US 4941092

I'll verify the patent's current status and any litigation/docket records before summarizing.

US 4,941,092 — Verified Summary

Caveat on search coverage: I queried for both the patent and any CAFC 2026 docket activity. I found no CAFC 2026 docket, appeal, IPR, or district-court proceeding involving US 4,941,092. I cannot confirm from an authoritative docket source that none exists; I can only report that my searches returned none. The patent has been expired since its anticipated expiration date of 2007-07-10, which makes active 2026 litigation unlikely but not impossible (e.g., past-damages or license disputes).

Bibliographic data (per the Google Patents record provided)

Assignee note: The "Current Assignee" field on the Google Patents page lists Fujifilm Holdings Corp — this is a present-day corporate-successor listing, not an assignment of record in the prosecution history. The recorded assignment (Reel/Frame 004560/0833, effective 1986-05-19) is to Fuji Photo Film Co., Ltd.

Cited prior art of record (only 2 references): US 4,720,786 (Fuji Photo Film, "Method of compensating for offset distortion in rows of electrophoretic patterns") and EP 0155777 A1 (listed on the page under the title "Carrying handle and article having such a handle" — I am reporting this literally as it appears; the title appears anomalous for this field, but I will not auto-correct it).

Abstract (as issued)

A signal-processing method for determining nucleic-acid base sequence from digital signals corresponding to an autoradiograph of plural resolved rows (lanes) formed by resolving radioactively labeled base-specific DNA or RNA fragments on a support medium, comprising: (1) detecting at least two bands continuously in the lower part of each row and numbering consecutively from the lower end; (2) obtaining a correlation of the inter-band distance in the resolving direction with band number and predicting positions of undetected bands from that correlation; (3) detecting at least one band based on the predicted positions and numbering it; (4) recomputing the correlation using previously detected bands plus the newly detected band, and re-predicting; and (5) repeating (3) and (4) until all bands on the rows are detected.

Plain-language overview of the independent claims

Claim 1 — Iterative "predict-and-recheck" band detection (pitch-based).
The core method. Read the autoradiograph as digital data of multiple lanes (four sequencing lanes). Start at the bottom of each lane, where bands are well-separated, and find at least two bands in a row; number them 1, 2, … from the bottom up. Compute how the spacing between adjacent bands (the "pitch") relates to band number — in practice a straight-line (or curved) fit. Extrapolate that fit upward to predict where the next (higher, denser-packed) band should appear. Look at the lanes around that predicted position and assign the band to the lane that actually shows a signal there — this is how two or three fused/overlapping "combined bands" get split into separate calls. Then fold the newly found band into the data set, re-fit the spacing-vs-number relationship, and predict the next one. Repeat, one band at a time, until every band on the rows is accounted for. The lane (G, A, T, or C) on which each band lands, read bottom-to-top, gives the DNA sequence.

Claim 12 — Same iterative method, but keyed to migration distance instead of pitch.
Structurally identical to claim 1 (steps 1–5, same repeat-and-refit loop), differing in the correlation variable: instead of correlating the distance between bands with band number, it correlates the resolved/migration distance of the band itself with band number, and predicts the next band's absolute migration distance by extrapolating that fit. This is a parallel independent claim capturing the equivalent alternative implementation described in the specification.

Claim 23 — One-shot ("en bloc") version.
A trimmed-down method: detect and number at least two bands in the lower part of each row (step 1), compute the distance-vs-band-number correlation and predict undetected band positions from it (step 2), and then detect all remaining bands at once on the basis of those predictions (step 3) — no iterative re-fit loop. The specification describes this as the "third method," where the correlation is obtained once rather than sectionally.

Dependent claims worth flagging:

• Claims 2/13/24 — band detection by extracting a one-dimensional waveform along each lane and finding signal maxima.
• Claims 3/14/25 — the correlation is a regression line or regression curve, extrapolated to predict.
• Claims 4/15 — in the re-fit step, the lowest-numbered band is dropped so the fit window stays a constant size (keeps the relation linear/simple).
• Claims 5–6 / 16–17 — detection granularity: one band per iteration vs. a fixed batch (spec: 5–10 bands per operation, ≤ half the initial data set).
• Claims 7/18/27 — pre-step correction for the "smiling" phenomenon (lane-distance correction from band inclination).
• Claims 8/19/28 — pre-step correction for offset distortion between lanes.
• Claims 9/20/29 — the four-lane G/A/T/C embodiment.
• Claims 10/21/30 and 11/22/31 — data-acquisition alternatives: stimulable phosphor sheet read by stimulating rays, or conventional X-ray-film radiography read photoelectrically.

Text note: Claim 23 as rendered reads "detecting al bands on the resolved rows" — I am reproducing this literally rather than correcting it to "all bands."

Uncertainty flags

• I could not retrieve the USPTO PatentCenter/Public PAIR record or the printed claim set independently of the Google Patents text you supplied; claim text above follows your supplied authoritative document.
• One search hit referenced USPTO application 06/917,609 (Hara, filed 1986-10-10, Examiner Kim T. Bui) — that is a different application in the same family, not this patent's application (06/866,355). I mention it only to flag the distinction, per your instruction not to conflate similar numbers.
• No CAFC 2026 activity found for 4,941,092; absence of search results is not affirmative proof of no litigation.

I'll research litigation involving US Patent 4,941,092. Let me run several targeted searches.

Initial broad searches returned unrelated results. Let me query litigation-specific databases directly.

No litigation hits yet. Let me try more case-database-specific searches.

Summary answer

I found no litigation involving U.S. Patent No. 4,941,092. Based on the searches I was able to run, the '092 patent appears never to have been asserted in a U.S. district court or at the Federal Circuit, and I found no ITC, PTAB, or foreign proceedings naming it either.

Because the requested fields (plaintiff, defendant, jurisdiction, case number, filing date, outcome) presuppose at least one case, here is the honest status:

What I searched

• Broad queries for "US patent 4941092 litigation," "4,941,092" patent lawsuit, "4,941,092" infringement complaint district court, and "4941092" patent asserted defendant case number.
• Litigation-aggregator style queries (portal.unifiedpatents.com litigation 4941092, courtlistener, justia dockets).

