One of six NEJM papers · all six
An adenoviral antigen behind vaccine-induced thrombotic thrombocytopenia
Wang et al. 2026. Adenoviral Inciting Antigen and Somatic Hypermutation in VITT. New England Journal of Medicine, 10.1056/NEJMoa2514824. Cited by 20 works (20 October 2026). Drawn from the CC BY 4.0 author manuscript, PMC12900036, deposited under the DFG and others open-access condition; the version of record on nejm.org is not CC BY and was not used.
How to read these graphs
A discourse graph breaks a piece of research into typed one-sentence nodes and draws how they bear on one another. The grammar is that of Discourse Graphs (Chan and Akamatsu). Read each graph left to right: a question, the experiments that addressed it, the results they produced, and the hypothesis every result is drawn against.
- QUE Question
- What the work set out to answer.
- HYP Hypothesis
- The answer the authors expected. Every result earns one edge against it.
- EXP Experiment
- One attempt at the question: a trial arm, an assay, a survey, an analysis.
- RES Result
- One finding from an experiment, with where in the source it comes from.
- ISS Issue
- An open question or untried branch the work leaves for someone else.
- addresses
- An experiment addresses the question.
- produces
- An experiment produces a result.
- supports
- A positive result that backs the hypothesis.
- opposes
- A negative result that counts as evidence against the hypothesis, or a refutation of an earlier claim.
- informs
- A partial or mixed result that bears on the hypothesis without settling it.
- consistent with
- A negative or inconclusive result that does not count either way: no positive control passed, or there is no sensitivity or power statement, so the absence of an effect cannot be told from a measurement that failed.
- relation
- A thin dotted line: a link the source itself states between two experiments, such as one being derived from another.
A dashed box marks a result the rule above counts as uninformative, or an issue. In the drawing, click a node to jump to its text below.
Adenoviral inciting antigen and somatic hypermutation in VITT (NEJM, 2026)
Read left to right. The question (QUE) is addressed by 5 experiments (EXP), each producing results (RES), and each result bears on the hypothesis (HYP). 11 support it, 0 oppose it, 1 inform it. Issues (ISS) hang below: what the paper leaves open.
HYP, in the authors’ words (Introduction): “we hypothesized that anti-PF4 antibodies likely emerge from a memory anti-adenoviral response”
- supports positive finding
- opposes a negative with a control and a stated sensitivity, or a failed replication
- informs bears on the question without deciding the hypothesis
- thin grey a relation between two experiments
| Node | Text | Type | Outcome | Edge |
|---|---|---|---|---|
| EXP | Mass-spectrometric sequencing of serum anti-PF4 antibodies from 21 patients with VITT, with light-chain hypervariable gene sequencing in 100 patients and two pre-vaccination plasma samples. Every VITT anti-PF4 antibody sequenced carried a restricted IGLV3-21*02/*03 light chain with a shared somatic K31E mutation that was absent from germline DNA. | wet-lab | positive | |
| RES | All 21 patients had anti-PF4 antibodies with an IGLV3-21*02/*03 light chain, a heavy-chain ED motif and a shared lysine-to-glutamic-acid mutation at position 31. n=21 patients (18 after ChAdOx1 nCoV-19-S, 3 after Ad26.COV2.S); K31E present in all; somatic origin confirmed by germline IgG sequencing in 18 subjects with available DNA Results, Characterization of VITT antibodies; Table 1; Figure 1A · “a striking shared somatic mutation was observed in the light chains across different patients” | positiveinformative | supports HYP | |
| RES | The permissive IGLV3-21*02/*03 germline allele was found in 99 of 100 patients with VITT, against a stated background population frequency of 20 to 60 percent. 99/100 patients with VITT; expected background frequency 20 to 60% depending on ethnicity; confirmatory genome sequencing of an additional 82 patients Discussion; Results (confirmation in 82 additional patients) · “it was present in 99 of 100 patients with VITT (versus the expected background frequency of 20 to 60%” | positiveinformative | supports HYP | |
| RES | Pre-vaccination plasma from two people who later developed VITT held only non-pathogenic polyclonal anti-PF4 antibodies with germline K31, with one rare K31E variant at 2 percent of light-chain peptides. n=2 repository plasma donors; prototypical IGLV3-21*02 light chain with germline K31; K31E present in 2% of LCDR1 peptides in one donor; VITT onset day 5 and day 7 after vaccination Results, Characterization of VITT antibodies; Discussion · “a rare somatic variant with a K31E mutation was observed, present in 2% of the LCDR1 peptides” | positiveinformative | supports HYP | |
| EXP | Two recombinant VITT anti-PF4 IgG1 antibodies (CR22046, CR23004) compared in vitro with their germline back-mutants (E31K) and an allele swap (D50Y-CR22046) for PF4 binding and platelet activation. Reverting position 31 to germline lysine, or swapping the position-50 allele residue, stripped the antibodies of PF4 binding and of platelet-activating capacity. | wet-lab | positive | |
| RES | Both germline back-mutants lost most PF4 binding and needed very high concentrations to activate platelets, whereas the parent antibodies bound PF4 but not PF4-heparin complexes and activated platelets PF4-dependently. E31K-CR22046 and E31K-CR23004: major reduction in chemiluminescence PF4 binding; platelet activation only at antibody concentrations of 120 ug/mL or above Results, Assessment of somatic hypermutation; Figure 1C and 1D · “showed a major reduction of binding to PF4 ... required very high concentrations” | positiveinformative | supports HYP | |
| RES | Replacing the IGLV3-21*02/*03 aspartic acid at position 50 with the tyrosine encoded by IGLV3-21*01 or *04 abolished PF4 binding and platelet activation. D50Y-CR22046 versus CR22046: loss of PF4 binding and of platelet activation (no numeric value given in the text) Results, Assessment of somatic hypermutation; Figures 1C and 1D · “This resulted in loss of PF4 binding and platelet activation” | positiveinformative | supports HYP | |
| EXP | Recombinant CR22046 or its germline back-mutant given intravenously at 1.5 ug/g bodyweight/day for 5 days to mice transgenic for human PF4 and human FcgammaRIIa with mouse PF4 knocked out. The VITT antibody caused thrombocytopenia and thrombosis in most mice; the germline back-mutant largely did not. | wet-lab | positive | |
| RES | CR22046 produced a rapid, pronounced fall in platelet count and thromboses, including cerebral venous sinus and splanchnic vein thromboses, in 13 of 15 transgenic mice. platelet count change -79.3%; thrombosis in 13/15 mice (86.7%); dose 1.5 ug/g bodyweight/day for 5 consecutive days Results, Assessment of somatic hypermutation; Figures 1E and 1F · “a rapid and pronounced platelet count decrease (-79.3%) and thromboses in 13/15” | positiveinformative | supports HYP | |
