per strand of each sequence being sought, under hybridizing conditions, so a primer extension product complementary to each strand is made — the primers being chosen so that the extension product of one primer, once separated from its complement, can itself template synthesis from the other primer;
• (b) denaturing to separate the extension products from their templates;
• (c) repeating primer extension using the separated single strands as templates, thereby amplifying the target;
• (d) adding a labeled probe that hybridizes to the amplified sequence (or a mutation of it); and
• (e) determining whether hybridization occurred.

Steps (a)–(c) may be run sequentially or simultaneously, and (b)–(c) may be cycled as often as needed (the specification's table shows exponential accumulation of the discrete target: 2ⁿ − n − 1 copies after n cycles, versus linear accumulation of "long products").

Claim 16 — allele-specific detection applied to sickle cell anemia. The same amplification scheme, but tied to the β-globin region containing the polymorphic MstII/DdeI site associated with sickle cell anemia, followed by hybridization with a labeled probe complementary to normal β-globin, digestion with the restriction enzyme for the sickle-specific site, and detection of a restriction fragment correlated with disease status — the "oligomer restriction"/RFLP-like read-out illustrated in Figs. 6–9.

Claims 2–15 and 17 are dependent: e.g., claim 2 (repeat b–c at least once), claim 4 (double-stranded input, denatured), claim 5 (single-stranded input), claims 6–9 (DNA or RNA target with oligodeoxyribonucleotide primers), claim 10 (restriction sites on primer 5′ ends, digestion, separation, then probing), claims 11–12 (deletions/mutations causing genetic disease; specifically sickle cell anemia), claim 13 (restriction digest + electrophoresis + Southern blot), claim 14 (pathogen or oncogene target), claim 15 (enzyme selected from E. coli DNA polymerase I, Klenow fragment, T4 DNA polymerase, reverse transcriptase, or a heat-stable enzyme active at ~65–90 °C).

Independent claims beyond 1 and 16 — uncertainty flagged. The specification discloses three further statutory embodiments: (i) a process for cloning the amplified sequence into a vector using primers bearing restriction sites on their 5′ ends; (ii) a process for synthesizing a longer fragment from a shorter "core" fragment using primers whose 5′ ends encode the appended sequences; and (iii) diagnostic kits (multicontainer unit with primers, polymerizing agent, four dNTPs, probe, and detection means). The reexamination record confirms that claims 19–25 were present and were reexamined. However, I was unable to retrieve verbatim claim text for claims 18–25 within my search budget, so I will not paraphrase them as if quoting. If the exact wording of those independent cloning/kit claims matters, the B1 reexamination certificate and the printed patent column should be pulled directly.

---

4. Prosecution / reexamination history

• Reexamination: Two requests for reexamination of the '195 patent were filed — Control No. 90/001902 (Dec. 6, 1989) and 90/001954 (Mar. 9, 1990); the second was a du Pont filing. The PTO granted reexamination and issued a decision on August 23, 1990 upholding the '202 and '195 patents in their entirety, rejecting the argument that PCR was anticipated by Kleppe et al. and Panet & Khorana. The reexamination certificate (B1) issued November 27, 1990, with the reexamination request effective date recorded as December 6, 1989.
• Other post-issue events in the INPADOC feed (fee payments, entity-status changes) are routine; the patent expired in 2005.

---

5. Litigation / docket search

No 2026 USPTO or CAFC docket activity involving US 4,683,195 was found. Searches of the CAFC 2026 materials and district-court filings returned no case naming this patent number. This is consistent with the patent having expired in 2005 — any enforcement proceeding would be time-barred for ongoing infringement, so live 2026 Federal Circuit activity on the '195 patent is not expected.

Historic enforcement activity the searches did surface (context only, not 2026):

• E.I. du Pont de Nemours & Co. v. Cetus Corp., N.D. Cal., declaratory judgment filed Aug. 1, 1989, challenging the validity of the '202 and '195 patents under §§ 102/103 based on the Khorana-era prior art. The district court upheld validity (reported as a Feb. 27, 1991 ruling in San Francisco), following the PTO's Aug. 23, 1990 reexamination decision.
• Hoffmann-La Roche, Inc. v. Promega Corp., N.D. Cal. (from Oct. 1992), Hoffmann-La Roche v. Promega, 323 F.3d 1354 (Fed. Cir. 2003) — this concerned the Taq polymerase patent ('818), not the '195 patent, but the opinions cite '195 as one of Roche's PCR patents.
• Molecular Diagnostics Laboratories v. Hoffmann-La Roche Inc., 402 F. Supp. 2d 276 (D.D.C. Dec. 1, 2005) — antitrust/Walker Process action; the court's footnotes list "Patents Nos. 4,683,195 and 4,683,202" among Roche's PCR patents.
• The Google Patents record carries a Darts-IP "Family has litigation" flag for family 27124781, which I could not open directly.

---

6. Explicit uncertainties

1. Verbatim text of claims 18–25 (independent cloning/kit/fragment-synthesis claims) is not verified here.
2. The scientific/technical effect statements (e.g., 1.4 fmol of 110-bp product at 70% per-cycle efficiency; >1000-fold sensitivity gain) are as reported in the patent's own Examples, not independently re-verified.
3. Assignment entries after 1997 are drawn from an INPADOC feed and may be unrelated portfolio filings; treat the Bioniche/Laurus entries with caution.
4. Date discrepancy: the system header gives today's date as 2026-09-27 while your instruction states April 26, 2026. I report both as noted; this does not affect any finding above, since no 2026 activity on the '195 patent was located under either date.
5. The title is rendered both as "and/or-cloning" (patent face, with hyphen) and "and/or cloning" (some indices); I have quoted the patent face version literally.

I'll search for litigation records involving US 4,683,195.

Let me dig deeper into specific cases involving this patent number.

Bottom line

US 4,683,195 (Mullis et al., "Process for amplifying, detecting, and/or-cloning nucleic acid sequences," Cetus Corp., filed 1986-02-07, issued 1987-07-28) is one of the two foundational PCR process patents — the co-filed sibling being US 4,683,202. It appears in the litigation record almost always as part of the PCR process patent family, typically asserted or challenged alongside the '202 patent, and sometimes alongside US 4,889,818 (the Taq polymerase patent). Legal status today: Expired – Lifetime (anticipated expiration 2005; Google Patents lists 2005-03-28 anticipated expiration).

Below is what I could actually ground in sources, with confidence levels flagged. Note: Google Patents' own record for the patent carries a Darts-IP "Family has litigation" link (family 27124781, https://patents.darts-ip.com/?family=27124781), but I could not retrieve the underlying case list from that source, so the entries below come from court opinions, secondary legal literature, and contemporaneous reporting.

---

Cases identifying or naming the '195 patent

1. E.I. du Pont de Nemours & Co. v. Cetus Corp. (declaratory judgment — validity challenge)

• Plaintiff: E.I. du Pont de Nemours & Co. (Wilmington, DE)
• Defendant: Cetus Corporation (Emeryville, CA)
• Jurisdiction: U.S. District Court for the Northern District of California (San Francisco)
• Case number: Not confirmed in my sources
• Filing date: on or about August 1, 1990 (DuPont filed a declaratory judgment action after threatening Cetus)
• Patents at issue: US 4,683,195 and US 4,683,202 — expressly named in contemporaneous reporting
• Outcome: On February 27, 1991, the federal district court in San Francisco ruled in Cetus's favor — the PCR patents were held valid and DuPont was held not to infringe. DuPont reportedly had not appealed as of the reporting date.
• Source: Chinese-language report "Cetus赢得PCR专利战的胜利," Biotechnology Bulletin (生物技术通报) 1991(9):19, which states DuPont argued the 1987 US PCR patents "第4683195和4683202号" should not have issued to Cetus, and that the San Francisco federal district court upheld the patents on Feb. 27, 1991 (https://biotech.aiijournal.com/CN/abstract/abstract2052.shtml).
• Confidence that '195 was at issue: High. The '195 number is explicitly named. Confidence in case number: Low — I did not find one and will not guess.

