idea-003 · desirability deck
Every slide states a condition that would have to be true, then reports where it stands using the candidate file's own claim and its own confidence tag. A deck never upgrades a tag, invents a number, or recommends anything, and there is no ask slide.
The chair: the user. For this product that is three people in one US laboratory running a tumor-informed ctDNA (circulating tumour DNA) minimal residual disease (MRD (molecular residual disease)) assay: the laboratory director who signs the limit-of-detection claim, the assay-development scientist who validates it, and the technologist who loads a control on every patient run-day. A fourth, smaller group is the assay developer that does not report patient results at all (a kit maker, a pharma or CRO assay group).
How to read this: every slide is a condition, not a conclusion. A bracketed claim reference points into knowledge-base/candidates/idea-003.md: the number is the claim's position in its ## Claims list, and the tag is copied from that line and never adjusted here. A [no claim] marker means nothing in the file speaks to the condition at all. See docs/deck-spec.md.
What changed since the 2026-09-15 build: that deck was built against the P0 the original shape shape, an RUO (research use only) reagent for assay development. The candidate is now at pivot P2 clinical QC material at launch. Its intended use, quoted from the file, is a "Quality control material ... to monitor the analytical performance ... of tumor-informed circulating tumor DNA minimal residual disease assays, during assay validation and in routine run-by-run quality control", supplied "as a multi-level panel plus a daily-use single-level control", with levels below 0.1% VAF (variant allele frequency) "labelled as derived by dilution rather than measured". So the user's moment is no longer only the validation study. It is now also every run-day. Rebuilt from the file, not patched.
Would have to be true: A US laboratory running a tumor-informed MRD assay on patient specimens would have to be meeting its daily control obligation and its limit-of-detection validation with in-house contrived material that its director can name as inadequate below 0.01% VAF, rather than with material it considers good enough.
Where it stands: The moment itself exists by regulation. CLIA (Clinical Laboratory Improvement Amendments) requires control procedures over the complete analytic process and, for each molecular amplification procedure, two control materials at least once each day patient specimens are assayed claim 7verified. A laboratory introducing a test not subject to FDA (US Food and Drug Administration) clearance must establish analytical sensitivity before reporting results claim 76verified. What the product needs on top of that is dissatisfaction, and that is on record and unchecked: that laboratories build in-house contrived controls and find them inadequate below 0.01% VAF claim 23unconfirmed. The customers' own assays report detection limits deep in that range, for example an LOD95 of 3.45 ppm for NeXT Personal claim 36unconfirmed and 0.008% VAF at 60 ng for the Invitae assay claim 37unconfirmed. The one verified figure of this kind is a heme assay at 120 ng of input, above this product's own assumed input range claim 127verified. The file also carries an argument for why a plasma-format control fits this moment better than a buffer one: a control supplied as purified DNA is added after extraction and so does not monitor it claim 183unconfirmed.
What would settle it: Structured customer-discovery interviews with directors of US laboratories running tumor-informed MRD assays, asking what they run as a daily control today and where it fails. The adoption connector refuses this class of claim by design, so a human has to run the interviews. claim 125unconfirmed's list of companies named in MolDX article A58376 is a starting list, though it is itself unconfirmed.
If it's false: If directors consider their in-house controls adequate, there is no moment to sell into. Slides 2 to 6 become moot, and the moat argument on viability slide 5 loses its referent.
Would have to be true: At least one in ten such laboratories would have to replace its in-house run control with this product's single-use daily control vial on every patient run-day, at a net $400-$900 per vial. That includes laboratories that could instead dilute the buffer-format Seraseq ctDNA MRD Panel Mix's 0.05% level into their own plasma at roughly $80 a run.
Where it stands: This is claim 177unconfirmed almost word for word, and it is unverifiable by design claim 177unconfirmed. The every-run-day cadence the revenue arithmetic assumes is stated as its own condition on viability slide 1 claim 187unconfirmed. For the panel rather than the vial, the older behaviour change still applies: a laboratory would have to anchor the limit-of-detection claim it publishes to an external supplier's lot claim 25unconfirmed.
What would settle it: A letter of intent or a paid pilot from one laboratory, or failing that, interview answers that name the control they would drop. Neither exists in the file.
If it's false: The daily control carries most of the revenue in the file's own planning arithmetic claim 178unconfirmed, and without it the product falls back to a one-time validation panel. Viability slide 3 loses its larger leg.
Would have to be true: The laboratory's CAP inspector, the MolDX technical-assessment reviewer who stands between the laboratory's test and Medicare coverage, and the laboratory's own QA would each have to accept third-party material with derived sub-0.1% values as limit-of-detection evidence. None of them could require something this product does not provide.
Where it stands: CLIA is silent in both directions. It names no required material, source or traceability condition for establishing analytical sensitivity claim 76verified, and its only traceability language is for calibrators, and only "if possible" claim 77verified. Every layer above CLIA is unread. CAP: COM.40400 requires an LoD and names no material claim 78unconfirmed, COM.40300 permits reference material as one option among several claim 79unconfirmed, and what the current MOL.36118 requires is open claim 144unconfirmed. Whether an inspector accepts a third-party panel at all is open claim 24unconfirmed. MolDX: L38779 (a Medicare local coverage determination number) requires a technical assessment claim 126unconfirmed, whose LoD-material criteria are unread claim 137unconfirmed, and the Palmetto FAQs on that assessment are not attached claim 165unconfirmed. The BLOODPAC consensus is claimed to require that contrived samples pass a functional characterization study before use in analytical validation claim 164unconfirmed. That cuts both ways: a product arriving with that study done would be worth more, and one without it would be one more thing the laboratory has to do. Inside the laboratory, whether a scientist can buy it as a consumable without a capital process is also open claim 26unconfirmed.