| RES | The germline back-mutant caused thromboses in only 3 of 12 mice and left the platelet count profile indistinguishable from the negative control. E31K-CR22046: thrombosis in 3/12 mice; platelet count profile the same as the negative control Results, Assessment of somatic hypermutation; Figures 1E and 1F · “the back-mutated E31K-CR22046 caused thromboses only in 3/12 mice” | positiveinformative | supports HYP | |
| EXP | Search for the inciting antigen by affinity-purifying antibodies from VITT sera against intact ChAdOx1 virions and against five recombinant adenoviral proteins (penton, pIIIa, pV, pVI, pVII) and matching their clonotypes to the anti-PF4 fingerprint. Only antibodies purified against core protein VII contained a clonotype matching the patient's own anti-PF4 fingerprint; virions and the other four proteins did not. | wet-lab | partial | |
| RES | No light- or heavy-chain motif unique to pathogenic anti-PF4 clonotypes appeared in antibodies isolated on intact ChAdOx1 virions, and neither serum nor recombinant anti-PF4 antibodies bound coated ChAdOx1 or Ad26 virions. no matched motifs detected on intact virion-purified IgG; no binding of serum-derived or recombinant anti-PF4 antibodies to ChAdOx1 or Ad26 virion-coated ELISA plates (Figure S4); no sensitivity or sample-size statement given Results, Epitope mimicry between adenovirus core protein pVII and PF4; Figure S4 · “These findings argue against an inciting antigen on the adenoviral surface” | negative-no-associationinformative | informs HYP | |
| RES | IgG purified from four VITT sera against recombinant ChAdOx1 pVII was oligoclonal but contained a clonotypic species matching the anti-PF4 fingerprint of the same patient, while anti-penton, anti-pIIIa, anti-pV and anti-pVI did not. 4 VITT sera with matched anti-PF4 and anti-pVII clonotypes by heavy- and light-chain CDR3 barcodes; 4 control adenoviral proteins with no matching clonotype Results, Epitope mimicry; Figure 2A · “contained a clonotypic species that matched anti-PF4 antibody fingerprints from the same patient” | positiveinformative | supports HYP | |
| RES | Anti-PF4 and anti-pVII antibodies were reciprocally cross-reactive, the germline back-mutant lost PF4 reactivity while binding pVII more strongly, and a healthy vaccinee's anti-pVII antibodies were not cross-reactive with PF4. two-way cross-reactivity by ELISA for bead-purified anti-PF4, anti-pVII and recombinant anti-PF4 antibodies; E31K-CR22046 lost PF4 reactivity with enhanced pVII binding; healthy donor anti-pVII non-cross-reactive (no numeric values in the text) Results, Epitope mimicry; Figure 2B1, 2B2, 2B3 · “lost reactivity to PF4 while showing enhanced binding to pVII compared with its K31E counterpart” | positiveinformative | supports HYP | |
| EXP | Epitope mapping of VITT anti-PF4 antibodies on a library of 61 overlapping 15-mer peptides spanning ChAdOx1 pVII, with the Ad26 counterpart peptide for comparison. Binding localised to one basic 15-mer linear epitope in the middle of pVII, conserved between the chimpanzee and human adenovirus vectors. | wet-lab | positive | |
| RES | Anti-PF4 IgG purified from three VITT sera bound most strongly to a single 15-mer peptide, RYARAKSRRRRIARR, in the middle of ChAdOx1 pVII, and the recombinant antibodies reproduced this, with the back-mutant binding it more strongly. 61 15-mer peptides overlapping by 11 amino acids; strongest binding to peptide 32 (RYARAKSRRRRIARR); 3 VITT sera plus 2 recombinant antibodies; E31K-CR22046 showed increased binding to the 15-mer Results, Epitope mimicry; Figure 3A and 3B · “Strongest binding occurred to a 15-mer peptide reflecting a single linear epitope in the middle of the ChAdOx1 pVII” | positiveinformative | supports HYP | |
| RES | The epitope is nearly identical in ChAdOx1 and Ad26 pVII, and anti-PF4 antibodies from patients vaccinated with either vector bound both peptides with similar strength. peptide 32 nearly identical in ChAdOx1 and Ad26 pVII; similar binding strength for sera from VITT after ChAdOx1 (VITT1) and after Ad26.COV2.S (VITT38, VITT40); predicted alpha-helix analogous to the PF4 epitope Results, Epitope mimicry; Figure 3C and 3D; Figure S5 · “is highly conserved in ChAdOx1 and Ad26” | positiveinformative | supports HYP |
- ISS The human-specific IGLV3-21*02/*03 requirement rules out mechanistic studies of VITT induction in current animal models; transgenic mice carrying those alleles would be needed. (next-step · Discussion, limitations)check and design
- ISS Additional cofactors cannot be excluded: adenoviral nucleic acids, proteins associated with pVII, vector-expressed spike protein, or other genetic and environmental factors. (open-question · Discussion, limitations)check and design
- ISS Anti-PF4 disorders triggered by other viruses, such as cytomegalovirus, probably involve different viral proteins and different antibody binding epitopes, none of which are identified. (open-question · Discussion, limitations)check and design
- ISS Safer adenoviral vector vaccines might be built by replacing pVII with analogues that do not activate IGLV3-21*02/*03 B cells; this is proposed but untested. (next-step · Discussion, closing paragraph)check and design
Extraction notes: Publication year read as 2026 from the citation string in the stripped header (NEJM volume 394, pages 669 to 683); the header is run together and the month could not be read reliably. The paper is a mechanistic study, not a trial, although it carries trial registrations (EU PAS EUPAS45098, DRKS00025738) for the underlying patient cohort; no power or sample-size justification appears anywhere in the text, so no finding is given the valence 'opposes'. The virion result is classed negative-no-association with valence 'informs' rather than 'opposes' for two reasons: it bears against a surface antigen rather than against the paper's own hypothesis, and the text gives no sensitivity statement for the virion ELISA, so the negative rests on the parallel pVII positive signal acting as an internal positive comparator. The split into five experiments is my reading: the paper presents the work as three result sections, and I separated the mouse work from the in vitro recombinant antibody work, and the antigen screen from the peptide epitope mapping, because they are distinct attempts with distinct controls. Numbers for the D50Y allele swap, the reciprocal cross-reactivity ELISAs and the healthy-donor control are reported in the text only as presence or loss of binding with reference to figures, so no values could be copied. Table 1 reached the text as a run-together block; the per-patient ELISA optical densities are legible in it but were not used, since the Results prose carries the claims. Table 1 is listed as covering 21 patients but the block is truncated mid-row at patient VITT39, so the per-patient tallies by vaccine (18 ChAdOx1, 3 Ad26) are taken from the Results prose rather than counted from the table. The three healthy vaccinees with non-pathogenic anti-PF4 antibodies are described in the text as lacking both the allele and the mutation; I folded that control into the cross-reactivity finding rather than giving it a finding of its own, to stay within the finding budget. The authors state the pVII epitope is predicted to form an alpha-helix analogous to the PF4 epitope; that is a prediction (Figure S5), not a measurement.