2. Molecular Diagnostics Laboratories v. Hoffmann-La Roche Inc. (antitrust / Walker Process — '195 named in the record)

• Plaintiff: Molecular Diagnostics Laboratories (direct-purchaser class)
• Defendants: Hoffmann-La Roche Inc. and Applera Corporation (successor to PE/Perkin-Elmer)
• Jurisdiction: U.S. District Court for the District of Columbia
• Case number: Not confirmed in my sources
• Filing date: Complaint filed on or about September 23, 2003 (class period ran from Sept. 23, 2000); the Dec. 1, 2005 opinion denied Rule 12(b)(6)/9(b) dismissal motions
• Patent actually asserted/central: US 4,889,818 (Taq), not '195. However, the court's opinion contains an express record statement: "Roche holds other patents related to the PCR process, among them Patents Nos. 4,683,195 and 4,683,202." (M.D.L. v. Hoffmann-La Roche, D.D.C., Dec. 1, 2005, footnote 2)
• Outcome/status: Motion to dismiss denied (2005); the case later settled on September 26, 2008, with defendants paying $33 million to the direct-purchaser class; approved Dec. 19, 2008, including ~$11 million in attorney's fees.
• Sources: https://www.courtlistener.com/opinion/[2426221](/patent/2426221)/molecular-diagnostics-laboratories-v-hoffmann-la-roche-inc/ ; https://scholarworks.law.ubalt.edu/cgi/viewcontent.cgi?article=1746&context=all_fac
• Confidence: High that '195 is named in the litigation record; this case was not an infringement action on '195 — it was a Walker Process antitrust action over the Taq '818 patent.

3. Hoffmann-La Roche, Inc. v. Promega Corp. (PCR/Taq licensing dispute — '195 part of the licensed family)

• Plaintiffs: Hoffmann-La Roche, Inc. and Roche Molecular Systems, Inc.
• Defendant: Promega Corporation
• Jurisdiction: U.S. District Court for the Northern District of California
• Case number: No. 93-1748 (N.D. Cal.) (as cited in the D.D.C. MDL opinion); an earlier related ruling appears at 33 U.S.P.Q.2d 1641 (N.D. Cal. June 13, 1994)
• Filing date: Roche filed in October 1992 (Oct. 27, 1992 per the MDL opinion) asserting infringement of US 4,889,818 and breach of the Cetus–Promega license agreement (a license covering Roche's PCR process patents, i.e., '195/'202)
• Outcome: District court found the '818 patent unenforceable for inequitable conduct (Dec. 7, 1999); Fed. Cir. vacated in part and remanded, 323 F.3d 1354 (Fed. Cir. 2003); on remand the district court again held the '818 patent unenforceable, 319 F. Supp. 2d 1011 (N.D. Cal. 2004).
• Sources: https://law.justia.com/cases/federal/appellate-courts/F3/323/1354/[575875](/patent/575875)/ ; https://www.courtlistener.com/opinion/[2316783](/patent/2316783)/hoffman-la-roche-inc-v-promega-corp/
• Confidence that '195 itself was asserted: Low. The asserted patent was '818; '195/'202 are the PCR process patents licensed in the same agreement.

---

Cases involving the PCR process patents where I could NOT confirm '195 was specifically asserted

4. Applera Corp. (Perkin-Elmer / Roche PCR license) v. MJ Research Inc.

• Plaintiff: Applera Corp. (successor to PE Biosystems / Perkin-Elmer); Roche licenses the PCR process patents
• Defendant: MJ Research, Inc.
• Jurisdiction/case number: Not confirmed by me (MJ Research litigation proceeded in more than one district; I do not have a verified case number)
• Filing: Not confirmed
• Patents: The opinion I retrieved centers on the '610 thermal-cycler patent and refers generally to "the PCR process patents" and the '493 patent in the induced/willful infringement analysis. It does not name '195 in the portion retrieved.
• Outcome: Jury found willful infringement; district court upheld the verdict, including inducement of infringement of the PCR process patents.
• Source: Applera Corp. v. MJ Research Inc. (D. Conn./N.D. Cal. opinions), https://www.courtlistener.com/opinion/[2431944](/patent/2431944)/applera-corp-v-mj-research-inc/
• Confidence that '195 was asserted: Unverified. I would want the docket and infringement contentions to confirm.

5. Roche enforcement campaign against PCR users (1990s)

• Contemporaneous reporting confirms Roche's position that it owned "U.S. Patent Numbers 4,683,195 and 4,683,202 which cover the PCR process and U.S. Patent Number 4,889,818 which covers Taq DNA Polymerase," and that users required a license (Bio.net/Diagnost mailing list posts, Nov. 1993 and June 1995: http://www.bio.net/hypermail/METHDS-REAGNTS/1993-November/008757.html ; http://iubioarchive.bio.net/bionet/mm/diagnost/1995-June/000141.html). These are licensing/enforcement assertions, not a specific court case I can cite with a docket number.

6. Roche v. Organon Teknika — not applicable to the US '195 patent

• Business-wire reporting describes Roche suing Organon Teknika in The Hague, Netherlands over a European PCR patent (NASBA HIV tests). Different patent, different jurisdiction — included only to explain why it should NOT be counted as '195 US litigation.
• Source: https://insights.citeline.com/MT004681/In-Brief-Roche-v-Organon-Teknika/

7. Board of Trustees of Stanford University v. Roche Molecular Systems — not applicable to '195

• This Supreme Court case (563 U.S. 776) concerned Stanford's HIV-quantification patents, not the '195 patent; Cetus PCR rights are background only.
• Source: https://www.courtlistener.com/opinion/[218133](/patent/218133)/board-of-trustees-of-the-leland-stanford-junior-university-v-roche/

---

Honest gaps

I could not verify any case number, PACER docket, or current litigation docket tied specifically and uniquely to US 4,683,195 from the sources available to me. I also could not retrieve the Darts-IP litigation list linked from the Google Patents family page. Specific things I do not know with confidence and will not fabricate:

• The case number in DuPont v. Cetus (N.D. Cal., filed ~Aug. 1, 1990).
• Whether the '195 patent (as opposed to '202) was separately asserted in any 1990s Roche infringement suit.
• Whether the '195 patent was ever the subject of a reexamination, IPR/PGR, or inter partes reexamination — the patent expired in 2005, well before AIA post-grant proceedings, so any challenge would have been pre-AIA (reexamination or district court validity).

Net conclusion: The only litigation I can ground in a source that expressly names US 4,683,195 is (a) the 1990–91 DuPont v. Cetus declaratory judgment in N.D. Cal. (patents upheld, no infringement), and (b) the '195 patent's appearance in the record of Molecular Diagnostics Laboratories v. Hoffmann-La Roche (D.D.C.), where it is named as a related Roche PCR patent rather than the asserted patent. Otherwise, '195's known litigation history is intertwined with, and generally subsumed within, the '202 PCR process patent and the '818 Taq patent disputes (Roche v. Promega; the Applera/MJ Research thermal-cycler and PCR-process litigation).

I'll verify the structured-data default (no AIA trials) against public sources before writing the memo.

Confirming there's no AIA trial history and checking the non-AIA post-grant history (the B1 certificate) for context.

One more check on the B1 reexamination certificate contents and the pre-AIA validity challenge (the closest analogue to an AIA trial here).

Proceedings overview

Zero AIA trial proceedings are on file against US 4,683,195 — 0 active, 0 with claims invalidated, 0 with claims sustained, 0 settled, and 0 institution denials. That is the canonical USPTO Open Data Portal answer, and my independent web search surfaced no IPR, PGR, or CBM naming the '195 patent (the hits were all unrelated patents, e.g., U.S. 8,421,195 and 9,662,273). The defensive posture this gives a defendant is not "the patent is hardened by PTAB wins," and it is not "the troll's claims are canceled." It is simpler and stronger: the '195 patent is expired, and it never lived long enough to be subject to an AIA trial. Any demand letter citing it is either stale or incompetent.