The only hits that contained the literal string "4941092" were not litigation: an unrelated Kenya IEBC voter-registration PDF (a coincidental 7-digit ID) and a 1990 Biotechnology Advances volume reproducing the '092 patent's abstract in an issued-patent listing. Everything else returned by the "patent litigation" queries concerned unrelated patents with similar-looking numbers (e.g., Unwired Planet's '092 = U.S. 6,321,092; Catch Curve's '926 = U.S. 4,994,926) — I have excluded those, per your instruction not to conflate similar numbers.

Important caveats on this conclusion

1. Tool limits, not absence of a database, ended my search. I was cut off before I could directly hit PACER, CourtListener's RECAP docket search, Unified Patents' Docket Navigator/Portal caselist, or RPX Insight. My negative finding rests on general web search, which is not a substitute for a docket-level search. Treat this as "no publicly indexed litigation found," not as a certified clearance.
2. Docket databases have weak coverage for 1980s–1990s cases. Even if the '092 patent had been asserted in, say, 1991–1994, PACER coverage is sparse for that era and aggregators like Unified Patents/Docket Navigator largely focus on post-2000 NPE activity. A pre-1995 suit could plausibly exist and be invisible to my searches.
3. The patent's legal status makes litigation unlikely. The '092 patent issued 1990-07-10, and Google Patents lists its anticipated expiration as 2007-07-10 (status: "Expired – Lifetime"; last maintenance fee shown was the 12-year payment on 2001-09-19). Any infringement action would have had to be filed and pursued well before 2007, which further reduces the likelihood of findable records.
4. The claims are procedural/method claims for signal processing of autoradiographic band patterns — a laboratory-data-processing method, not a commercial product. That profile makes it an unattractive assertion target relative to the imaging-hardware patents in the same Fuji Photo Film family.
5. Assignee identity: original assignee Fuji Photo Film Co., Ltd.; Google Patents lists the current assignee as FujiFilm Holdings Corp. A corporate acquisition record is not litigation.
6. No PTAB/IPR or reexamination found. Note that IPRs were unavailable for pre-AIA patents of this vintage in practice; no reexamination certificate is indicated either.

Related context (not litigation)

• The '092 patent is one of a family of Fuji Photo Film signal-processing patents (e.g., the "similar documents" list includes U.S. 4,802,101; 4,894,786; 4,939,667; 4,951,559; 4,972,325; and EP 0,199,327 / EP 0,242,629). These are family members, not suits.
• The '092 patent's own cited prior art includes U.S. 4,720,786 (Fuji Photo Film, "Method of compensating for offset distortion in rows of electrophoretic patterns"), which corresponds to the offset-distortion correction discussed in the '092 specification at claims 8, 19, and 28. Citing a patent is not litigation.
• Patent-office assignment records show the inventors (Hara, Hashiue, Tanaka) assigning to Fuji Photo Film in 1986 — an assignment, not a dispute.

Recommended next steps (if you need certainty)

1. PACER docket search on U.S. Patent 4,941,092 across all district courts.
2. CourtListener/RECAP and Unified Patents Portal – Litigation (https://portal.unifiedpatents.com/litigation/caselist), filtered to patent number 4941092.
3. Docket Navigator (paid) litigation-by-patent search — the most complete source for this purpose.
4. USPTO Patent Assignment Search and USPTO Patent Trial and Appeal Board (PTAB) and Office of Enrollment records for any contested assignment or ethics proceedings (a dispute could exist even without an infringement suit).
5. Derwent LitAlert / Westlaw LitAlert, which are the standard sources for pre-2000 U.S. patent litigation that may not be in PACER.

Bottom line: On the evidence available to me, there is no known litigation involving U.S. Patent 4,941,092. I could not verify this against a complete docket database, so I would treat it as a well-supported negative finding rather than a definitive one.

I'll verify PTAB activity with targeted searches before writing this up.

The initial search returned "'092 patent" hits that are unrelated patents (an HVAC blower-motor patent and a cached-data-storage patent) — I need to confirm that. Let me run targeted queries on the actual patent.

Let me run a final pair of checks for any Federal Circuit appeal or AIA trial record referencing this specific patent.

PTAB / AIA Trial Analysis — US 4,941,092

Proceedings overview

Zero AIA trial proceedings are on file for US 4,941,092 — the structured "PTAB proceedings on file" block (USPTO Open Data Portal ingest) returns no IPR, PGR, or CBM, and my web searches surfaced none. Breakdown: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials. Defensive posture: there is no PTAB roadmap to hand you. The good news for a defendant is that this patent expired 2007-07-10 and is therefore, as a practical matter, unenforceable against future conduct — but you cannot say "claims 1–5 have been canceled," because no tribunal has ever touched any claim of this patent. Any validity attack would have to be built from scratch in district court, and any infringer's own IPR would have to be filed against an expired patent with no recoverable damages on the table.

No proceedings to itemize

Because the count is zero, there are no {PROCEEDING_NUMBER} — Petitioner v. Patent Owner entries to write. Reporting any proceeding number here would be fabrication. I am deliberately not doing that.

A critical data-integrity flag on my own searches. Querying "092 patent" returns a flood of PTAB petition text from different patents that share the "092" shorthand:

• An HVAC blower-motor / ECM patent (claims to "characteristics of an HVAC enclosure," "motor having multiple torque settings," Beifus as prior art) — this is a post-2015 patent, not a 1990 Fuji patent.
• A cached-data-storage patent (claims to "route the read command to the cache memory," Heil/Nakayama prior art; IPR2023-00121 and IPR2023-00122 were filed against it in the VARTA-related campaign) — again a completely different patent.

These are not US 4,941,092 and must not be conflated with it. They illustrate exactly the failure mode the instructions warn against: "092" is a three-digit fragment, not a patent identity. The only identifiers unique to this patent are US 4,941,092 / application 06/866,355 / Fuji Photo Film.

What my searches did and did not establish:

• ✅ No PTAB E2E / P-TACTS / PTAB-API record located for 4,941,092.
• ✅ No Federal Circuit or CourtListener opinion referencing 4,941,092 located.
• ⚠️ Absence of results is not affirmative proof of absence. I could not run a direct structured PTAB-API query with the tool available to me, and free-text search for an old patent number is weak evidence. Treat "no AIA trials" as high-confidence but not certified.