Each open issue: checked, designed, and sent to a lab in mock
These pages are written for the next researcher who finds an open issue, not for the authors. Opening an issue shows, in order: whether anyone has already addressed it, what a registry of negative results holds on it, an experiment designed to close it with a control and a sensitivity statement, and what sending that experiment to a laboratory service would look like. Under each is a place to say what is wrong with it.
ISS The human-specific IGLV3-21*02/*03 requirement rules out mechanistic studies of VITT induction in current animal models; transgenic mice carrying those alleles would be needed. (next-step · follow-up: nothing found)
The authors say they cannot study how vaccine-induced immune thrombocytopenia and thrombosis (VITT) gets started, only what the finished antibody does, because the permissive immunoglobulin lambda variable 3-21 (IGLV3-21) alleles are human and no mouse carries them; they propose building a mouse that does.
1. Has anyone done this since?
- Status
- nothing found
- Works since
- None found that bears on this issue.
- Searched
- All 20 citing works in the supplied citing-works file were read by title, and Europe PMC was searched for IGLV3-21 together with transgenic mouse, for VITT together with mouse model or murine model, and for humanisation of the mouse immunoglobulin loci; no work reports a mouse carrying IGLV3-21*02 or *03, and the only VITT animal work found is the passive-transfer model the paper itself already uses.
- Still open
- Nobody has yet raised a VITT antibody from scratch in any animal, so the whole induction arm of the mechanism, boosting plus somatic hypermutation, remains untested outside human serum.
- Registry query
- 0 of 49 records in the open node mention any of: iglv3-21, adenoviral core protein vii, platelet factor 4, humanised transgenic mouse, somatic hypermutation, immunothrombosis. The open node holds biology and chemistry so far, so this is the expected answer; a private node of a clinical group would be asked the same question.
2. A proposed experiment
- Question
- Can a mouse carrying a human IGLV3-21*02 light-chain allele, human platelet factor 4 (PF4) and human Fc gamma receptor IIa develop platelet-activating anti-PF4 antibodies after priming with adenoviral core protein VII and boosting with an adenoviral vector?
- Hypothesis
- Priming and boosting against adenoviral core protein VII in mice that carry IGLV3-21*02, human PF4 and human Fc gamma receptor IIa will produce platelet-activating anti-PF4 antibodies bearing a position-31 lysine-to-glutamic-acid change.
- Attempt type
wet-lab- Design
- Cross a knock-in line carrying the human IGLV3-21*02 light-chain segment onto the hPF4-hFcgammaRIIa transgenic background used in the paper. Four arms run the same schedule: knock-in mice primed with recombinant ChAdOx1 core protein VII in adjuvant and boosted with the ChAdOx1 vector; knock-in mice primed with an irrelevant protein and boosted the same way; littermates carrying IGLV3-21*01 instead of *02; and hPF4-hFcgammaRIIa mice with the ordinary mouse light-chain repertoire. Read out, at each bleed and at necropsy, platelet count, anti-PF4 chemiluminescence, the PF4-induced platelet activation assay, thrombosis at unusual sites, and mass-spectrometric plus B-cell-receptor sequencing of the light chains to ask whether a position-31 acidic residue appears. Analysis compares the proportion of mice with thrombosis across arms and the proportion of sequenced clones carrying the mutation.
- Controls
designatedA passive-transfer arm giving recombinant CR22046 at 1.5 micrograms per gram bodyweight per day for 5 days, the schedule and dose from the paper.
Rules out: That the mouse background has lost the ability to thrombose at all, which would otherwise make any null in the immunisation arms unreadable.genetic-backgroundLittermates carrying the IGLV3-21*01 allele, whose position-50 tyrosine abolished PF4 binding in the paper's D50Y experiment, put through the identical immunisation schedule.
Rules out: That any anti-PF4 response seen is a generic consequence of adenoviral boosting rather than something the permissive allele allows.no-treatmentKnock-in mice primed with an irrelevant protein and boosted with the vector, plus unimmunised knock-in mice bled on the same days.
Rules out: Background anti-PF4 reactivity and platelet-count drift unrelated to the core protein VII response.
- Sensitivity
- Cannot be set from what is published for the immunisation arms, and that is the honest answer. The only effect size the record supplies is from passive transfer, 13 of 15 mice with thrombosis for CR22046 against 3 of 12 for the germline back-mutant, and that applies to the positive control arm alone. What is needed before group sizes can be fixed is a pilot estimate of how often immunised knock-in mice convert to a platelet-activating anti-PF4 response at all; in humans the event is of the order of one per hundred thousand vaccinations, and the mouse rate is unknown.
- Would count as
- Supported →
positive· opposed →negative-no-effect
Inconclusive if: The passive-transfer arm fails to produce thrombocytopenia and thrombosis, or the immunised mice raise no measurable anti-core-protein-VII antibody response at all, in which case nothing was tested about the hypermutation step. - Known before
- A mouse transgenic for human PF4 and human Fc gamma receptor IIa with mouse PF4 knocked out reproduces VITT when given a VITT antibody. (Mueller 2025, Research and Practice in Thrombosis and Haemostasis, doi:10.1016/j.rpth.2025.102925)
So: This is the background strain the knock-in must be crossed onto, and it already supplies the thrombosis readout, so only the light-chain allele has to be added. - Whole human immunoglobulin loci can be placed in mice, and such mice mount efficient antibody responses using human variable genes. (Lee 2014, Nature Biotechnology; Murphy 2014, Proceedings of the National Academy of Sciences, doi:10.1038/nbt.2825)
So: The humanisation technology exists, but these platforms carry one chosen set of human alleles, so the specific question of whether they carry IGLV3-21*02 or *03 has to be checked before building a new line. - IGLV3-21*01 is an inherited risk factor in chronic lymphocytic leukaemia through a single point mutation that enables autonomous B-cell receptor signalling. (Maity 2020, Proceedings of the National Academy of Sciences, doi:10.1073/pnas.1913810117)
So: The same gene with a different allele already has a known mutation-dependent phenotype, so the knock-in line may show B-cell effects of its own that have to be separated from the VITT readout. - The paper's own recombinant antibodies and their germline back-mutants are available and characterised for diagnostic and research use. (Mueller 2024, Vaccines; Wang 2024, Blood, doi:10.3390/vaccines13010003)
So: The positive control antibody does not have to be remade, which removes the main reagent risk from the control arm.
Searched: Europe PMC was searched for IGLV3-21 with transgenic mouse, for humanisation of mouse immunoglobulin loci, for VITT mouse and murine models, and the paper's own reference list was read for the strain and antibody sources; no published mouse carrying IGLV3-21*02 or *03 was found, and no attempt to induce VITT by immunisation in any species was found.