No AIA trial proceedings — and why there never could be

There is no proceeding to summarize, so I'm stating the negative plainly rather than padding the record:

• Timing makes IPR/PGR/CBM structurally impossible here. AIA trials began 2012-09-16. The '195 patent (earliest priority 1985-03-28; filed 1986-02-07; issued 1987-07-28) reached its anticipated expiration on 2005-03-28, per the Google Patents legal-status record. The patent was already dead for roughly seven years before the first IPR petition could lawfully be filed. PGR is categorically unavailable (it reaches only patents with effective filing dates on/after 2013-03-16). CBM is unavailable on subject matter and the program sunset on 2020-09-16.
• Why the absence is expected rather than suspicious. The "well-asserted patents eventually attract IPRs" heuristic does not apply to pre-1998 patents. The relevant enforcement era for this family ran 1989–2004, ending before AIA trials existed. The structured record does flag family litigation (Darts-IP family 27124781), which is consistent with the well-documented enforcement campaigns below — those were district-court and reexamination fights, not PTAB fights.

Non-AIA post-grant history (important context — these are not AIA trials)

I am listing these separately and labeling them precisely so nothing here is misread as a PTAB outcome.

Ex parte reexamination of the '195 patent (and the sibling '202 patent) — claims confirmed, 1990

• Type: Ex parte reexamination under 35 U.S.C. §§ 301–307 (pre-AIA). Not an AIA trial; no PTAB panel, no petitioner standing, no FWD.
• Trigger: E.I. du Pont de Nemours filed a declaratory-judgment invalidity action on 1989-08-01; the USPTO ordered reexamination of claims 19–25 of the '195 patent shortly afterward, and du Pont and a second requester filed their own requests, which the Office consolidated. See the du Pont/Cetus litigation record at CourtListener, E.D. Va. 2:19-cv-? docket doc. 48 and doc. 50.5.
• Grounds considered: anticipation and obviousness over Khorana-era "repair replication" art (Kleppe et al.; Panet & Khorana) — i.e., § 102/§ 103 challenges to the core amplification claims.
• Outcome: On 1990-08-23 the USPTO upheld the validity of the '202 and '195 patents in their entirety, rejecting the Kleppe/Panet & Khorana anticipation theory. Cetus publicly stated on 1990-08-27 that the Office would "issue certificates reaffirming the validity of all claims of two Cetus patents" (Cetus press release, Cohen AIDS collection, University of Michigan). Consistent with that, Google Patents records publication of the reexamination certificate US4683195B1 on 1990-11-27.
• Claim-level result: all claims confirmed; none canceled. Per the Cetus announcement, the certificates reaffirmed validity of all claims.
• Caveat / verify: at least one publicly filed license definition refers to "B2 4,683,195" (alongside "B1 4,683,195"), which would imply a second reexamination certificate. I could not confirm a B2 in this session and I will not assert it. If claim status matters to you, pull the certified copy from USPTO Patent Center and check the file wrapper.
• Appeal: none — ex parte reexamination certificates are not appealable PTAB decisions. The parallel district-court fight was the du Pont declaratory-judgment action; I retrieved the briefs/opinion excerpt but not the final disposition in this session, so I am flagging its ultimate outcome as unverified rather than guessing.
• Defensive value today: nil as an invalidity tool (the Office confirmed every claim), but the historical value is that it shows the strongest prior art of the era — Kleppe/Panet & Khorana — was already run and rejected. Do not lead with that art.

Later enforcement (context for why no AIA trial exists)

• Roche/Cetus asserted the '195 patent (among '202, '188, '493, '756, '675, '610) in district court through the 1990s and early 2000s, including against MJ Research — see the District of Connecticut opinion in Applera v. MJ Research (2004-02-24), which discusses contributory-infringement theories on the '195 patent. All of this predates the AIA trial regime.

Strategic summary

Claim status: nothing canceled, nothing sustained in an AIA forum, and nothing left to assert. No claim of US 4,683,195 has ever been canceled by the PTAB, because no IPR was ever instituted. The only claim-status event of record is the 1990 ex parte reexamination, in which the Office confirmed all claims (B1 certificate published 1990-11-27, per the Cetus announcement and Google Patents). So the "surviving claims" question is academic: the patent expired 2005-03-28. There is no narrow-surviving-claims story here — there is a term story. Note the related family members in the structured data carry more recent reexamination certificates (e.g., the '188, '493, '756 patents), and if a counterpart patent or a continuing application is what's actually being asserted, that is a different analysis and you should pull its own prosecution/post-grant history.

Estoppel landscape: § 315(e)(2) is a non-issue. Estoppel attaches only to a petitioner that obtained an FWD — and there is no petitioner. Nothing is barred; a defendant could theoretically raise any ground. That said, an IPR is the wrong tool: on an expired patent, the only possible relief is pre-expiration damages, and 35 U.S.C. § 286 bars recovery for infringement more than six years before the complaint is filed. A suit filed today over a patent that expired 2005-03-28 cannot reach any infringing act within the damages window. The available "prior-art grounds" are therefore irrelevant — the dispositive defense is expiry and the statute of limitations, not art.

Pattern signals: none of the usual tells are present. No repeat petitioner, no serial IPRs, no PTAB appeals, no defensive aggregator (no Unified Patents involvement — and Unified did not exist in any relevant form during this patent's life). The litigation-family flag in the structured data is litigation, not PTAB activity. Historically, the patent owner was the aggressor (Cetus, then Roche), not a party defending against PTAB challenges. If someone is now asserting the '195 patent, the anomaly is on their side.

Recommended next steps

• No FWD exists to link or quote. I will not manufacture a disposition. For negative confirmation, query PTAB E2E by patent number and application number 06/828,144; the result set is empty.
• If you received a demand letter citing US 4,683,195: verify the number as written and check its status. The patent's anticipated expiration is 2005-03-28 (Google Patents legal-status record; authoritative patent text). Any assertion is at minimum time-barred under § 286 and cannot support injunctive relief. First question back to the sender: identify the assertedly infringing acts and their dates.
• If the real target is a family member still in force (the structured data lists many continuations, e.g., US 4,800,159, 5,176,995, 5,386,022, 6,040,166, 5,594,123, 6,197,563, 6,514,736 — all from the same Cetus/Roche chain), pull that patent's post-grant history separately. AIA trial activity, if any, will be attached to the live claims, not to the expired '195.
• Verify the B1/B2 certificate question in Patent Center before relying on "all claims confirmed" as a claim-status statement; my sourcing for that is the 1990 Cetus press release plus the Google Patents B1 publication entry, both of which are strong but neither of which is the certificate itself.

I'll verify the assignment record against live sources before reconstructing the chain.

Key records surfaced. Now verifying the anomalous Laurus/Bioniche entries and the correspondent of record.

US 4683195 — Ownership & Assignment Reconstruction

Patent: US 4,683,195 A — "Process for amplifying, detecting, and/or-cloning nucleic acid sequences"
Appl. No. 06/828,144 · filed 1986‑02‑07 · priority 1986‑01‑30 · published 1987‑07‑28 · granted 1990‑11‑27 (INPADOC also lists a US4683195 (B1) record dated 1990‑11‑27)
Current assignee of record: Roche Molecular Systems, Inc. (Roche Diagnostics) · Legal status: Expired – Lifetime

Source note / limitation: I could reach the Assignment Center mirror data through Google Patents legal events and Espacenet INPADOC for this patent, and those do resolve reel/frame numbers (used below). Assignment Center's live UI at https://assignmentcenter.uspto.gov/ and https://assignment.uspto.gov/patent/index.html could not be queried directly in this pass, and I was unable to retrieve the correspondent-of-record field for any reel/frame entry. Correspondent names below are therefore marked not captured rather than guessed. Reel/frame numbers, conveyance types, execution dates, and recording dates below are as reported by those mirror sources.

---

Inventors

Unusual patterns:

• Inventor departure — yes, but isolated. Mullis left Cetus in 1986, the same year as the 1986‑02‑07 filing (to Xytronyx, later an independent consultant). Under the "all inventors leave within 12 months" heuristic this is half a signal: only one of six departed. Here it did not precede a fire‑sale — it preceded Cetus keeping and commercializing the portfolio for another five years.
• Inverse of the usual troll precursor. Rather than inventors scattering, the whole team followed the asset: when Cetus sold the PCR business to Roche for $300M in 1991, Cetus' ~125‑person PCR division (including Erlich and Saiki) moved to Roche's Alameda, CA PCR facility. Continuity of inventors into the acquiring operating company is a strong anti‑NPE tell.
• Signing dates on the original assignment run 1986‑02‑07 to 1986‑02‑14 (reel 004526/0934) — i.e., executed within days of filing, as expected for an employee invention‑assignment program.