What would settle it: A human reading the current CAP Molecular Pathology checklist, the full text of the BLOODPAC paper (PMID (PubMed identifier) 41533993), the Palmetto GBA MolDX technical-assessment FAQs and the text of LCD (local coverage determination) L38779. None of these is reachable by a connector in this repo.
If it's false: If any of these reviewers requires material this product is not, the product is excluded at that gate. If all of them are silent, nobody outside the laboratory requires it, and the moat argument on viability slide 5 loses its strongest lever.
Would have to be true: The laboratory would have to accept three costs at once: a net price of $1,500-$3,000 per panel and $400-$900 per daily vial; a label that restricts every level below its precision threshold to qualitative detection only; and the one-time work of designing a patient-specific assay panel against the control's certified variant list.
Where it stands: The prices are claim 149unconfirmed, a willingness-to-pay claim that no connector can reach claim 149unconfirmed. The qualitative-only restriction is a design condition on the label: inter-run precision is claimed only where at least 100 molecules are expected per extraction claim 161unconfirmed. The panel-design burden follows from how tumor-informed assays work: an assay only sees the variants it was designed to track. The file puts the supply half on record, that each lot ships with tumor-proxy and normal-proxy DNA and that one external laboratory has designed a panel against a lot claim 162unconfirmed. Whether laboratories will do that routinely, and keep that panel in their QC (quality control) workflow, is a separate question this deck put on record claim 185unconfirmed. The continuity risk of a single external supplier is still open claim 25unconfirmed.
What would settle it: Interviews that put the price, the qualitative-only label and the panel-design step in front of the same director at once. Then a pilot in which one external laboratory designs a panel against a pilot lot, which would also settle the second half of claim 162unconfirmed.
If it's false: If laboratories will not design a panel against the control, the daily control cannot work with tumor-informed assays at all, because the assay will not detect variants it does not track. See feasibility slide 1.
Would have to be true: A laboratory director would have to believe that the native-plasma matrix behaves like a patient sample in their own extraction and assay. They would also have to accept values below 0.1% labelled as derived by dilution, not measured, as good enough for the limit-of-detection claim they publish.
Where it stands: The published record is mixed, and part of it cuts against the matrix premise as a selling point. In a study of six commercial materials, none resembled native cfDNA (cell-free DNA) in all metrics claim 87unconfirmed, but its authors concluded that similarity to native cfDNA mainly aids protocol development, and that well-defined variants are preferable for determining sensitivity and precision claim 90verified. A laboratory testing this product would test extraction first. Some kits preferentially recover spiked mononucleosomal DNA over endogenous cfDNA claim 92verified, and concordance on commercial contrived material falls at 0.1% VAF claim 94verified. Whether this product's commutability holds across assay chemistries claim 31unconfirmed and extraction methods claim 135unconfirmed is not demonstrated. Whether directors accept derived labelling at all is a condition this deck put on record claim 186unconfirmed. Whether they would pay more for an ISO (International Organization for Standardization) 17034 producer is open claim 129unconfirmed.
What would settle it: A commutability study against patient plasma at two or three laboratories using their own extraction and assay, and interview answers specifically on derived-by-dilution labelling.
If it's false: The premium over the buffer-format incumbent that viability depends on disappears claim 153unconfirmed, and the product is a more expensive Seraseq.
Would have to be true: When directors of US laboratories running tumor-informed MRD assays are asked, at least some would have to name the control they run today, describe where it fails, and say they would switch at the stated price. Hearing none do so would be the falsifying observation.
Where it stands: Nothing has been asked yet. Claims 23 and 177 state the answer the product needs, and both are unverifiable by design claim 23unconfirmed claim 177unconfirmed. The file records no interview, letter of intent or pilot with any laboratory [no claim].
What would settle it: Named customer-discovery interviews run by a human, recorded, and attached to the candidate so that claims 23, 177, 185 and 186 can be re-read against them.
If it's false: If no director names a failure of their current control, slides 1 to 5 are moot, and so is the daily-control half of the P2 clinical QC material at launch shape.
Every condition above with nothing verified behind the condition itself. Slides 1, 3 and 5 carry verified context (CLIA's control and sensitivity requirements, and three published findings), but in every case the condition the slide turns on rests on unconfirmed or unverified lines. Six of six conditions.
The rubric scores no factor for whether a user wants the product. A list this long is the expected result, and it says as much about the pipeline as about this idea.
If only one thing from this chair could be checked: whether at least one in ten US laboratories running tumor-informed MRD assays would use this daily control as their run-by-run external control, at $400-$900 per vial, in place of in-house material, including material they could make by diluting the buffer-format incumbent into their own plasma for roughly $80 a run claim 177unconfirmed.
This one rather than the others because P2 clinical QC material at launch moved the product's centre of gravity from a one-time validation panel to a consumable used every run-day. Every other desirability condition either feeds this one (slides 3 to 5) or matters only if it holds (slide 2's cadence). Naming it is not a recommendation, a gate, or a kill. It is the answer to "check what first."
_shared/plain-language.md