- A mouse transgenic for human PF4 and human Fc gamma receptor IIa with mouse PF4 knocked out reproduces VITT when given a VITT antibody. (Mueller 2025, Research and Practice in Thrombosis and Haemostasis, doi:10.1016/j.rpth.2025.102925)
- Feasibility
- A transfusion-medicine or platelet-immunology group with an animal facility, working with a mouse genome-engineering service and with the group holding the hPF4-hFcgammaRIIa line · 18 to 30 months, most of it line generation and breeding before any immunisation starts · high, because a new knock-in line plus crossing onto a double transgenic background plus multi-arm immunisation is a multi-year animal programme
Blockers: A knock-in carrying one human light-chain segment may not pair productively with mouse heavy chains, so no usable repertoire may emerge; Animal ethics approval for a study whose endpoint is thrombosis; Access to the hPF4-hFcgammaRIIa line, which the paper obtained from two named collaborators rather than a repository; Somatic hypermutation at one specific position may be too rare to observe in achievable cohort sizes - Strongest objection
- Putting one human light-chain allele into a mouse does not reconstitute the human germinal centre, the human anti-adenoviral memory pool, or the heavy chains that pair with that light chain to make the acidic paratope, so a mouse that never produces the position-31 mutation would say nothing about whether the mutation can be induced in people. A specialist would argue that the informative experiment is a human one: deep sequencing of the pVII-specific memory B-cell compartment in genotyped vaccinees.
- Not derivable from what is published
- Whether existing humanised immunoglobulin mouse platforms already carry IGLV3-21*02 or *03
- Adjuvant, priming dose and prime-to-boost interval for recombinant core protein VII
- The expected rate of conversion to a platelet-activating anti-PF4 response after immunisation, and therefore the cohort size
- Whether mouse heavy chains can form the acidic ED paratope the human antibodies use
3. Send it to a lab
- Send to a lab
- Serology and sequencing arm only: receive serial mouse serum and plasma, run anti-PF4 chemiluminescence and anti-core-protein-VII enzyme-linked immunosorbent assay (ELISA) in duplicate against VITT patient serum as the plate positive and naive mouse serum as the plate negative, run the PF4-induced platelet activation assay on donor platelets, then affinity-purify anti-PF4 immunoglobulin G and report light-chain sequences around position 31. The immunisation and necropsy cannot be quoted by a bench service and need an animal facility.
- The lab would need
- Serial serum and plasma from each arm, volumes and bleed days specified
- Recombinant ChAdOx1 core protein VII and the ChAdOx1 vector for the coating antigens
- VITT patient serum or recombinant CR22046 as the assay positive control
- Fresh donor platelets and ethics cover for the activation assay
- Animal ethics approval and the knock-in line, held by the requester
What this step would be: the design posted as a proposed job to a connected laboratory’s API, with the requester as approver and the result returned as a record. No lab is connected and no payment exists. The button says so.
Tell us what is wrong with it
Nothing here is sent anywhere by itself. The copy button puts your answers on the clipboard, and the email link opens your own mail client.
ISS Additional cofactors cannot be excluded: adenoviral nucleic acids, proteins associated with pVII, vector-expressed spike protein, or other genetic and environmental factors. (open-question · follow-up: partly addressed)
The authors have shown that core protein VII is sufficient to explain the antibody's specificity, but not that it is sufficient to drive the response, and they list the other vaccine and host components they could not rule out.
1. Has anyone done this since?
- Status
- partly addressed
- Works since
- Platelet factor 4 antibody persistence and long-term pathogenicity in vaccine-induced immune thrombotic thrombocytopenia (Kanack 2026, Journal of Thrombosis and Haemostasis, doi:10.1016/j.jtha.2026.03.025)
Followed 6 patients with Ad26.COV2.S-associated VITT to a median of 244 days and one ChAdOx1 patient beyond 4 years, found no new anti-PF4 clones after the acute phase, and reported one patient whose antibodies used a kappa rather than a lambda light chain, which means the restricted light-chain route is common but not universal and leaves room for a host factor the paper did not identify. - Lack of association between vaccine-induced immune thrombocytopenia and thrombosis and HLA loci in a large cohort (Schoenborn 2026, British Journal of Haematology, doi:10.1111/bjh.70419)
Reports no association between VITT and human leukocyte antigen loci in a large cohort, which closes off the most obvious additional genetic cofactor; it is not in the supplied citing-works file and shares authors with the paper, so it may be concurrent work rather than a follow-up.
- Platelet factor 4 antibody persistence and long-term pathogenicity in vaccine-induced immune thrombotic thrombocytopenia (Kanack 2026, Journal of Thrombosis and Haemostasis, doi:10.1016/j.jtha.2026.03.025)
- Searched
- The citing-works file was read in full, Europe PMC was searched for IGLV3-21 with VITT or platelet factor 4, for VITT with cofactors and with vaccine components, and the abstracts of the two New England Journal of Medicine citing items and three Journal of Thrombosis and Haemostasis citing items were retrieved.
- Still open
- No work has tested whether core protein VII on its own, without adenoviral DNA, without the proteins it travels with and without spike, is enough to boost an anti-core-protein-VII memory response into an anti-PF4 one.
- Registry query
- 0 of 49 records in the open node mention any of: adenoviral core protein vii, memory b cell, platelet factor 4, cofactor, adenoviral dna, spike protein. The open node holds biology and chemistry so far, so this is the expected answer; a private node of a clinical group would be asked the same question.
2. A proposed experiment
- Question
- Which components of an adenoviral vector are needed to boost IGLV3-21*02/*03 memory B cells specific for core protein VII into producing platelet-activating anti-PF4 antibodies?
- Hypothesis
- The core protein VII epitope alone is enough to boost memory B cells from IGLV3-21*02/*03 donors into secreting anti-PF4 reactive antibodies, and adding adenoviral DNA, associated core proteins or spike protein does not change the proportion that do.
- Attempt type
wet-lab- Design
- Genotype healthy blood donors from a standing repository for IGLV3-21 and for anti-adenovirus serostatus, and recruit donors who are IGLV3-21*02/*03 positive and adenovirus seropositive, with IGLV3-21*01/*04 donors as the comparison genotype. From peripheral blood mononuclear cells, sort memory B cells and culture them under a standard activation cocktail with one antigen condition per well: the 15-mer peptide RYARAKSRRRRIARR, recombinant core protein VII, core protein VII combined with adenoviral genomic DNA, intact ChAdOx1 vector, intact Ad26.COV2.S vector, a vector expressing spike alone, and no antigen. After culture, assay supernatants for anti-core-protein-VII and anti-PF4 immunoglobulin G by ELISA and chemiluminescence, run the PF4-induced platelet activation assay on positive supernatants, and sequence the light chains of responding wells to score position-31 acidic residues. Compare the proportion of donors with an anti-PF4 reactive supernatant across antigen conditions and genotypes.
- Controls
designatedVITT patient serum and recombinant CR22046 run on every anti-PF4 plate and in every platelet activation run.
Rules out: That a null across all antigen conditions is an assay failure rather than an absence of anti-PF4 production.internal-positiveA recall antigen each donor is known to respond to, run in parallel wells, to show the memory B-cell culture is capable of producing specific antibody from that donor's cells.
Rules out: That a donor's cells simply did not survive or respond in culture.no-treatmentDonors carrying IGLV3-21*01 or *04, and no-antigen wells from every donor.
Rules out: Anti-PF4 reactivity that arises from the culture conditions or from a genotype-independent polyclonal response.