---

Original assignee

Cetus Corporation, 1400 Fifty‑Third Street, Emeryville, California (recorded on reel 004526/0934).

• Primary line of business: First‑generation biotech; recombinant therapeutics (Proleukin/IL‑2, beta‑interferon) and research/diagnostic products. PCR and the Perkin‑Elmer Cetus Instruments joint venture (Cetus 49%, Perkin‑Elmer 51%) were its research‑tools business.
• Did it ship products embodying the claims? Yes. Cetus commercialized PCR reagents/services and the PCR instrumentation JV distributed thermal cyclers; it also licensed PCR broadly (e.g., licensed the sibling Taq patent to Promega in June 1990; exclusively licensed PCR diagnostics to Roche in 1989).
• Litigation by the original assignee: Cetus successfully defended the PCR patents against a DuPont validity challenge (favorable ruling February 1991, N.D. Cal.) before selling them.
• Current status: Dissolved as an independent entity. After an FDA advisory‑committee rejection of Proleukin (July 1990), Cetus announced in July 1991 a merger into Chiron Corporation (stock swap ≈$660M); the sale of the PCR business to Hoffmann‑La Roche for $300M cash plus up to $30M in royalties was a condition of that merger. The Cetus name disappeared. This was a distressed‑merger, not a Chapter 7/11 bankruptcy sale.
• Nature of the 1991 transfer: a $300M purchase of a complete operating PCR business (patents, manufacturing, and staff) — see LA Times 1991‑07‑23 and Nature 1991 coverage.

---

Assignment timeline

Five records are indexed against US 4,683,195. Three are chain‑of‑title conveyances; two are a lender's security interest and its release.

• 1986‑02‑07 (executed; latest signing date 1986‑02‑14) / recorded 1986‑03‑03 — Reel 004526/0934

  • Conveyance: Assignment of assignors' interest
  • Assignor: Mullis, Kary B.; Erlich, Henry A.; Arnheim, Norman; Horn, Glenn T.; Saiki, Randall K.; Scharf, Stephen J.
  • Assignee: Cetus Corporation, 1400 Fifty‑Third Street, Emeryville, CA
  • Correspondent: not captured (the mirror legal‑status feeds omit this field; requires the reel/frame image or Assignment Center detail view)
  • Context: Original employee‑inventor assignment to the operating company; standard practice, executed within days of filing.

• 1991‑12‑11 (executed) / recorded 1992‑01‑17 — Reel 005974/0222

  • Conveyance: Assignment
  • Assignor: Cetus Corporation
  • Assignee: Hoffmann‑La Roche, Inc., New Jersey
  • Correspondent: not captured
  • Context: Acquisition of the entire PCR business (asset + personnel + IP) for $300M plus up to $30M royalties, executed as the condition to the Cetus/Chiron merger — an operating‑company acquisition, not a fire‑sale. Note the execution date is dated to completion of the sale (Dec 1991), with recording ~5 weeks later.

• 1997‑01‑08 (executed) / recorded 1997‑01‑27 — Reel 008321/0096

  • Conveyance: Assignment of assignors' interest ("ASSIGNMENT OF ASSIGNORS INTEREST")
  • Assignor: Hoffmann‑La Roche Inc.
  • Assignee: Roche Molecular Systems, Inc., New Jersey
  • Correspondent: not captured
  • Context: Internal corporate reorganization — same Roche group, moving the PCR/diagnostics IP into the diagnostics operating subsidiary (Roche Molecular Systems, Alameda/Branchburg). No change in ultimate control.

• 2005‑12‑02 (executed) / recorded 2005‑12‑14 — Reel 016891/0196 ⚠️ anomalous entry

  • Conveyance: Security Agreement (grant of security interest — not a transfer of title)
  • Assignor: Bioniche Life Sciences Inc., Ontario
  • Assignee: Laurus Master Fund, Ltd., New York
  • Correspondent: not captured
  • Context: Flag for verification. Bioniche is a Canadian human/animal‑health biopharma with no PCR business and no plausible interest in this patent; Laurus Master Fund was a distressed‑debt lender that took blanket patent security interests. Bioniche's three‑year US$7M secured convertible term note from Laurus is corroborated as real (BioSpace, 2007‑04‑18; Bioniche 2009 AIF), but the inclusion of US 4,683,195 in that collateral schedule is not explained by any Bioniche product or business. The execution date also post‑dates the anticipated expiration date recorded on Google Patents (2005‑03‑28). This looks like a mis‑indexed collateral‑schedule patent number. Either way it is a lien, not a conveyance of title, and it never moved ownership from Roche.

• 2009‑07‑14 (executed) / recorded 2009‑10‑28 — Reel 023438/0096

  • Conveyance: Release by Secured Party
  • Assignor: Laurus Master Fund, Ltd. (as agent) and Valens U.S. SPV I, LLC c/o Valens Capital Management, LLC
  • Assignee: Bioniche Life Sciences Inc.
  • Correspondent: not captured
  • Context: Release/termination of the 2005 security interest following repayment of the Laurus facility (Bioniche repaid the convertible term note with shares in April 2007 and converted the revolver with Valens in 2008). Closes out the anomalous collateral entry; no title change.

No chain‑of‑title assignment has been recorded since 1997. Roche Molecular Systems, Inc. remains the owner of record.

Related legal events (non‑assignment):

• The patent's sibling and family patents were heavily litigated: Hoffmann‑La Roche v. Promega (N.D. Cal., filed 1992) — the court held the sibling Taq patent '818 unenforceable for inequitable conduct (Roche I 1999; affirmed in part, Fed. Cir. 2003, 323 F.3d 1354; unenforceability re‑entered 2004) while expressly leaving the '195 and '202 PCR patents intact for Roche to enforce. The 2008 $33M direct‑purchaser antitrust settlement (D.D.C. MDL) concerned Taq pricing, not this patent.
• Google Patents flags "Family has litigation" for this family (large worldwide family: EP 0200362, EP 0509612, EP 0502589, DK 171161, CA 1291429, AU 591104, and continuations US 4,800,159 / US 5,176,995 / US 5,386,022 / US 5,594,123 / US 6,047,166 / US 6,197,563, etc.).
• Term: Google Patents records an anticipated expiration of 2005‑03‑28, consistent with 20 years from the earliest US application in the priority chain (Ser. No. 716,975, filed 1985‑03‑28). I could not retrieve the USPTO term record to confirm whether the pre‑URAA "greater of 17‑years‑from‑grant" rule (which would give 2007‑11‑27) or a terminal disclaimer governs — treat the 2005 date as unverified but plausible. Either way the patent is long expired and unassertable today.

---

Timeline diagram

timeline
    title Ownership of US 4683195
    1985 : Earliest priority application filed 28 Mar 1985
    1986 : Application filed 07 Feb 1986
         : Six inventors assign to Cetus Corporation
    1987 : Patent published 28 Jul 1987
    1990 : Patent granted 27 Nov 1990
    1991 : Cetus sells PCR business to Roche for 300M
    1992 : Assignment to Hoffmann La Roche recorded
         : Roche sues Promega over PCR patents
    1997 : Internal transfer to Roche Molecular Systems
    2005 : Security interest recorded by Laurus Master Fund
         : Anticipated expiration recorded
    2009 : Security interest released to Bioniche

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. No assignee in the chain carries an "IP / Holdings / Licensing / Ventures" suffix. The only post‑1997 record is a lender's security interest (reel 016891/0196, 2005‑12‑02) and its release (reel 023438/0096, 2009‑07‑14) — liens, not title. Title has rested with Roche Molecular Systems since reel 008321/0096 (1997).