You can verify independently:

• P-TACTS (successor to PTAB E2E): https://ptacts.uspto.gov/ptacts/
• PTAB API v2 (bulk/structured): https://developer.uspto.gov/api-catalog/ptab-api-v2
• CourtListener docket search: https://www.courtlistener.com/?type=o&q=%224941092%22

Strategic summary

Claim status: entirely UNTESTED. Claims 1–31 stand exactly as issued on 1990-07-10. Not one claim has been canceled, disclaimed, or held unpatentable in any AIA trial. There is no FWD, no certificate canceling claims, and therefore no narrowed "surviving claim set" to point at. If someone asserts this patent today, every claim — including independent claims 1, 12, and 23 — is live-on-paper and has never been construed by the Board.

Estoppel landscape: empty. Section 315(e)(2) estoppel is petitioner-specific; with no petitioners, no estoppel bars anyone from raising any ground. Any prior art of record is thin — the patent cites only US 4,720,786 (Fuji's own offset-distortion patent) and EP 0155777 A1 as being of record. A defendant is free to run § 102/§ 103 on any reference, and § 101/§ 112 theories are likewise untested. Practically, though, the real defense is temporal: the patent issued 1990-07-10 and, under the pre-URAA 17-years-from-issue term, expired 2007-07-10. With no infringement possible after expiry, and 35 U.S.C. § 286 capping back-damages at six years before a complaint is filed, a suit filed in 2026 can recover essentially nothing regardless of validity. That makes an IPR an economically irrational expenditure for a defendant — you would be paying fees to invalidate a patent that can no longer be infringed.

Pattern signals: none — and the pattern is the story. There is no repeat petitioner, no serial IPR campaign, no Patent Owner appeal history at the CAFC, and no defensive-aggregator (Unified Patents, etc.) trail. That is expected for a 1985-priority, 2007-expired patent in a narrow, superseded technology (film/phosphor autoradiograph band-calling for Sanger/Maxam sequencing). Modern sequencing is fluorescence-based capillary electrophoresis; the commercial incentive to invalidate this patent evaporated when the market and the term both ended. The absence of IPRs is therefore not a signal that the patent is "hardened" — it is a signal that attacking it was never worth anyone's money.

Recommended next steps

1. If a demand letter or complaint cites US 4,941,092, lead with expiry, not invalidity. The patent's term ran 1990-07-10 → 2007-07-10 (pre-URAA, 17 years from issue). Confirm the maintenance-fee history in USPTO PatentCenter — the record shows fee payments at years 4, 8, and 12 (2001-09-19), with expiration thereafter; a lapse for non-payment of the year-16 fee would only shorten an already-dead term. Either way, no post-2007 infringement exists.
2. Invoke § 286 in any damages theory. 35 U.S.C. § 286 limits recovery to six years before filing. A plaintiff asserting this patent after 2013 has no compensable window. This alone should dispose of most monetization attempts.
3. If you nonetheless need an invalidity position (e.g., for a declaratory-judgment race or a licensing standoff about past conduct), start from the two of-record references — US 4,720,786 and EP 0155777 A1 — plus the sibling Fuji family members identified in the patent's own "Similar Documents" list (US 4,802,101; US 4,894,786; US 4,972,325; US 4,958,281). Note the prior summary's flag that EP 0155777 A1 is titled "Carrying handle and article having such a handle" on the Google Patents page — a title that is anomalous for this field and should be verified against the EPO register before relying on it. No estoppel constrains you, so the field is open.
4. If you want to close the verification gap, run the structured PTAB-API v2 query (/trials/proceedings filtered on patentNumber=4941092) and a CourtListener docket query, then confirm the negative. I could not complete that structured query with the search tool available, so I am reporting "no AIA trials on file" as high-confidence, unverified-by-structured-query, per the instruction not to overstate.

Bottom line for a defendant: this is an expired, never-challenged 1990 sequencing-algorithm patent. There is no IPR history to exploit and no IPR history to fear. Your motion to dismiss on term/§ 286 grounds is stronger than any PTAB strategy.

I'll pull the assignment-record evidence before writing anything.

Ownership & Assignment Reconstruction — US 4,941,092

Evidence basis and its limits. The only reel/frame-level record I could retrieve for this patent is the original 1986 assignment, as rendered in the Google Patents legal-events table supplied with this analysis (Reel 004560/0833). I could not reach a live USPTO Assignment Center record for 4,941,092 within this session; the searches I ran returned the Espacenet/INPADOC ST.27 feed, which gave a second Fujifilm reel/frame (018904/0001) — but that record is attached to a different Fujifilm application, not to 4,941,092 (details below). I mark that clearly rather than implying it covers this patent.

Inventors

Pattern check — no anomaly. The assignment was executed 1986-05-19 (recorded 1986-05-23), i.e. four days before the US filing date of 1986-05-23 and alongside the JP priority filings of the same day. That is a standard pre-filing "hereby sell, assign and transfer" employment assignment, not a departure signal. There is no evidence any of the three inventors left Fuji within 12 months of filing, and no inventor-held patent filings surfaced. Hashiue, in particular, reads as a career Fuji radiation-imaging/imaging-plate inventor: he also appears on Fuji's US 4,882,488 (radiation image read-out) and US 4,569,903 (optical data recording with silver halide media), i.e. the same stimulable-phosphor instrument division that produced this patent.

Original assignee

Fuji Photo Film Co., Ltd., No. 210 Nakanuma, Minami-Ashigara-shi, Kanagawa 250-01, Japan (assignee address of record per Reel 004560/0833).

• Primary line of business: photographic film, photographic chemistry, and — directly relevant here — X-ray/medical and industrial imaging equipment. The patent's own dependent claims 10/21/30 recite a divalent-europium-activated barium fluorobromide (BaFBr:Eu²⁺) stimulable phosphor sheet, i.e. the imaging plate that Fuji commercialized as the hardware core of its bio-imaging analyzer line for autoradiography.
• Did they ship a product embodying the claims? Partly, and on inference. Fuji indisputably shipped the data-acquisition platform (imaging-plate read-out instruments) that the dependent claims describe, and the patent's own specification describes reading sequencing autoradiographs off that platform rather than off X-ray film. Whether the specific claim-1/12/23 signal-processing steps (sectional regression on inter-band pitch, iterative re-prediction) shipped in a commercial analysis software build is not something I could verify and I will not assert it. Mark this as platform confirmed, method implementation unverified.
• Current status: Operating, now within a reorganized group. On 2006-10-01 Fuji Photo Film Co., Ltd. converted to a holding-company structure and was renamed FUJIFILM Holdings Corporation; the operating businesses were taken over by newly established FUJIFILM Corporation (26-30 Nishiazabu 2-chome, Minato-ku, Tokyo). There is no bankruptcy, no Chapter 7/11, and no fire-sale in this chain.