Vocabulary note: The closest fitting kind, but this is really a donor group chosen for a genotype that should not permit the response, not an untreated arm.
- Sensitivity
- Cannot be derived from what is published. The design needs the frequency of core-protein-VII-specific memory B cells in an IGLV3-21*02/*03 donor and the proportion of wells that secrete detectable specific antibody in this culture format, and neither number appears in the paper or in the works found. A pilot on a small number of donors, reporting wells positive for anti-core-protein-VII per million memory B cells, would set the donor and well numbers needed to call a condition negative.
- Would count as
- Supported →
positive· opposed →negative-no-effect
Inconclusive if: The recall antigen control produces no specific antibody from a donor's cells, or no condition including the intact vectors yields any anti-core-protein-VII response, in which case the culture never reproduced a boost. - Known before
- VITT anti-PF4 antibodies do not cross-react with SARS-CoV-2 spike protein. (Greinacher 2021, Blood, doi:10.1182/blood.2021012938)
So: Spike is unlikely to be the mimicking antigen, so the spike-only condition tests whether spike contributes as an adjuvant-like cofactor rather than as a target. - ChAdOx1 virions bind PF4 electrostatically through the hexon surface, which was proposed as a route to VITT. (Baker 2021, Science Advances, doi:10.1126/sciadv.abl8213)
So: A competing surface-binding mechanism exists and the paper's virion results argue against it, so the intact-vector conditions must be compared against purified core protein VII to separate antigen presentation from charge-driven complexing. - Vaccine constituents including host cell proteins and ethylenediaminetetraacetic acid form complexes with PF4 and increase capillary leak, proposed as a contributor to the VITT immune response. (Greinacher 2021, Blood, doi:10.1182/blood.2021013231)
So: These are vaccine components rather than vector proteins, so a complete cofactor experiment would need a formulation arm, which this cell-culture design does not cover and should say so. - ChAdOx1 and Ad26.COV2.S differ in their protein content, including core protein amounts. (Michalik 2022, Haematologica, doi:10.3324/haematol.2021.280154)
So: The two intact-vector conditions are not interchangeable, and the comparison between them is itself informative about dose of core protein VII. - Most blood donors carry antibodies against many adenovirus types, measured longitudinally in a standing donor repository. (Wang 2024, Viruses, doi:10.3390/v16111747)
So: Adenovirus-seropositive donors are easy to find and can be selected on documented serostatus, so the memory compartment this design needs should be present in most recruited donors.
Searched: Europe PMC was searched for the spike cross-reactivity and vaccine-constituent papers by digital object identifier, for ChAdOx1 with platelet factor 4 and binding or hexon, and for VITT with cofactor terms; a search for an existing memory B-cell culture assay for anti-PF4 production found none, which is why the culture format here is specified only in outline.
- VITT anti-PF4 antibodies do not cross-react with SARS-CoV-2 spike protein. (Greinacher 2021, Blood, doi:10.1182/blood.2021012938)
- Feasibility
- A platelet immunology group with access to a genotyped blood donor repository and a B-cell culture facility, working with a vector supplier · 9 to 18 months · moderate, because the genotyping, donor recruitment and culture work are routine but the antigen panel and the sequencing readout are not
Blockers: Access to clinical-grade or research-grade ChAdOx1 and Ad26.COV2.S vector material; Ethics and consent for genotyping blood donors for an immunoglobulin allele; Memory B-cell culture may not reproduce a germinal-centre boost at all, which limits what a null means - Strongest objection
- A memory B-cell culture is not a germinal centre, and the somatic hypermutation this mechanism turns on happens in one. A specialist would say the assay can show which antigens recall an existing anti-core-protein-VII clone, which is worth knowing, but cannot show whether the position-31 mutation arises, so a null for the cofactor conditions would be a null about recall and not about the cofactor question as the authors posed it.
- Not derivable from what is published
- The memory B-cell culture format, cytokine cocktail and duration that would give a reliable antigen-specific readout here
- The frequency of core-protein-VII-specific memory B cells in IGLV3-21*02/*03 donors
- Whether research quantities of Ad26.COV2.S and a spike-only vector can be obtained now that the vaccines are withdrawn from several markets
- The adenoviral DNA preparation and amount that would represent the nucleic acid cofactor fairly
3. Send it to a lab
- Send to a lab
- Receive cryopreserved peripheral blood mononuclear cells from genotyped donors. Sort memory B cells, culture them in a standard activation cocktail across seven antigen conditions in replicate wells, then assay supernatants by anti-core-protein-VII and anti-PF4 ELISA plus anti-PF4 chemiluminescence, with VITT serum as the plate positive and no-antigen wells as the plate negative. Report per-well optical densities, flag wells above the plate negative, and return flagged supernatants frozen for platelet activation testing and light-chain sequencing.
- The lab would need
- Cryopreserved peripheral blood mononuclear cells with IGLV3-21 genotype and adenovirus serostatus for each donor
- Recombinant core protein VII, the 15-mer peptide, adenoviral genomic DNA, and ChAdOx1, Ad26.COV2.S and spike-only vector material
- VITT patient serum or recombinant CR22046 as the assay positive control
- The donor-specific recall antigen for the culture positive control
- Ethics approval and donor consent, held by the requester
What this step would be: the design posted as a proposed job to a connected laboratory’s API, with the requester as approver and the result returned as a record. No lab is connected and no payment exists. The button says so.
Tell us what is wrong with it
Nothing here is sent anywhere by itself. The copy button puts your answers on the clipboard, and the email link opens your own mail client.
ISS Anti-PF4 disorders triggered by other viruses, such as cytomegalovirus, probably involve different viral proteins and different antibody binding epitopes, none of which are identified. (open-question · follow-up: partly addressed)
The mechanism is shown for two adenoviruses, and the authors say plainly that a clinically similar disorder after other viral infections must have its own mimicking protein, which nobody has looked for.
1. Has anyone done this since?
- Status
- partly addressed
- Works since
- Human Adenovirus B3-Induced Immune Thrombocytopenia and Thrombosis (Irsara 2026, New England Journal of Medicine, doi:10.1056/nejmc2605960)
A correspondence item, with authors shared with the paper, extending the disorder to natural infection with human adenovirus B3; Europe PMC holds no abstract for it, so whether it tested the core protein VII epitope or the position-31 mutation could not be verified, and in any case it stays within the adenoviruses rather than reaching other viral families. - Platelet-Activating Anti-Platelet Factor 4 Disorders (Warkentin 2026, New England Journal of Medicine, doi:10.1056/nejmra2502459)
A review by two authors of the paper that classifies the anti-PF4 disorders and states that viral triggers other than adenovirus occur, but identifies no mimicking protein for any of them, so it frames the gap rather than filling it.
- Human Adenovirus B3-Induced Immune Thrombocytopenia and Thrombosis (Irsara 2026, New England Journal of Medicine, doi:10.1056/nejmc2605960)
- Searched
- The citing-works file was read in full, the two New England Journal of Medicine citing items were retrieved from Europe PMC, and Europe PMC was searched for anti-PF4 with cytomegalovirus and with viral infection and thrombocytopenia, and for anti-PF4 disorders without heparin exposure.