2. Known asserter in the chain — NOT PRESENT. Assignees are Cetus → Hoffmann‑La Roche → Roche Molecular Systems. None appears on the Acacia / Marathon / IV / Wi‑LAN / Conversant / Vringo / Pendrell / Round Rock / Spangenberg‑type lists, and none surfaces in Unified Patents or RPX high‑frequency‑plaintiff directories. Roche is a pharmaceutical/diagnostics operating company. Laurus Master Fund is a distressed‑debt lender, not an NPE — securitizing loan collateral is not assertion activity.

3. Repeat correspondent across the chain — UNCLEAR (not determinable). I could not obtain the correspondent-of-record field for any of the five reel/frame entries from the sources reachable in this pass. This is the one check that genuinely needs the Assignment Center detail view or reel/frame images; if a single correspondent recurs across reels 004526/0934 → 005974/0222 → 008321/0096, that would be routine corporate/outside counsel for Cetus and Roche respectively and would still not be an NPE finding absent an assertion record.

4. Cascading transfers — NOT PRESENT. Three title transfers across 39 years (1986, 1992, 1997), with the last two five years apart. The only tight cluster is the lien + release pair (2005/2009), which is lender mechanics on a non‑Roche entity and does not move title.

5. Pre-litigation transfer — PRESENT, but benign. The Cetus→Roche assignment was executed 1991‑12‑11 and recorded 1992‑01‑17; Roche sued Promega in 1992. That is within six months of first assertion. However, the transfer was a $300M purchase of a complete operating division (patents + manufacturing + 125 employees), not venue/standing engineering, and Roche asserted the patent in its own name as an operating competitor in the PCR reagents market. Scored present on timing, negative on the pattern.

6. Bankruptcy fire-sale — NOT PRESENT. Cetus was distressed (IL‑2 rejection, $61.5M FY1990 loss) but merged into Chiron; it did not file Chapter 7/11 and its patents were not sold in a bankruptcy proceeding. The 1992 Roche transfer was a negotiated $300M sale of a going concern. The Laurus/Bioniche financing is a distressed‑microcap lender relationship involving a different company, with the security interest released rather than foreclosed.

7. Privateering — NOT PRESENT. No transfer out of Roche to an assertion vehicle; Roche itself enforced the family (Roche v. Promega, and the family-level litigation flagged by Google Patents). The recorded chain never leaves the Roche group after 1997.

8. Defensive aggregator — NOT PRESENT. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. This patent was not neutralized by defensive aggregation; it was enforced and then simply expired.

---

Verdict

Operating-company assertion.

The entire recorded chain of title runs Cetus → Hoffmann‑La Roche (reel 005974/0222, executed 1991‑12‑11) → Roche Molecular Systems (reel 008321/0096, executed 1997‑01‑08) and has not moved since; the only later records are a lender's security interest and its release (reels 016891/0196 and 023438/0096), which are not conveyances. The current owner ships products embodying the claims (Roche Molecular Systems PCR reagents and instruments, staffed in part by the inventors who transferred with the business) and asserted the patent family itself against an actual competitor, Promega, in 1992 — the opposite of the shell‑LLC pattern. No NPE signal is present on any of the eight checks; the only caution flag is the data anomaly that a Bioniche/Laurus security‑interest record is indexed against this patent number, which should be resolved by pulling the reel 016891/0196 image before treating this patent's legal‑status table as clean.

Verify at: https://assignmentcenter.uspto.gov/ — search patent number 4683195 (mirror listing: Google Patents legal events at https://patents.google.com/patent/US4683195/en ; INPADOC legal status at https://worldwide.espacenet.com/patent/search?q=US4683195).

I'll verify the record and pull the cited prior art for US 4,683,195 specifically.

Let me pull the patent's front-page "References Cited" list and cross-check the citation counts.

I need the actual front-page citation list. Let me pull the patent PDF text and the Google Patents citation tables.

Prior-Art Analysis — US 4,683,195

Patent under analysis: US 4,683,195 A (also US4683195B1) — "Process for amplifying, detecting, and/or-cloning nucleic acid sequences." Appl. No. 06/828,144; filed 1986‑02‑07; priority 1986‑01‑30 (Ser. No. 824,044), itself a division of Ser. No. 791,308 (filed 1985‑10‑25), itself a CIP of Ser. No. 716,975 (filed 1985‑03‑28); issued 1987‑07‑28; B1 reexamination certificate 1990‑11‑27; expired.

Search scope confirmation. I searched the literal string 4683195 / 4,683,195. Every hit I used resolves to this record. I did not substitute 4,683,202 (the co-filed sibling PCR patent), 4,889,818, 4,965,188, 4,800,159, 8,421,195, or 9,662,273 — each of which surfaced in the same result sets and was discarded. One caution: Espacenet's page for US4683195 reports "aproximadamente 6.257 documentos citados" — that is the forward-citation count (later patents citing '195), not prior art. Do not confuse the two.

Date note. The system header says 2026‑09‑29; your instruction says April 26, 2026. As in the earlier sections, I treat April 26, 2026 as the report date. Immaterial to anything below.

Builds on prior sections. Claim text, status, ownership, and the § 103 analysis are in the earlier sections and are not repeated. This section corrects one item and completes another — see § 7.

---

1. First, an important limitation on "each patent citation"

You asked me to look at each patent citation on the '195 patent. I have to be straight about what I could and could not retrieve.

What I could not retrieve: I could not obtain the verbatim front-page "References Cited" box of the printed '195 patent — i.e., the list of U.S. Patent Documents as it appears between the Abstract and the first claim column. The fetched full text supplied via Google Patents (reproduced in the record above) contains the description, the abstract, the claims, and the legal events, but not the front-matter citation box. So I will not recite a "U.S. PATENT DOCUMENTS" list as if I had seen it.

What the record does tell me:

• The archived Google Patents bibliographic frame for US4683195 lists "非特許引用 (12)" — i.e., 12 non-patent citations — and, in that same header block, no patent-citation count. That frame normally prints a patent-citation count when one exists. This is consistent with a hypothesis I cannot yet confirm: that the '195 front page carries a non-patent-literature-only citation box. Treat as hypothesis, not finding.
• The earlier sections' claim that the reexamination covered claims 19–25 is now corroborated verbatim by the N.D. Cal. opinion: "PTO ordered reexamination of all individual claims of the '202 patent pertinent to this motion and of claims 19–25 of the '195 patent" — and "found no new question of patentability with respect to claims 1–18 and 26, and declined to reexamine those claims."

What I therefore did instead of guessing: I built the reference set from three grounded sources, and I label each reference's provenance explicitly so you can tell examiner-cited art from spec-admitted art from third-party-asserted art. That distinction matters legally (§ 102 vs. § 103; "admitted prior art" weight) and it is exactly the distinction that decided this patent.

---

2. Patent documents in the '195 record

2a. U.S. patents named in the specification [SPEC]

Every one of these is cited in the '195 description as background art, which makes each an admission against interest and available under § 102(b) if published more than one year before the earliest filing date (i.e., before 1984‑03‑28). Publication dates below are given to the precision I can defend.

2b. The references that actually mattered — asserted by third parties, not cited by the examiner [REEXAM]/[DJ]

This is the headline finding of this section. The two references that generated the entire validity fight over this family — and the ones the PTO upheld the claims against — were not applicant-cited and, on the record I retrieved, not part of the '195 citation box. They entered via third-party reexamination requests filed by DuPont and Hoffmann‑La Roche in 1989–90 and via DuPont's declaratory-judgment action:

Note on the Kleppe citation: some secondary sources render it as J. Mol. Biol. 66 (1971); the N.D. Cal. opinion and the '188 front page both give 56:341–361. I use 56 and flag the discrepancy rather than silently picking one.

---

3. Non-patent literature (the bulk of the 12-item count)

---

4. § 102 anticipation map — which claim(s) each reference could reach

4a. The threshold problem every challenger hit

No independent claim of '195 is a bare amplification claim. From the claim text now verified (Wayback Google Patents snapshot, 26 claims):

• Claim 1 = (a) anneal/extend on each strand + (b) denature + (c) extend again + (d) add a labeled oligonucleotide probe + (e) determine hybridization.
• Claim 16 = amplification + hybridization to a normal β‑globin probe + digestion with the sickle-site restriction enzyme + detecting a correlated fragment.
• Claim 19 = synthesis of a fragment from a shorter "core" fragment using primers with non-complementary 5′ extensions.
• Claim 26 = cloning into M13 of an HLA DQ/DR/DP Class II α/β sequence using primers with different 5′ restriction sites, in DMSO at 35°–40 °C.