Contradiction to flag: the previously generated summary records Google Patents' "Current Assignee" field as Fujifilm Holdings Corp. That conflicts with the corporate record — the 2006 reorganization made Holdings a parent, and Fujifilm's own asset transfers of the period ran out of Holdings into FUJIFILM Corporation (see Reel 018904/0001 below). Google's assignee field is a present-day successor inference, not an assignment of record; the recorded chain, as far as it is documented, terminates at Fujifilm Corporation, not Holdings.

Assignment timeline

Chronological, with the verification status of each link stated explicitly.

1. 1986-05-19 (executed/effective) / recorded 1986-05-23 — Reel 004560/0833

• Conveyance: Assignment ("ASSIGNMENT OF ASSIGNORS INTEREST")
• Assignor: Hara, Makoto; Hashiue, Masakazu; Tanaka, Kazuyoshi (individual inventors)
• Assignee: Fuji Photo Film Co., Ltd., Japan
• Correspondent: Not retrievable. The Google Patents legal-events rendering supplies only assignor names, reel/frame, and effective date — it does not expose the correspondent. I could not obtain the attorney/agent of record for this recording and will not guess. (See note under Signals §3.)
• Context: Standard pre-filing employment assignment — inventors to employer, four days before the 1986-05-23 US filing.
• Verification: Confirmed in the supplied Google Patents record.

2. 2006-10-01 — corporate reorganization (effective), recording at USPTO unconfirmed

• Conveyance: Change of name and/or merger, not an arm's-length transfer of title
• Assignor: Fuji Photo Film Co., Ltd.
• Assignee: FUJIFILM Holdings Corporation (same legal entity, renamed; operating businesses moved to new FUJIFILM Corporation)
• Correspondent: Unknown for this patent.
• Context: Internal reorganization only. No consideration changes hands; economic ownership stays inside the Fujifilm group.
• Verification: Corporate fact confirmed (Fujifilm's own corporate history; ADR name-change notice effective 2006-10-02; contemporaneous trade press). Whether a name-change record was filed at USPTO against 4,941,092 specifically is not confirmed.

3. effective 2007-01-30 / recorded 2007-02-15 — Reel 018904/0001

• Conveyance: Assignment ("ASSIGNMENT OF ASSIGNORS INTEREST")
• Assignor: FUJIFILM HOLDINGS CORPORATION (FORMERLY FUJI PHOTO FILM CO., LTD.)
• Assignee: FUJIFILM CORPORATION, Japan
• Correspondent: Not established for this patent.
• Context: Internal group reorganization — the same pattern shown on the GB side, where Fujifilm Corporation replaced Fujifilm Holdings "by virtue of deed of assignment." Title moves down from the renamed parent to the operating company.
• Verification: ⚠️ This reel/frame is documented via Espacenet/INPADOC for a sibling Fujifilm matter (US 10/328,141, pub. US 2006/152562 A1), not for US 4,941,092. Fujifilm ran this as a bulk portfolio-wide recording, and it very probably sweeps in 4,941,092 — but "very probably" is not a recorded fact for this patent, and I am not entering it as one. Treat the current record owner of 4,941,092 as Fujifilm-group, entity-level status unconfirmed.

No further assignments of any kind. No security agreement, no license, no release, no correction, no transfer to a licensing entity. Google's legal events show only maintenance-fee payments (1994-01-10 year 4; 1998-01-12 year 8; 2000-12-09 payor-number entry; 2001-09-19 year 12), no post-grant conveyance other than the 1986 grant of the patent itself. The patent's term ran 17 years from issuance and lapsed at expiration on 2007-07-10, three weeks after the 2007-06-… filing cycle and months after the reorg recording — so the group still held it at expiry.

Timeline diagram

timeline
    title Ownership of US 4941092
    1986 : Filed by Fuji Photo Film
         : Reel 004560 frame 0833 assignment
    1990 : Patent issued
    2006 : Parent renamed Fujifilm Holdings
    2007 : Group asset transfer to Fujifilm Corp
         : Patent expires

(The 2007 asset transfer is charted for completeness of the corporate chain; as noted above it is documented on a sibling matter, not verified against this patent's own reel/frame.)

NPE / troll-pattern signals

Inverse-signal note. Signal 8's inverse is informative here: the patent lapsed for failure to reach a further term (expired 2007-07-10) rather than being sold into either an assertion vehicle or a defensive aggregator. Nothing about the asset was monetized in the secondary market at any point — which is itself consistent with a portfolio kept purely as internal product support.

Verdict

Insufficient data.

For US 4,941,092 itself, the only assignment I can verify at reel/frame level is the original inventor-to-employer recording, Reel 004560/0833, executed 1986-05-19 and recorded 1986-05-23 — the definition of "only the original assignment." The 2007 intra-group transfer is documented at Reel 018904/0001, effective 2007-01-30, recorded 2007-02-15, but on a sibling Fujifilm application, so the current record owner cannot be confirmed from this patent's own chain.

That said, nothing in the available record supports any NPE classification, and I want that stated plainly rather than hidden behind the "insufficient data" label: zero of the eight NPE signals are present, every assignee in every documented link is a named operating Japanese imaging company, and the asset was never asserted and expired on 2007-07-10. If the 2007 group recording is confirmed to cover this patent, the correct re-classification is Defensive / non-asserting (a never-asserted asset held to expiry by an operating group), not any NPE tier.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent number 4941092; the legacy interface at https://assignment.uspto.gov/patent/index.html queries the same records). Two specific items to pull when you do: (a) the correspondent of record on Reel 004560/0833, which my sources omit; and (b) whether Reel 018904/0001 — or any companion Fujifilm name-change recording effective 2006-10-01 — actually lists property US 4,941,092.

I'll verify the USPTO record and each cited reference directly rather than relying solely on the supplied text.

Let me verify the two cited references in more detail and check the specification-referenced documents.