- Still open
- For every non-adenoviral trigger, including cytomegalovirus, the mimicking protein, the epitope and whether the antibody uses the same light-chain allele and the same position-31 change are all unknown.
- Registry query
- 0 of 49 records in the open node mention any of: anti-pf4 antibodies, cytomegalovirus, molecular mimicry, epitope mapping, clonotype matching, virus-associated thrombocytopenia. The open node holds biology and chemistry so far, so this is the expected answer; a private node of a clinical group would be asked the same question.
2. A proposed experiment
- Question
- In anti-PF4 disorders that follow a non-adenoviral viral infection, is there a mimicking viral protein, and do the antibodies carry an acidic residue at light-chain position 31 on an IGLV3-21*02/*03 background?
- Hypothesis
- Anti-PF4 antibodies from patients with a non-adenoviral virus-associated anti-PF4 disorder are clonally matched to antibodies against a basic, alpha-helical protein of that virus, and carry an acidic residue at light-chain position 31.
- Attempt type
mixed
Vocabulary note: Case ascertainment through a registry is observational, the epitope mapping is wet-lab, and the candidate antigen shortlist is computational; the three arms share one decision rule, which is why this is labelled mixed rather than split.- Design
- Through an anti-PF4 disorder registry and the reference laboratories that run platelet activation testing, ascertain patients with platelet-activating anti-PF4 antibodies and a documented recent non-adenoviral viral infection, cytomegalovirus first because a case is already published. For each patient, affinity-purify anti-PF4 immunoglobulin G, sequence heavy and light chains by mass spectrometry, and sequence germline light-chain genes from blood. In parallel, score the proteome of the implicated virus computationally for short basic stretches predicted to form alpha helices, as the core protein VII epitope does, and synthesise overlapping 15-mer peptides across the top candidates. Test the purified anti-PF4 antibodies against those peptides and against the recombinant candidate proteins, then test antibodies purified on the best candidate protein for a clonotype matching the same patient's anti-PF4 fingerprint. The decision rule is the paper's: a shared clonotype plus reciprocal cross-reactivity.
- Controls
reference-caseAn adenovirus-associated VITT serum with a known answer, run blind through the whole pipeline, which must recover both the core protein VII epitope and the position-31 change.
Rules out: That a failure to find a mimicking protein for a new virus is a failure of the pipeline rather than an absence of mimicry.no-treatmentSera from donors seropositive for the same virus who have no anti-PF4 disorder, carried through the peptide and protein binding assays.
Rules out: That binding to a basic viral peptide is a normal feature of the antiviral response rather than something specific to the patients.
Vocabulary note: The closest available kind, but this is a disease-free seropositive donor group, not an untreated arm; the vocabulary has no kind for an exposed but unaffected human comparator.backgroundNo-antibody wells, an irrelevant peptide of matched length and charge, and uncoated wells on every peptide plate.
Rules out: Charge-driven sticking of immunoglobulin to highly basic peptides, which is the obvious artefact for this epitope class.
- Sensitivity
- Cannot be powered, and should not pretend to be. The published base is a single cytomegalovirus-associated case, so the achievable sample is whatever a registry can ascertain over the study window. One patient with a clonally matched viral antibody would be hypothesis-generating rather than conclusive, and an absence across a handful of patients cannot distinguish no mimicry from too few cases. What is needed to set a real target is the ascertainment rate of non-adenoviral virus-associated anti-PF4 cases per year across the participating reference laboratories, which is not published.
- Would count as
- Supported →
positive· opposed →inconclusive-underpowered
Inconclusive if: The adenoviral reference case does not come back through the pipeline with its known epitope and mutation, or fewer than a handful of non-adenoviral cases can be ascertained, in which case the absence of a match carries no weight. - Known before
- Anti-PF4 mediated thrombocytopenia and thrombosis after acute cytomegalovirus infection shows both heparin-induced-thrombocytopenia-like and VITT-like features. (Nicolson 2025, British Journal of Haematology, doi:10.1111/bjh.20092)
So: This is the index case the design starts from, and its mixed serological phenotype is a warning that the antibody may not use the VITT light-chain route at all. - Anti-PF4 immunothrombosis occurs without proximate heparin or adenoviral vector exposure. (Schoenborn 2023, Blood, doi:10.1182/blood.2023022136)
So: A pool of cases without an obvious trigger already exists in reference laboratories, which is where ascertainment for this study should begin. - PF4-associated immune thrombocytopenia and thrombosis is broader than heparin and vaccines, and the naming and classification of triggers is unsettled. (Makris 2025, British Journal of Haematology, doi:10.1111/bjh.20125)
So: Case definitions differ between centres, so the ascertainment criteria must be stated in advance or the cohort will not be comparable. - Antibody epitopes on PF4 in VITT were mapped to a restricted region, and the mimicking adenoviral peptide is predicted to form an analogous alpha helix. (Huynh 2021, Nature, doi:10.1038/s41586-021-03744-4)
So: The PF4 side of the mimicry is already defined, which is what makes a structure-led shortlist of candidate viral peptides possible rather than a blind proteome screen. - Mass-spectrometric clonotype matching of two affinity-purified antibody populations is a workable way to demonstrate mimicry at the clonal level. (Wang 2024, New England Journal of Medicine; Wang 2022, Blood, doi:10.1056/NEJMc2402592)
So: The method transfers directly, so the new work is the candidate antigen shortlist rather than the assay.
Searched: Europe PMC was searched for anti-PF4 with cytomegalovirus, parvovirus and viral infection, for anti-PF4 disorders without heparin exposure, and for the cytomegalovirus index case and the PF4 epitope paper by digital object identifier; no study was found that searched any non-adenoviral viral proteome for a PF4-mimicking epitope.
- Anti-PF4 mediated thrombocytopenia and thrombosis after acute cytomegalovirus infection shows both heparin-induced-thrombocytopenia-like and VITT-like features. (Nicolson 2025, British Journal of Haematology, doi:10.1111/bjh.20092)
- Feasibility
- A reference laboratory network for anti-PF4 testing together with an antibody proteomics group and a structural bioinformatics collaborator · 24 to 36 months, limited by case ascertainment rather than by laboratory work · moderate, because the assays are established and the cost sits in multi-centre ascertainment and peptide synthesis
Blockers: Case numbers: non-adenoviral virus-associated anti-PF4 disorder is rare and the trigger is often not documented; Enough acute-phase serum per patient for affinity purification of two antibody populations; Consent for germline sequencing alongside the serological work; Attributing a virus at all, since serology may not establish which infection preceded the disorder - Strongest objection
- A specialist would object that the shortlist is built from one example of mimicry, a short polybasic alpha helix, so the screen can only find more of the same, and that polybasic viral peptides bind immunoglobulin non-specifically often enough that any hit needs the clonotype match to mean anything. They would also point out that the cytomegalovirus case already looks serologically mixed, so PF4 complexed with a polyanion, rather than a viral protein, may be the antigen there and the whole mimicry frame may not apply.