So the pure cycling content is a sub-step of each independent claim. That is precisely why DuPont's § 102(b) motion was aimed at '202 claims 1–4, 6–9, 11, 16, 19 and 20 — and its '195 attack was pleaded only under § 103, against claims 1–4, 6, 11, 12, 14 and 15 (N.D. Cal., C 89‑2860 MHP, 1990 WL 305551). No § 102 anticipation ground was ever pressed against any claim of '195 in either forum. That is a documented, not inferred, finding.

4b. Reference-by-reference

4c. § 102(g) — the "prior invention" theory

DuPont also ran a § 102(a)/(g)-flavoured theory that PCR was in fact invented in Khorana's lab between 1969 and 1974: Ruth Kleppe and Ian Molineux testified for DuPont, and Stuart Linn's notes of a late-1960s conference presentation reportedly showed Kleppe's diagrams reciting the cycle. Because the conference proceedings were not published and the notes were not admitted as evidence, that theory could not be reduced to a § 102(b) printed publication and never ripened. It is worth recording because it shows the challenger's only route to § 102 against '195's core claim was an unpublished-oral-disclosure/misappropriation theory — i.e., a § 102(a)/(g) theory with a proof problem, not a documentary one. Do not lead with it.

---

5. Why every § 102 theory failed — the enablement asymmetry

The PTO and the district court each rejected anticipation on the same ground, and it is a § 102-specific one:

the publications are "too indefinite and uncertain to anticipate PCR, in part because they fail to specify the precise nature of the primers to be used, the size of DNA duplex to be copied, or specific reaction conditions required for DNA synthesis." (N.D. Cal., 1990 WL 305551, 19 U.S.P.Q.2d 1174)

That is the In re LeGrice / Titanium Metals rule: an anticipatory reference must be enabling. It is a § 102 shield only. A § 103 challenger escapes it, because for obviousness the specification's own enablement of the claimed genus supplies what the reference lacks — the asymmetry I flagged in the earlier obviousness section. This is why the '195 record reads as it does: § 102 failed, and the fight was § 103.

Supplementary data points: Kleppe's own Fig. 11 plateaued at ~8,000 cpm against a ~12,000 cpm exponential prediction — the reference failed to show exponential amplification. And the Khorana progress report contains an express teaching away: complete replication "could not be realized." Both cut against anticipation, and under KSR they cut less against obviousness.

---

6. Non-AIA post-grant outcome on the citations (context, not a PTAB action)

• Reexamination scope for '195: claims 19–25 only. The Office found no new question of patentability as to claims 1–18 and 26 and declined to reexamine them. So claim 1 of '195 was never reexamined — ever. This corrects the "PTO upheld the '195 patent in its entirety" framing used in the earlier sections (correction restated in § 7).
• Result on the examined claims: the Examiner upheld claims 19–25 of '195 and all challenged claims of '202 against Kleppe and Panet & Khorana, and "found unpersuasive the assertions of the Kornberg and Sambrook declarations that Panet & Khorana teaches PCR."
• Effect in litigation: because the reexam certificate issued, the N.D. Cal. court held DuPont's burden of proving invalidity was "heavier than it would have been in the absence of the reexamination," and denied summary judgment on both § 102 and § 103.
• Timing point that reinforces the earlier sections: the Office's decision (1990‑08‑23) and the B1 certificate (1990‑11‑27) both long predate the AIA. There is no IPR/PGR/CBM anywhere in this record.

---

7. Corrections and completions to the earlier sections

1. COMPLETION — '195 claims 18–26 are now verified. The earlier sections flagged claims 18–26 as unverified. Retrieval resolves them:

• Claim 18 — dependent on claim 16; positive and negative controls.
• Claims 19–25 — independent claim 19 covers synthesizing a nucleic acid fragment from a shorter existing "core" fragment using primers whose 5′ ends supply non-complementary left/right segments; claims 20 (repeat (b)–(c) ≥5×), 21 (recursive core), 22 (core from two partially complementary oligos), 23 (purify before step (e)), 24 (do not purify), 25 (four dNTPs + polymerization agent) depend from it. These are the claims the PTO actually reexamined.
• Claim 26 — independent, M13 cloning of an HLA DQ, DR or DP Class II α/β polymorphism, primers with two different 5′ restriction sites, amplification in DMSO at 35°–40 °C, orientation-specific ligation.

1. CORRECTION — there is no kit claim in US 4,683,195. The earlier summary inferred that a diagnostic-kit claim existed in this patent because the specification discloses one. The verified claim set (1–26) contains no kit claim. The multicontainer-unit kit invention is claimed in the European family member — EP 0502589 ("Kit zur Verwendung bei der Amplifizierung und beim Nachweis von Nukleinsäuresequenzen") / DE3650763 — not in the US '195. Do not attribute kit claims to this patent.
2. CORRECTION CONFIRMED — reexamination was partial, not "all claims." § 6 above. The "PTO confirmed all claims of '195" framing must not be used for claim 1, because the Office never examined it.
3. CORRECTION CARRIED — DuPont filing date is 1989‑08‑01, case C 89‑2860 MHP, per the primary opinion ("Shortly after du Pont filed its declaratory judgment action on August 1, 1989…"). The litigation section's "August 1, 1990" is the error; the PTAB section's 1989 date is right.
4. NEW — the citation box's most prominent non-art entries. US 4,582,789 and 4,617,261 are cited in the '195 specification but issued after the '195 filing date. Anyone reading the citation box as a prior-art list will over-count the art by two.
5. NEW — the decisive references were not cited by the examiner. Kleppe et al. 1971, Panet & Khorana 1974, and the Khorana NSF grant proposal reached this patent only through third-party reexamination requests and the DJ action. If you are reconstructing the '195 art-of-record, the examiner-cited universe and the litigated-art universe are different sets, and the litigated set is the one that matters.

---

8. Bottom line

For anticipation specifically, this patent is unusually clean, and the reason is structural rather than lucky. Every independent claim of '195 bolts a detection, cloning, or fragment-synthesis step onto the amplification cycle. The only references that disclose the cycle (Kleppe 1971; Panet & Khorana 1974; the Khorana grant proposal) disclose nothing of the detection/assembly overburden — so no single reference reads on any claim. Falkow supplies the detection half and no cycle at all. Conner supplies the allele discrimination and no cycle. Maniatis/Messing supply the cloning architecture and no cycle. Every real combination is § 103, and it was pleaded and decided as § 103.

The strongest reference against the broadest concept is Kleppe et al. 1971, because of its express statement that "the whole cycle could be repeated" and because it used E. coli DNA polymerase I — but against this claim set it anticipates nothing, and the PTO and the district court so held on enablement grounds.

The most relevant reference to the claims as issued is US 4,458,066 (Caruthers solid-support synthesis), because it is what converts the recited primers from a research curiosity into a routine reagent — and therefore what forecloses the argument that the '195 genus was non-enabled.

---

9. What you should verify before relying on any of this

10. Confidence ledger

Per the operating rule, no patent number above has been auto-corrected or normalized. US 4,356,270, 4,293,652, 4,403,036, 4,416,988, 4,458,066, 4,358,535, 4,582,789, 4,617,261, 4,683,195, 4,683,202, 4,800,159, 4,889,818, 4,965,188, 5,176,995, EP 63,879, EP 0502589, EP 0200362, and DE3650763 are each reproduced exactly as they appear in the source text.

I'll ground the obviousness analysis in the actual cited prior art and the reexamination/litigation record.

Obviousness Analysis — US 4,683,195 under 35 U.S.C. § 103

Builds on the prior three sections (bibliographic, litigation, PTAB/post-grant). I do not repeat claim-construction or status data there; where this analysis depends on them I cross-reference. Date note: the system header and your instruction still disagree (2026‑09‑27 vs. April 26, 2026); I treat April 26, 2026 as the report date. This is a substantive, retrospective § 103 analysis of expired claims — not a live challenge vehicle (see § 7).