Scope and method

I re-verified the target record against independent mirrors (Justia, FreePatentsOnline, uspto.report) rather than relying only on the Google Patents text you supplied. All three independently reproduce the same two "References Cited" items, so the citation list below is confirmed against the printed patent's own front-page art list:

• Justia "Referenced Cited" table — 4720786 | January 19, 1988 | Hara and 0155777 | August 1984 | EPX → https://patents.justia.com/patent/[4941092](/patent/4941092)
• FreePatentsOnline claim text → https://www.freepatentsonline.com/4941092.html
• uspto.report grant record → https://uspto.report/patent/grant/4941092

Important framing note before the §102 analysis. US 4,941,092 is a pre-AIA patent (US filing 1986-05-23, JP priority 1985-05-23, issued 1990-07-10). Its §102 analysis therefore runs under pre-AIA §§102(a), (b), (e), (g) — not the AIA's §§102(a)(1)/(a)(2). The relevant date arithmetic is set out in each reference's entry below.

The prior art of record for US 4,941,092 — exactly two references

Nothing else is "of record." There are no other U.S. patent citations and no non-patent-literature citations on the '092 front page.

Reference 1 — US 4,720,786 (the only substantively relevant citation)

Full citation: Hara, M., Method of compensating for offset distortion in rows of electrophoretic patterns, U.S. Patent 4,720,786, issued Jan. 19, 1988. Assignee: Fuji Photo Film Co., Ltd. (Google Patents: https://patents.google.com/patent/US4720786 ; claims and "Referenced Cited": https://patents.justia.com/patent/4720786)

Dates and the §102 timing problem (flagged explicitly):

• JP priority 1985-04-19 — roughly one month before '092's JP priority of 1985-05-23.
• Issued 1988-01-19 — after '092's 1986-05-23 U.S. filing. That means it is not available as a §102(a)/§102(b) "patent or printed publication" against '092.
• The only viable §102 route is §102(e) ("a patent granted on an application for patent by another filed in the United States before the invention thereof by the applicant").
• I could not confirm the actual U.S. filing date of the '786 application from the sources I was able to reach. This matters: under pre-AIA §102(e), the reference's prior-art date is its U.S. filing date, not its foreign priority date (In re Hilmer). If the '786 U.S. filing is ~April 1986 (i.e., after '092's ~May-1985 invention/priority date but before '092's own 1986-05-23 filing), the "before the invention by the applicant" prong may fail, and '786 would then be usable only under §103 (obviousness), not §102. I flag this as an unresolved date question rather than asserting §102(e) applies.
• "By another" is satisfied on its face: '786 names Hara alone; '092 names Hara, Hashiue, and Tanaka. Different inventive entities, so '786 is not disqualified on that ground (it is, however, commonly owned by Fuji, which bears on §103(c)-type considerations and double-patenting, not on §102).

Brief description of the disclosure ('786):
Directed to correcting offset distortion — the whole-lane positional shift caused by uneven slot geometry, inconsistent sample loading, or incomplete urea washout. The '786 method: detect at least two well-separated bands in the lower part of each lane and number them from the lower end; obtain a correlation between band number and migration/resolved distance for each lane (usually near-linear → straight-line regression; a regression curve where the relation is locally non-linear); take the difference in migration distance between lanes (e.g., as a difference in regression-line intercept) as the lane-to-lane locational deviation; and shift the lane waveforms accordingly. It expressly contemplates a sectional, iterative mode: correct a given section, then use the modified correlation to correct the next section, adding the newly detected band and excluding the lowest-numbered band to keep the data window constant.

Which '092 claims '786 potentially anticipates under §102 (with the dependency caveat applied):

'092's independent claims 1, 12, and 23 each require a prediction of undetected band positions followed by band detection on the basis of the predicted positions. '786 is framed around correcting lane positions, not predicting band locations to resolve fused bands. So whether '786 reaches the independent claims depends on how the Examiner read '786's upper-region band-detection step — I do not have the full '786 specification text and cannot assert that it expressly discloses "predicting positions of undetected bands." Treat the independent-claim mapping as possible, not established.

The dependent claims are where the mapping is clean and largely one-to-one:

Because claims 2–11 (and 13–22, 24–31) are dependent on claims 1/12/23, a §102 rejection over '786 would have to include the independent-claim elements as well. In practice the realistic exposure is (a) '786 as a §103 primary reference against claims 1/12/23 when combined with the combined-band problem, and (b) '786 as a §102 reference against the added subject matter of the dependent offset/inclination claims (7/8, 18/19, 27/28) — which, notably, are the very features '092's specification attributes to Fuji's co-pending Japanese applications 60(1985)-74899, 60(1985)-85275 and 60(1985)-85276, i.e., the '786 family.

Verification links used: Google Patents US4720786 (specification excerpt confirming the band-space/correlation approach and Figs. 4–6 regression lines/curves); Justia claims page (verbatim claim set 1–19).

Reference 2 — EP 0,155,777 A1 ("Carrying handle…")

Full citation: Erik Bock, Carrying handle and article having such a handle, European patent application EP 0 155 777 A1, priority 1984-02-29, published 1985-09-25. Applicant/inventor: Erik Bock. Google Patents: https://patents.google.com/patent/EP0155777A1/en

Brief description (as verified — reproduced literally, not auto-corrected):
A one-piece injection-moulded or extruded plastics carrying handle for containers such as buckets or pails. The handle has integral thinner-section folding lines/hinges adjacent each end portion, at least two folding lines at an angle to each other near each end, permitting the handle to swing between an upright carrying position and a folded-down position flush with the container side. Classification: Mechanical Engineering / B65D-type container hardware. The full text at the link is entirely about handle geometry — resilient plastics, folding lines, container-filling-line clearance.

§102 analysis:
None of '092's claims (1–31) can be anticipated by EP 0 155 777 A1. The reference is in an unrelated art (container handles) and discloses no autoradiography, no electrophoresis, no digital signal processing, and no band-detection or correlation method. There is no element overlap with any '092 claim.

Anomaly flag (per your instruction to interpret identifiers literally and not auto-correct):
EP 0 155 777 A1 is printed in the '092 front page art list, and both Google Patents and Justia independently reproduce "0155777" there. Yet the document it resolves to is a carrying-handle application by Erik Bock — a mismatch that strongly suggests a citation error in the printed '092 patent itself (a mis-transcribed EP number). I am reporting the number exactly as printed and noting that I did not locate a corrected or intended reference. If an EP number was garbled, the actually-intended reference is unknown from the record I could reach. (Also note the date discrepancy: Google gives publication 1985-09-25, while Justia's table renders it "August 1984" — I am reporting both literally.)