- Not derivable from what is published
- How many non-adenoviral virus-associated anti-PF4 cases are ascertainable per year
- Which viruses beyond cytomegalovirus should be in scope
- The structural criteria and thresholds for shortlisting candidate viral peptides
- Serum volume needed per patient for two affinity purifications plus peptide testing
3. Send it to a lab
- Send to a lab
- Receive patient serum and a shortlist of candidate viral proteins and peptides. Affinity-purify anti-PF4 immunoglobulin G on PF4-coupled beads and candidate-protein-specific immunoglobulin G on protein-coupled beads, digest and run liquid chromatography with tandem mass spectrometry, and report complementarity-determining-region 3 clonotype barcodes for both populations plus the residue at light-chain position 31. Run overlapping 15-mer peptide ELISA plates for each candidate protein, with the adenoviral reference serum, an irrelevant basic peptide and no-antibody wells on every plate.
- The lab would need
- Acute-phase patient serum in a volume sufficient for two affinity purifications
- Recombinant candidate viral proteins and overlapping 15-mer peptide libraries
- PF4 for bead coupling, and the adenoviral reference serum for the pipeline control
- Virus-seropositive, anti-PF4-negative donor sera for the comparator group
- Ethics approval, consent and the germline sequencing arm, held by the requester
What this step would be: the design posted as a proposed job to a connected laboratory’s API, with the requester as approver and the result returned as a record. No lab is connected and no payment exists. The button says so.
Tell us what is wrong with it
Nothing here is sent anywhere by itself. The copy button puts your answers on the clipboard, and the email link opens your own mail client.
ISS Safer adenoviral vector vaccines might be built by replacing pVII with analogues that do not activate IGLV3-21*02/*03 B cells; this is proposed but untested. (next-step · follow-up: nothing found)
This is the authors' own closing proposal, stated as a possibility and not tested anywhere in the paper; what they leave unspecified is whether the epitope can be changed at all without breaking a protein that the virus needs.
1. Has anyone done this since?
- Status
- nothing found
- Works since
- None found that bears on this issue.
- Searched
- The citing-works file was read in full; Europe PMC was searched for adenoviral vector vaccines with redesign, safer and next-generation terms together with VITT, for core protein VII with vector, capsid, mutant, deletion and engineered terms, and for core protein VII with packaging and chromatin; a web search for engineering or replacing core protein VII after this paper returned only commentary and institutional news items repeating the proposal.
- Still open
- Nobody has built or tested a vector whose core protein VII epitope is altered, so it is not known whether the epitope can be ablated at all, nor whether a vector that loses it still packages its genome, expresses its transgene and immunises.
- Registry query
- 0 of 49 records in the open node mention any of: adenoviral vector, core protein vii, epitope ablation, vaccine safety, genome packaging, platelet factor 4. The open node holds biology and chemistry so far, so this is the expected answer; a private node of a clinical group would be asked the same question.
2. A proposed experiment
- Question
- Can the RYARAKSRRRRIARR epitope in adenoviral core protein VII be mutated so that VITT antibodies no longer bind it, while the vector still packages its genome, expresses its transgene and raises an antibody response to it?
- Hypothesis
- There is at least one substitution in the core protein VII epitope that abolishes binding of VITT anti-PF4 antibodies while leaving virion yield, transgene expression and transgene immunogenicity indistinguishable from the parent vector.
- Attempt type
wet-lab- Design
- Build a substitution series across the 15-mer epitope in core protein VII of a replication-defective ChAdOx1-type vector, neutralising the basic residues singly and in blocks, plus a swap of the whole stretch for a charge-reduced analogue. For each construct measure, against the unmodified parent in the same run: binding of recombinant CR22046 and CR23004, of their germline back-mutants, and of VITT patient-purified anti-PF4 immunoglobulin G, by ELISA on the recombinant mutant protein and on the corresponding 15-mer peptides; virion yield per production run; genome packaging and transgene expression after infection of a permissive cell line; and, for constructs that pass, transgene antibody titres in mice on the schedule the vector literature uses. The decision rule is loss of VITT antibody binding to the plate background with the parent positive on the same plate, and packaging, expression and titre within the parent's run-to-run range.
- Controls
internal-positiveThe unmodified parent vector and its recombinant core protein VII on every binding plate and in every production run.
Rules out: That a loss of binding reflects a bad plate, a bad antibody lot or a failed production run rather than the substitution.reference-casePeptide 32 from the paper's own 61-peptide library as the known-positive peptide, with its Ad26 counterpart.
Rules out: That the peptide synthesis or coating step, rather than the mutation, explains a loss of signal.non-targeting-constructA vector carrying a point substitution elsewhere in core protein VII, outside the epitope, built and produced alongside the series.
Rules out: That any change in yield, packaging or expression is a generic consequence of mutating core protein VII at all.backgroundNo-antibody wells, an irrelevant basic peptide of matched length, and the germline back-mutant antibodies, which bind core protein VII more strongly and so mark the opposite direction of the specificity shift.
Rules out: Charge-driven non-specific immunoglobulin binding to a polybasic peptide.
- Sensitivity
- Not derivable as a numeric threshold from the published record, because the paper reports its binding assays graphically without optical density values or a limit of detection. What can be stated is the form the threshold must take: a substitution counts as having ablated the epitope only when VITT antibody binding falls to the no-antibody background while the parent construct on the same plate gives its usual signal, with the assay's own repeatability measured across replicate plates first. Setting the numeric cut-off needs the plate-level variability of the chemiluminescence and ELISA assays in the performing laboratory, which is a pilot measurement, not a literature value.
- Would count as
- Supported →
positive· opposed →negative-not-achievable
Inconclusive if: The parent vector or peptide 32 fails to be bound by the VITT antibodies in a run, or no construct in the series can be produced at all, so that binding was never testable on a packaged vector. - Known before
- The epitope is a short polybasic stretch in a protein that condenses and chaperones the viral genome and mediates its nuclear transport. (Kulanayake 2021, Viruses, doi:10.3390/v13030388)
So: The residues the design would neutralise are the ones that bind DNA, which makes loss of function the expected result and makes the negative outcome class a realistic one. - Adenovirus maturation establishes transcription-competent packaging of the genome, with core protein VII central to that state. (Baumgartl 2025, EMBO Reports, doi:10.1038/s44319-025-00598-z)
So: Packaging and early transcription are the readouts most likely to break, so both must be measured rather than assumed from virion yield alone. - Human adenovirus core protein VII antagonises the viral RNA sensor retinoic-acid-inducible gene I and so suppresses type I interferon production. (Yu 2026, Cell Reports, doi:10.1016/j.celrep.2026.117676)
So: Core protein VII has an innate-immune function of its own, so altering it could change vector immunogenicity for reasons unrelated to the epitope, which is why a transgene titre arm is needed. - Post-translational modifications on core protein VII matter during early infection. (Arnold 2025, Journal of Virology, doi:10.1128/jvi.01462-24)
So: Substitutions in this region may remove a modification site rather than only a binding surface, which is a second route to a broken vector and an alternative explanation for any loss of fitness. - Next-generation adenoviral vector vaccine development proceeds mainly by serotype selection, capsid modification and delivery changes. (Sampson 2025, Human Vaccines and Immunotherapeutics, doi:10.1080/21645515.2025.2514356)
So: Core protein engineering is not an established route, so this design should be compared against the cheaper alternative of choosing a serotype whose core protein VII lacks the epitope. - Recombinant adenovirus core protein VII can be expressed and purified to a published protocol. (Athukorala 2024, Journal of Virological Methods, doi:10.1016/j.jviromet.2024.114907)
So: The mutant proteins for the binding arm can be made without method development, so the binding screen can run ahead of, and gate, the vector building.