---

1. Scope of the "Prior Art section" and the legal frame

What the page's Prior Art section actually contains. The fetched record supplies (i) prior-art keyword tags — nucleic acid, sequence, primer, primers, synthesized — and (ii) a "Definitions" extract whose snippets are drawn verbatim from the specification's own "Description of Related Disclosures" (Narang phosphotriester; helicase/RecA strand separation per Cold Spring Harbor Symposia Vol. XLIII, Kuhn et al.; Radding; the pBR328:HbA/HbS MstII digests; the 40‑mer probe purification). Google Patents does not present an independent examiner-cited reference list on this page. My reference universe is therefore drawn from three grounded sources: (A) the specification's admitted prior art, (B) the court-verified reexam record (Kleppe; Panet & Khorana; Khorana NSF Grant Proposal), and (C) the art the patent itself cites as known. No reference is invented.

Governing law. Application 06/828,144 was filed 1986‑02‑07; the claims are governed by pre-AIA § 103. Any validity analysis of those claims today still applies KSR Int'l v. Teleflex (2007) — KSR construed pre-AIA § 103 — which replaced the rigid teaching/suggestion/motivation test that controlled in 1991 with an expansive, "predictable variation of known elements" standard. Net effect: the obviousness case against these claims is materially stronger today than the case Cetus defeated. The EPO counterpart decisions (T 78/96, T 216/96) are useful comparative data but are not U.S. § 103 law.

Critical date. The § 176/§ 102(b)/§ 103 reference date for the subject matter disclosed in parent Ser. No. 716,975 (filed 1985‑03‑28) is March 1984 — the framing du Pont itself used, and the one the court adopted ("obvious to one of ordinary skill in March 1984"). I adopt March 1984 as the POSITA knowledge date.

POSITA. A molecular biologist/immunologist (Ph.D. level) with routine access to oligodeoxynucleotide synthesis (Beaucage & Caruthers 1981; U.S. Pat. No. 4,458,066), E. coli DNA polymerase I and its Klenow fragment, T4 DNA polymerase and reverse transcriptase (all recited in claim 15), the four dNTPs, restriction enzymes, and standard cloning/probing manuals (Maniatis et al., Molecular Cloning (1982)).

---

2. The reference set, with each reference's actual teaching

---

3. The structurally important point about the claim set

Claim 1 contains no minimum cycle number and no exponential-amplification requirement. Per the claim summary in the earlier section, steps (b)–(c) "may be repeated," and it is claim 2 that adds "repeat at least once." If that reading is correct, claim 1 is met by a single denaturation followed by a single primer-extension round on each strand — i.e., precisely the first cycle Kleppe et al. describe and the strand-by-strand replication Panet & Khorana actually performed. Du Pont pressed exactly this structurally at trial: "Cetus chose to write its claim so that they cover an amplification process that involves only two or three cycles… contain none of the conditions that they now contend are so important to practicing PCR." (closing argument transcript, E.I. du Pont v. Cetus, N.D. Cal. C 89‑2860 MHP). The patent's own arithmetic table (2ⁿ − n − 1) is the challenger's best exhibit: once the cycle is defined, exponential accumulation is arithmetic, not invention.

Caveat (carried forward): the earlier section could not retrieve verbatim text for the '195 claims beyond the set it paraphrased. The "no cycle minimum" point is drawn from that summary and from the '202 claim-1 argument; verify against the printed claim column before relying on it.

---

4. Combinations that would render the claims obvious, with motivation

Set A — Claim 1 (and claims 2–9, 13–15): Kleppe 1971 + Falkow U.S. 4,358,535

• Mapping. Kleppe supplies (a) two primers on opposite strands, extension products, the express statement that the extension product can recycle as template ("the whole cycle could be repeated"), (b) heat denaturation, and the molar-excess teaching. Falkow supplies (d)/(e): a labeled probe hybridized to the target and a determination step — Falkow's entire disclosed purpose is detecting pathogen nucleic acid in a clinical sample.
• Motivation (three independent legs).

1. Express suggestion in the reference. "The whole cycle could be repeated, there being added every time a fresh dose of the enzyme" is the paradigm case of a TSM-qualifying statement — and under KSR it is more than enough.
2. Recognized, articulated problem. The specification's own Background concedes the state of the art: target sequences "may be only a small portion of the DNA… difficult to detect," and "little research has been conducted on amplifying the target sequence so that it is present in quantities sufficient to be readily detectable." An applicant's admission that amplifying the target is the desired remedy is a heavy § 103 admission. Falkow supplies the very sensitivity problem (a "relatively small number of pathogenic organisms" in a clinical sample) that amplification solves.
3. All elements known, each performing its known function, with no change in their respective operating principles. The patent claims, at bottom, a new use of a coupling of two known techniques — the classic pre-KSR "combination of old elements" and a fortiori a KSR predictable variation.

• Enablement overlay. The reexam record shows the Examiner rejected anticipation because Kleppe/Panet & Khorana are not enabling (no primer sequences, no duplex sizes, no conditions). That defeats § 102 but is a much weaker shield under § 103 so long as the claimed genus is enabled by the specification's disclosure combined with routine skill — the familiar asymmetry that du Pont exploited.

Set B — Claim 1, alternative: Panet & Khorana 1974 + Kuhn 1978/Radding 1982 + Maniatis 1982

• Motivation: Panet & Khorana already teach cycling on an immobilized template; Kuhn/Radding teach helicase/RecA denaturation in simultaneous (non-heat) mode; Maniatis teaches the buffer/enzyme/nucleotide conventions. This combination is the natural route to the claim's "steps (a)–(c) may be conducted… simultaneously" embodiment — a POSITA seeking an isothermal, automation-friendly format (cf. the patent's own "continuous"/immobilized-enzyme-column disclosure, which is itself a straightforward engineering extension of Panet & Khorana's cellulose-bound template).

Set C — Claims 11, 12, 16 (sickle cell / genetic disease): Set A + Conner 1983 + restriction polymorphism art

• Mapping. Claim 16 ties amplification of the β-globin region spanning the sickle polymorphism to (i) a labeled probe complementary to normal β-globin, (ii) digestion with the enzyme for the polymorphic site, (iii) detection of a restriction fragment correlated with disease status.
• Motivation. (1) The MstII/DdeI site alteration in the sixth codon of β-globin was itself known diagnostic art by 1984 (the patent's Fig. 6 is the conventional depiction of it; the S-allele abolishes the MstII site). (2) Conner et al. (1983), cited in the patent, already distinguished the alleles by oligomer hybridization — the patent's own Example 10 cites Conner and Leary as the state of the art it improves upon by amplification. (3) Adding amplification to an assay already known to work but known to be insensitive is the definition of an obvious improvement for a POSITA facing limited fetal/clinical sample mass. (4) The "oligomer restriction" read-out (Figs. 7–8) is an application of ordinary restriction mapping (Maniatis; claim 13's Southern blot is likewise admitted art).

Set D — Claim 10 and the cloning claims (19–25 as reexamined): Set A + Maniatis 1982 + Messing 1981

• Mapping. Restriction sites on the 5′ non-complementary ends of the primers; digestion; ligation into a vector.
• Motivation. (1) Adding a restriction site to the 5′ end of a synthetic oligomer was routine linker/adaptor practice (Maniatis pp. 390–401; U.S. Pat. Nos. 4,416,988 / 4,403,036 recite the same era's linker methodology). (2) The specification itself supplies the motivation by disparaging the prior methods — blunt-end ligation is "inefficient" and neither prior technique "would control for the orientation or multiplicity of insertion." Where the patent asserts a known problem with a known technique, the modification it adopts is presumptively obvious. (3) Different 5′ and 3′ sites to fix orientation is the standard directional-cloning convention, not an invention.