Documents the '092 specification relies on that are NOT prior art of record

These are cited in the body of '092 as background/incorporated material, not on the front page. They are not §102 art against '092 (same-applicant/co-pending), but they are the real technical neighbourhood of the invention and are worth having in the file:

• "U.S. Pat. No. 549,417" — cited in '092's Background section exactly as printed (six digits). I could not verify this number; as printed it is a six-digit string and may be a garbled rendering of a seven-digit patent number. I am not correcting it. I retrieved no record for a patent numbered 549,417 and am reporting the limitation explicitly rather than guessing.
• U.S. Application Ser. No. 664,405 and U.S. Application Ser. Nos. 568,877 and 730,034 — cited in '092 as the Fuji applications describing the stimulable-phosphor autoradiography and automatic DNA-sequence processing that '092 builds on.
• Related-generation documents in the German counterpart text (DE 3,781,307 T2, "Signal processing method for determining the basic structure of nucleic acids," https://patents.google.com/patent/DE3781307T2/en) instead recite Application Nos. 664,405 and 837,037, plus 917,606 and 917,609. Whether DE 3,781,307 T2 is the counterpart of '092 or of a sibling Fuji case, I could not confirm — flagging as unverified family attribution.
• Co-pending Japanese applications 60(1985)-74899 (smiling correction) and 60(1985)-85275 / 60(1985)-85276 (offset-distortion correction) — expressly identified in '092 as the source of the pre-correction techniques of claims 7/8, 18/19, 27/28. The offset-distortion pair corresponds to the US 4,720,786 family discussed above.

Bottom line

1. The single most relevant prior art for US 4,941,092 is US 4,720,786 (Hara / Fuji Photo Film, issued 1988-01-19, JP priority 1985-04-19). It is the same technical field, shares an inventor, and its claims map closely onto '092's dependent claims 2/3/7/8/9/10/11 (and the 13–22 and 24–31 counterparts), including the offset-distortion pre-correction that is the core of claims 8/19/28 and the iterative re-fit loop that resembles claims 1/12/23.
2. Whether '786 is §102 art at all is a live question. Because it issued after '092's filing, it can only be prior art under pre-AIA §102(e), whose date is '786's U.S. filing date and not its April-1985 JP priority date. I could not confirm '786's U.S. filing date, and if it postdates '092's invention date, §102(e) fails and '786 becomes a §103 reference only. I have flagged this rather than asserting anticipation.
3. EP 0 155 777 A1 appears to be an erroneous citation — it is a plastics carrying-handle application and anticipates no claim of '092.
4. There are no other patent citations of record for '092, so no other reference can be assessed as §102 art from the front page.

Caveats: (i) I was cut off by a tool step-limit before retrieving '786's U.S. filing date, the full '786 specification, or the '092 Image File Wrapper (examiner's actual §102/§103 rejections and any applicant remarks); a PatentCenter/Public PAIR look-up of application 06/866,355 and of '786 would resolve both the §102(e) date question and the true identity of the intended EP reference. (ii) Per your rule, I have reproduced "0155777," "549,417," and claim 23's "al bands" literally, without correction.

I'll verify the prior-art references from this page before building the §103 analysis, since dates and disclosures drive the whole comparison.

Key finding already: EP 0155777 A1 is genuinely a carrying-handle patent (Erik Bock, container handle) — the citation on the '092 record is anomalous. Let me pin down the dates and disclosures of the family references that matter for §103.

Obviousness Analysis — US 4,941,092 under 35 U.S.C. § 103

Note on the date discrepancy in the prompt: the task instructions state "Current Date: April 26, 2026," while the operating date for this session and the patent-record fetch is 2026-09-30. I flag this rather than silently picking one; nothing in the analysis below turns on the four-month difference.

Confirmation of a previously flagged item (no contradiction): The prior summary reported that the second reference of record, EP 0155777 A1, is listed under the title "Carrying handle and article having such a handle" and was described as "anomalous for this field." My verification confirms the anomaly is real, not a rendering error: EP 0155777 A1 (Erik Bock, priority 1984-02-29, published 1985-09-25) is genuinely a plastics container/bucket handle disclosure — integral hinges, folding lines, injection-moulded resilient strip (https://patents.google.com/patent/EP0155777A1/en). It has no relation to autoradiography, electrophoresis, or signal processing, and it cannot support any §103 combination against the '092 claims. Its presence on the face of the patent is a citation-record artifact.

1. Governing law and critical date

The application was filed 1986-05-23 with JP priority 1985-05-23. It is therefore governed by pre-AIA § 103. The critical date for art is at the latest 1985-05-23 (the foreign priority date, which the applicant can rely on for its own antedating purposes). Note the asymmetry that drives most of this analysis:

• For the applicant's own effective date, the foreign priority counts → 1985-05-23.
• For a reference patent's § 102(e) date pre-AIA, the foreign priority does not count — only the reference's actual U.S. filing date does (In re Hilmer, 359 F.2d 859 (CCPA 1966)).

This is not academic here: the one substantive reference of record claims a JP priority thirty-four days before the '092 priority date, but its § 102(e) date is its U.S. filing date, which I could not verify and which is almost certainly later than 1985-05-23. See §3.2.

2. Level of ordinary skill (PHOSITA)

A PHOSITA in May 1985 would hold a degree in analytical chemistry/biochemistry or in computer science, with ~2–4 years' experience in automated autoradiograph analysis, and would be familiar with: Maxam–Gilbert and Sanger–Coulson sequencing; gel electrophoresis band patterns; stimulable-phosphor autoradiography (Fuji's own platform); digitisation of autoradiographs into (x, y, z) signal matrices; one-dimensional waveform extraction and peak detection by sign-inversion of the first difference; and least-squares regression / polynomial curve fitting as standard numerical tools. Notably, this PHOSITA is not a hypothetical construct — on the '092 specification's own account, the entire Fuji signal-processing group (inventor Makoto Hara and colleagues) was working this exact problem in 1984–1986.