Searched: Europe PMC was searched for core protein VII with adenovirus in the title, for core protein VII with vector, capsid, mutant, deletion and engineered terms, for core protein VII with packaging and chromatin, and for next-generation adenoviral vector vaccines; a web search for replacing core protein VII to make safer vectors returned commentary on this paper, including a virology blog item of April 2026 and the authors' institutional news release, but no experimental work.
- The epitope is a short polybasic stretch in a protein that condenses and chaperones the viral genome and mediates its nuclear transport. (Kulanayake 2021, Viruses, doi:10.3390/v13030388)
- Feasibility
- An adenoviral vector laboratory or vaccine platform group with a platelet immunology collaborator for the antibody readouts · 12 to 24 months for the binding screen and vector production, longer if a mouse immunogenicity arm is included · high, because building and producing a substitution series of replication-defective vectors is expensive and most constructs are expected to fail
Blockers: Access to the ChAdOx1-type vector backbone, which is held commercially; Core protein VII is multifunctional and essential, so the mutable space may be empty; Core protein VII is both a particle component and a protein expressed from the delivered genome, so a mutation acts at two points at once; Supply of VITT patient-purified anti-PF4 immunoglobulin G, which is scarce - Strongest objection
- The epitope is the protein's polybasic DNA-binding core, so the most likely result is that every substitution that stops VITT antibodies binding also stops core protein VII condensing the genome, and the series yields no viable vector. A specialist would add that the cheaper and more likely route to a safer product is to pick an adenovirus serotype whose core protein VII does not carry this conserved stretch, or to leave the vector alone and screen recipients for the IGLV3-21 genotype, and would want the serotype survey done before any construct is built.
- Not derivable from what is published
- Which substitutions to try, since no mutational tolerance map of core protein VII in this region is published
- The numeric binding threshold that counts as ablation, because the paper's assays are reported graphically
- Whether any adenovirus serotype suitable for a vector lacks this conserved stretch
- The parent vector's run-to-run range for virion yield, packaging and transgene expression
- Whether the epitope is presented to B cells from the particle, from newly expressed protein, or both
3. Send it to a lab
- Send to a lab
- Express and purify the core protein VII substitution series to a published protocol, synthesise the matching 15-mer peptides, and run binding plates for recombinant CR22046, CR23004, their germline back-mutants and VITT patient-purified anti-PF4 immunoglobulin G, with the parent protein and peptide 32 as plate positives and no-antibody plus irrelevant basic peptide wells as negatives. Report per-construct optical densities with plate controls and replicate variability, and rank constructs by loss of VITT antibody binding. Vector production and animal work are separate and not part of this job.
- The lab would need
- The substitution list and expression constructs, or the parent core protein VII sequence to build them from
- Recombinant CR22046 and CR23004 with their germline back-mutants
- VITT patient-purified anti-PF4 immunoglobulin G, or VITT serum with PF4-coupled beads
- Synthesised 15-mer peptides including peptide 32 and its Ad26 counterpart
- The parent recombinant core protein VII as the plate positive
What this step would be: the design posted as a proposed job to a connected laboratory’s API, with the requester as approver and the result returned as a record. No lab is connected and no payment exists. The button says so.
Tell us what is wrong with it
Nothing here is sent anywhere by itself. The copy button puts your answers on the clipboard, and the email link opens your own mail client.
Designed on 21 October 2026 against outreach/nejm-experiments/brief.md. Grounded on: record, full-text, citing-works, prior-art. Full text: Europe PMC PMC12900036 fullTextXML, CC BY author manuscript. Designer notes: Follow-up status by issue: iss-1 not found, iss-2 partly addressed, iss-3 partly addressed, iss-4 not found. The citing-works file holds 20 works, all dated 2026, and most are reviews, commentaries or clinical descriptions rather than mechanistic follow-ups; abstracts were retrieved from Europe PMC for the five that looked relevant, and three of those five had no abstract deposited. Two placement calls worth checking: the British Journal of Haematology human leukocyte antigen paper is listed under iss-2 follow-ups although it is not in the citing-works file and shares authors with the paper, so it may be concurrent work, and the human adenovirus B3 correspondence is listed under iss-3 follow-ups although it stays within the adenoviruses and its content could not be verified without an abstract. Sensitivity could not be set numerically for any of the four designs, for three different reasons: no pilot conversion rate exists for the mouse and the B-cell culture designs, the case base for iss-3 is one published patient, and the paper reports its binding assays graphically with no optical density values or limit of detection. That last point is the single most limiting gap in the record. Vocabulary fit: the control kind list has no value for a human comparator group that is exposed but unaffected, which both iss-2 and iss-3 need, so no-treatment was used with a kindNote in each; iss-3 was labelled mixed because registry case ascertainment, mass spectrometry and a computational antigen shortlist share one decision rule, and none of the single attempt types covers that; iss-4 is labelled wet-lab although what is really being tested is whether an engineered reagent can exist, which is closer to a feasibility or constructability question than to a biological hypothesis, and the attempt type list has no value for that. All four designs are bench runnable in part, but the iss-1 job covers only the serology and sequencing arm and says so, since immunisation and necropsy need an animal facility. Numbers used in the designs are limited to those in the record: the 1.5 micrograms per gram per day dose for 5 days, the 13 of 15 against 3 of 12 thrombosis contrast, the 61-peptide library, and the peptide sequence RYARAKSRRRRIARR.
The other five
- BCG revaccination did not prevent sustained tuberculosis infection
- Balanced fluid versus saline in children with septic shock
- Neutralizing anti-interleukin-10 autoantibodies in inflammatory bowel disease
- Tumour-infiltrating clonal haematopoiesis
- Mitochondrial donation and preimplantation testing for mtDNA disease
Licence: the author manuscript of this paper is CC BY 4.0 in PubMed Central, so a graph of it is published here with attribution, exactly as the pilot’s records are. The NEJM version of record is under the publisher’s terms and nothing was taken from it; where the manuscript and the version of record differ, the manuscript is what is drawn. Method: one extraction pass against outreach/nejm-graphs/brief.md, every finding with a locator and a verbatim quote; the extraction notes above say what was inferred.