Set E — Claim 15 (enzyme selection): Set A + Kaledin 1980

• Motivation. Claim 15's Markush group (E. coli DNA polymerase I, Klenow, T4 DNA polymerase, reverse transcriptase) was the ordinary toolkit; the "heat-stable enzyme… 65–90 °C" alternative is the applicant's own cited reference (Kaledin et al.). Citing a reference in the specification is an admission of its prior-art status; using the heat-stable polymerase that reference supplies, for the elevated-temperature cycling the specification elsewhere describes, is an express design incentive (avoid enzyme replenishment each cycle — the patent itself states the problem: heat denaturation "inactivate[s] the agent for polymerization… necessary to replenish the agent" every cycle).

Set F — Claims 6–9 (RNA template) — Set A + reverse transcriptase (known cDNA art)

The claimed RNA- and DNA-RNA-hybrid templates are reached by substituting a known enzyme (reverse transcriptase, listed in claim 15) for a known purpose.

Set G — Kit claims (whichever of 18–26 they are) — Set A + Falkow + routine packaging

A kit is obvious where the components are the claimed reagents in packaged multi-container form; In re packaging-type claims of this era required no more than assembling known reagents for their disclosed use.

---

5. Why a POSITA would have combined them — the motivation, generalized

1. Express textual suggestion (Kleppe p. 360: "the whole cycle could be repeated").
2. Solved a recognized problem with a known tool — amplification is a substitute for the admitted art's cloning route (Maniatis pp. 390–401; U.S. Pat. Nos. 4,416,988/4,403,036), with the patent itself listing the advantages (no organism propagation, no unrelated sequence synthesis, speed).
3. Predictable quantitative result. The exponential law follows from the cycle by inspection; the patent derives no new physics — it tabulates algebra.
4. Complete, off-the-shelf reagent set by March 1984 (synthetic oligomers per Beaucage & Caruthers 1981 and U.S. Pat. No. 4,458,066; polymerases; dNTPs; heat and helicase/RecA denaturation).
5. Detection side already solved (Falkow; Ward EP 63,879; Conner 1983; Leary 1983) and merely bolted on.
6. Optimization claims (DMSO, 35–40 °C, restriction-site primers, elevated temperature) are routine parameter tuning of a known reaction — KSR territory. (Caveat: I could not verify the prior-art status of DMSO as a nucleic-acid-reaction helix-destabilizer in this session; flagging as an unverified but plausible line.)

---

6. Where the obviousness case breaks down (the counter-case Cetus actually won)

A defensible memo must state these as directly as the affirmative case:

1. Kleppe's own data show linear, not exponential, amplification. Kleppe's Fig. 11 plateau at 8,000 cpm instead of the ~12,000 cpm exponential prediction — i.e., two new "hot" strands per cycle, consistent with replication of only the original template (EPO T 78/96, ¶ 17; T 216/96). The single best reference fails to show the property the claims are about.
2. Express teaching away. The Khorana grant progress report: "Although limited synthesis (two to three rounds) could be achieved, extensive and complete replication of the starting DNA could not be realized." That is an affirmative statement of failure by the most skilled group in the field. The EPO Board credited it: the Khorana group "had itself abandoned this line of research some ten years before the priority date."
3. Non-enablement cuts harder here than the § 102 defense shows. The references specify no primer sequences, no duplex sizes, no conditions — meaning the "motivation" is arguably to a hypothesis, not to a process with a reasonable expectation of success (In re O'Farrell).
4. Ten-year dormancy and failure of others. Nobody of ordinary skill, including Khorana's post-docs, practiced it for a decade — strong objective evidence.
5. The EPO found the combination non-obvious in T 78/96 and T 216/96: "The skilled person would not have rendered more complex an experiment that already did not work in its simpler version…" and would not have combined repair replication with DNA sequencing art (D72) as "a completely different field." Persuasive as comparative reasoning; not binding U.S. law — and decided under a more TSM-like framework than KSR.*
6. U.S. adjudications went Cetus's way. The PTO's 1990‑08‑23 decision rejected Kleppe/Panet & Khorana; the N.D. Cal. court denied du Pont's summary judgment on both anticipation and obviousness, holding "secondary considerations support a finding of nonobviousness," citing commercial success, long-felt need, and failure of others (1990 WL 305551, 19 U.S.P.Q.2d 1174).
7. Under KSR, however, points 1–4 are weakened substantially: "obvious to try" is now permissible where there is a finite number of identified, predictable solutions and a design incentive — which is precisely the profile here (denature, anneal, extend, repeat).

Bottom line of the merits analysis: on the KSR standard, each independent claim family has at least one combination with a clean textual motivation (Sets A, C, D, E are the strongest and are each internally consistent). Claim 1 and claim 15 are the most vulnerable — claim 1 because it lacks any cycle minimum or exponential requirement, claim 15 because the heat-stable-enzyme reference is the applicant's own citation. The sickle-cell claims (11, 12, 16) and the cloning claims are the most defensible, because they add a specific diagnostic target and a specific cloning architecture that the art did not marry to amplification. None of this, however, is actionable today — see below.

---

7. Practical posture (why this analysis cannot be operationalized)

• The patent expired 2005‑03‑28; § 286 bars recovery more than six years pre-complaint, so no damages window exists.
• No AIA trial can reach it in practice — consistent with the prior section's finding of zero IPR/PGR/CBM. (Note for completeness: the Board has in some cases instituted on expired patents, so "expired" is not itself a jurisdictional bar; the bar here is practical — there is no live claim to cancel and no infringement exposure to defend.)
• Therefore the deliverable is a validity-strength assessment, not a petition. If a demand letter or a license audit cites US 4,683,195, the answer is the expiry/§ 286 point, not the art in § 4 above.

---

8. Contradictions and corrections to flag against the earlier sections

1. Du Pont filing date — CONTRADICTION WITHIN MY OWN PRIOR SECTIONS. The "Litigation summary" says the DJ action was filed "on or about August 1, 1990." The "PTAB challenges" section says "1989‑08‑01." The primary document resolves it: "Shortly after du Pont filed its declaratory judgment action on August 1, 1989…" (CourtListener, N.D. Cal. opinion, 19 U.S.P.Q.2d 1174). The 1989 date is correct; the 1990 date in the litigation summary is an error. The same source confirms case number C 89‑2860 MHP and trial before Judge Marilyn Hall Patel — which the litigation section listed as "not confirmed." It is now confirmed.
2. Reexamination scope for '195 — OVERBROAD AS TO '195. The earlier sections say the PTO "upheld the validity of the '202 and '195 patents in their entirety" and that "all claims [were] confirmed." The record says the PTO ordered reexamination of claims 19–25 only of the '195 patent, and "found no new question of patentability with respect to claims 1–18 and 26, and declined to reexamine those claims." Therefore claims 1–18 and 26 of the '195 patent were never reexamined at all — including core claim 1. The "PTO confirmed all claims" framing should not be used for '195 claim 1; the PTO never passed on it.
3. Du Pont's § 103 target set on '195. Per the same opinion, du Pont challenged claims 1–4, 6, 11, 12, 14 and 15 of '195 for § 103 obviousness (and claims 1–4, 6–9, 11, 16, 19, 20 of '202 for § 102(b)). This is the only claim-by-claim road map of the '195 attack in my sources and it confirms the independent/dependent structure I assumed: claim 1 independent, claims 2–4 dependent, 6 and 11–15 dependent. It also implies claim 5 and claims 16–26 were not challenged in that motion.
4. Outcome of the DJ action — still not fully verified. The Jan. 1991 transcript excerpt (jury trial, C 89‑2860 MHP) contradicts the litigation section's clean "Feb. 27, 1991 ruling upheld validity, no infringement" narrative only in tone: a jury trial in January 1991 is consistent with a February 1991 judgment, but I have not seen the judgment or verdict form. Treat the disposition as probable but unverified.
5. The four-genus EPO decisions are being used comparatively, not as authority. T 78/96 (30.10.2003) and T 216/96 concern the European counterpart (EP 200 362 family). They are labeled as such throughout § 6.

---

9. Confidence ledger

Per the operating rule, no patent number above has been auto-corrected. US 4,683,202, US 4,889,818, US 4,358,535, EP 63,879, and the U.S. Pat. Nos. 4,293,652 / 4,356,270 / 4,403,036 / 4,416,988 / 4,458,066 / 4,582,789 / 4,617,261 are each reported as they appear in the source text.