3. The prior art actually on the record of this patent

3.1 The only two references cited of record

3.2 US 4,720,786 — what it actually discloses

Fetched text (https://patents.google.com/patent/US4720786) shows it discloses, in the same words the '092 patent later uses:

• detection of "at least two bands in the lower part of each resolved row, and numbering all the bands consecutively from the lower end";
• "a certain correlation between the consecutive band's number and the resolved distance thereof for each row", which "is nearly linear so that the correlation can be approximated by a straight line";
• regression lines/curves of band's number (n) vs migration distance (y′) (its FIGS. 4–6);
• the explicit observation that bands are "sparse in the lower part … and become denser toward the resolution starting point in the upper part"; and
• a "third method" of successive correction, applied section-by-section because the relation "can be locally approximated by a straight line."

The gap: US 4,720,786 uses the band-number↔distance fit to compute and remove inter-lane offset distortion — it does not use the fit to predict the position of an undetected band within a lane in order to split fused bands. That is the only material difference between it and '092 claims 1/12/23.

The date problem: because it claims JP priority 1985-04-19 but issued only in January 1988, its § 102(a)/(b) dates are both after 1985-05-23, and under Hilmer its § 102(e) date is its U.S. filing date. I could not verify that U.S. filing date. If it is on or after 1985-05-23 (highly likely given a ~20-month pendency), US 4,720,786 is not prior art to the '092 patent at all and its appearance among the "Citations (2)" reflects an examiner/IDS citation of a co-pending sibling application, not an art rejection. I state this as a strong caution, not a verified conclusion.

3.3 The "Similar Documents" family — mostly not prior art

Structural finding: apart from US 4,720,786 (itself date-questionable under Hilmer) and the wholly irrelevant EP 0155777, nothing in the prior-art section of this record has a verified date before 1985-05-23. Every other listed document is either a Fuji sibling application published later in 1986–1990 or a third-party follower. A §103 rejection cannot be built on post-critical-date art, however close the subject matter.

3.4 What is usable: the applicant's own admissions

Pre-AIA law permits an applicant's specification to be used as an admission of the state of the art. The '092 specification contains several:

1. "there are proposed in U.S. Application Ser. Nos. 568,877 and 730,034 methods for automatically determining the base sequence of DNA by obtaining the autoradiograph as digital signals and subjecting the digital signals to appropriate signal processing" — an admission that automated digital signal-processing of sequencing autoradiographs was already known.
2. An admission that the "spaces between resolved bands are sparse in the lower part … and become denser toward the upper resolution-starting position" — i.e., the underlying physical regularity.
3. An admission that the "correlation between the bands' space … and the band's number can be locally represented by a straight line" — i.e., the fit is conventional.
4. An admission that the "separation of the bands is insufficient so that two or three bands are sometimes combined together" — i.e., the problem is an acknowledged, pre-existing one.

(Text note, per the "do not auto-correct" rule: the '092 specification cites "U.S. Pat. No. 549,417." A U.S. patent number in the 549,000s would issue c. 1896, which is plainly inconsistent with a stimulable-phosphor disclosure. I report the number literally as printed and flag it as a probable typographical error in the patent rather than correcting it.)

4. Element-by-element mapping of the independent claims

I use US 4,720,786 as primary reference ("R1"), the Hara/Fuji family (US 4,802,101; US 4,980,827; EP 0 199 327 B1; EP 0 242 629 B1) as secondary references ("R2"–"R5"), and applicant admissions ("AA"), purely for the sake of argument — i.e., assuming a challenger can establish the dates that I could not establish.

Claim 1 (iterative, pitch-based)

Claim 12 (iterative, resolved/migration-distance-based)

Identical to Claim 1 but keyed to absolute resolved distance rather than pitch. This is a closer match to R1 than Claim 1 is, because R1's correlation is literally "band's number (n) vs migration distance (y′)". The only remaining difference is the use of the fit (offset correction vs. position prediction).

Claim 23 (one-shot / en bloc)

Reduced to steps (1)–(3), with all bands detected at once. R1's first method already discloses doing the whole-row correction "en bloc," and R5 (EP 0 242 629 B1) discloses "detecting all bands on the resolved rows" by iteratively advancing a search-fiducial point and comparing a signal level to a threshold. The en bloc variant is the weakest independent claim on obviousness grounds, because step (3) requires no re-fitting at all.

5. Proposed combinations and the motivation to combine

Combination A (core): R1 (US 4,720,786) + AA.
Motivation: R1 and '092 address the same problem in the same art, from the same laboratory, and R1 expressly teaches that the band-number/migration-distance relation is (i) computable from the well-resolved lower region and (ii) "nearly linear." Once a PHOSITA has a fitted line through the lower-region data and the specification's own admission that the identical relationship holds locally in the upper region, extrapolating the line one band-step forward and searching for a signal near the extrapolated coordinate is the ordinary, predictable use of a regression fit — nothing more than the arithmetic of the fit. KSR Int'l v. Teleflex, 550 U.S. 398 (2007), at 417 ("a court must ask whether the improvement is more than the predictable use of prior art elements according to their established functions"). It is also, under the '092's own characterisation, the single predictable solution to the admitted "combined band" problem, which satisfies the "obvious to try" branch of KSR.

Combination B: A + R2 (US 4,802,101) + R3 (US 4,980,827).
Motivation: R2 supplies the mechanical step of "extracting digital signals along the resolving direction … and finding positions where the level … is maximum" (which is exactly dependent claims 2/13/24), and R3 supplies the lane-selection rule "compare the signal level at the maximum position … and determine that a band exists at the maximum position having the higher level," together with the explicit reason — the four base-specific groups are mutually exclusive, so a band cannot legitimately occupy two lanes. Both references are the same inventors, same assignee, same problem; combining them with a fitted-line prediction requires no change in the principle of operation of either.

Combination C: A + R5 (EP 0 242 629 B1).
Motivation: R5 teaches band detection by searching an interval for a peak exceeding a threshold and, where no peak is found, advancing by a given distance instead — which is the '092's fallback rule ("when a signal of the maximum level is not found out, the predicted position is decided to be the position of the next band"). R5 also derives the threshold from a running mean level, addressing the admitted Sanger–Coulson density gradient. This combination disposes of dependent claims 3/14/25 and 26.

Combination D: A + R4 (EP 0 199 327 B1) for the iterative loop.
Motivation: R4 discloses the repeat-and-refit structure — detect a band, add it to the data set, recompute the band-number↔distance correlation on the enlarged set, and repeat until all bands are found — as well as the two alternative refinements the '092 claims as dependents: a sectional (windowed) re-fit for the non-linear regime, and a straight-line local approximation. R4's stated reason for the sectional approach ("the correlation … can be locally approximated by a straight line and easily obtained when the bands employed therefor are not so many") is operationally identical to '092 dependent claims 5–6/16–17 (batch sizes of 5–10 bands, ≤ half the initial data) and to claim 4/15's constant-window re-fit.

Combination E: A + the applicant's own cross-referenced applications for the pre-processing corrections.
Motivation: The '092 specification expressly points to JP 60(1985)-74899 for the smiling correction and to JP 60(1985)-85275 / 85276 for the offset-distortion correction (these JP numbers correspond to the 1985-04-19 priority family of US 4,720,786 / EP 0 199 327). Dependent claims 7/18/27 (smiling) and 8/19/28 (offset) are therefore the applicant's own described pre-processing, pointed to by the applicant itself, and are obvious to practise in combination with the band-prediction step; the specification even concedes that "the separation of the combined bands is done while correcting the offset distortion, when the separation … is conducted sectionally."

Combination F: A + the admitted read-out art for claims 10/21/30 and 11/22/31.
Motivation: The '092 specification itself describes both acquisition routes (stimulable phosphor read by stimulating rays; X-ray film read photoelectrically) as conventional and cross-references U.S. App. Ser. No. 568,877 and Ser. No. 664,405 for them. These are the applicant's own admissions of convention and are not plausibly patentable differentiators.

Motivation-to-combine, generally: all references are in the same field of endeavour, address the same recognised deficiency (poor band resolution toward the top of the gel), and several are the work of the same inventor and assignee within a single multi-application programme — which is the strongest possible evidentiary basis for "the artisan would have looked to these teachings." There is no reference that teaches away from regression-based position prediction, and no change in principle of operation is required by any proposed combination.

6. Why a §103 case on this record is nonetheless weak — honest counter-assessment

1. The of-record art is thin and partly unusable. One cited reference is a bucket handle; the other is date-questionable under In re Hilmer. A challenger would need art outside this page to make a clean prima facie case.
2. The "prediction" step is where the references cut against the patentee only imperfectly. Every substantive reference locates bands by actually detecting a maximum or a threshold crossing. R1 uses its fit to correct geometry, never to posit a band at a location where no maximum exists. The '092 claims, by contrast, permit the predicted coordinate itself to be the band position when no maximum is found. A patentee can legitimately argue that the art's uniform insistence on detected peaks teaches away from the claimed fallback.
3. The extrapolation is into a regime the references say is non-linear. R1 and R4 both stress that the relation is only locally linear. A PHOSITA extrapolating a lower-region straight line into the dense upper region could be said to be courting exactly the error the art warns about — and the '092's own move to higher-order fits for the upper region (claim 3/14/25) concedes the point.
4. Ordinary skill vs. invention. The strongest patentee rebuttal is that the art treated the fit as a diagnostic of lane geometry, and the re-characterisation of the fit as a forward predictor / band splitter is a genuine change in use. Under pre-AIA practice this is a real, though not overwhelming, non-obviousness argument.
5. Secondary considerations are unattested. No evidence of unexpected results, long-felt but unsolved need, industry scepticism, or commercial nexus appears on this record. Licensing or adoption (if any) would need to be shown to be because of the claimed method rather than the underlying phosphor-imaging platform.
6. No verified litigation exists against which to test the claims (consistent with the litigation section already generated), so there is no judicial construction of "lower part," "correlation," "predicted positions," or "combined bands" to work from.

7. Alternative invalidity angles (not §103, but decisive if available)

• Obviousness-type double patenting (OTDP). US 4,720,786 and the '092 share inventor Hara, assignee Fuji Photo Film, and overlapping disclosures. Where a patent applicant's own earlier-issued or co-pending claim sets differ only in an obvious variation (fit-as-corrector vs. fit-as-predictor, or distance-vs-pitch), OTDP — including over a same-family continuation — is often the cleanest attack. Because the '092 was filed in 1986, pre-AIA § 103(c)'s common-ownership safe harbour is not available (it applies only to applications filed on or after 1999-11-29).
• § 112(b) indefiniteness. "the lower part" is defined in the specification only as "a region below nearly the middle of the support medium" — a relative, non-numerical boundary. Claim 23's "detecting al bands" (reported literally, per the no-auto-correct rule) is a typographical defect that would be construed, but it illustrates the claim-drafting looseness.
• § 112(a) written-description breadth. Claim 23's en bloc detection across all bands, without any stated mechanism for handling the admitted non-linearity, may exceed what the specification enables.

8. Bottom line

• On the prior art appearing in this patent's own record, a § 103 rejection of claims 1, 12 and 23 is not sustainable. EP 0155777 A1 is non-analogous (a container handle) and US 4,720,786 is (a) different in the one respect that matters — it uses the band-number/distance regression for inter-lane offset correction rather than intra-lane position prediction — and (b) very probably not even prior art given its 1985-04-19 JP priority versus its later U.S. filing date (In re Hilmer). Every other listed reference postdates the 1985-05-23 critical date.
• Assuming the date obstacles can be cured, the strongest combinations are: US 4,720,786 + the Hara/Fuji family (US 4,802,101, US 4,980,827, EP 0 199 327 B1, EP 0 242 629 B1) + the applicant's own admissions. The motivation is unusually strong because the references are the same artisans addressing the same admitted problem, and because the '092 specification itself supplies the two key premises of the obviousness case: that band spacing vs. band number is locally linear, and that fused bands in the upper region are the recognised deficiency. Converting a fitted line into a one-step-ahead position prediction is a predictable use of a known numerical technique (KSR), and the iterative re-fit is squarely disclosed by EP 0 199 327 B1's repeating embodiment.
• The best patentee defences are the Hilmer date defect, the non-analogousness of the sole other cited reference, and a teaching-away argument from the art's uniform reliance on detected peaks rather than predicted positions.

Uncertainties I could not resolve: the actual U.S. filing dates of US 4,720,786 and US 4,802,101; whether the "third embodiment" iterative language I attribute to EP 0 199 327 B1 also appears verbatim in US 4,720,786; the content of U.S. App. Ser. Nos. 568,877, 664,405 and 730,034 (cited in the specification but not on the patent's face); and the literal identity of "U.S. Pat. No. 549,417." Searches were terminated by step limits before these could be confirmed, so each statement above tied to those items should be treated as a working hypothesis rather than a verified fact.