{"id":"b7fa6351-e948-43e2-8b91-b694e71ec20b","arxiv_id":"2502.05388","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A single DNA origami triangular prism core with over 70% conserved staples can be programmed via overhang lengths and sequences to assemble shells and a 96-triangle toroid with nonuniform Gaussian curvature.","lead":"This paper introduces a modular DNA origami building block with a conserved scaffold core and swappable interface staples, allowing many different interaction types and binding angles to be programmed without redesigning the whole structure. Using this approach, the authors assemble closed shells and a hollow toroid with spatially varying curvature, a step toward complex curved nanostructures.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The toroid's positive-angle design silently switches from the reported 6.65°/bp calibration to an unexplained 7.18°/bp slope, without which the 109.5° target would fall outside the allowed 0–15 bp range.","rationale":"The reader's weakest assumption identifies the same load-bearing issue: the empirical angle calibration is assumed to transfer reliably to all multispecies assemblies, and the unexplained switch to 7.18°/bp is a concrete red flag. My stress-test sharpens this into a specific numerical consequence: the flagship toroid requires a 109.5° positive angle, which is not reachable under the main-text 6.65°/bp calibration within the stated 0–15 bp angle-domain range. The paper's own oxDNA result for +15δ (99.2°) independently suggests that the 7.18°/bp slope is not supported by the dimer-level data. If the true angles are ~10° off at the most curved toroid junctions, then the closed toroids observed by TEM could form because the flexible bonds accommodate errors, not because the geometric specificity is precisely programmed. This does not refute the qualitative achievement of assembling a toroid, but it does undermine the quantitative claim of precise, independent programming. The reader's CONDITIONAL verdict already reflects this concern, so I do not propose a different verdict. I also credit the paper's substantial independent evidence: cryo-EM reconstructions and multibody analysis, gel-based selectivity measurements, oxDNA simulations, and TEM tomography of the toroid. The concern is specifically about the calibration transfer step, which is the least secure link in the central argument.","tokens_in":32124,"tokens_out":8261,"duration_ms":95630,"concrete_test":"Recompute the toroid edge angles from SI Tables II–III and Fig. S20C using both the reported 6.65°/bp and 7.18°/bp positive slopes, then run oxDNA dimer simulations for the exact toroid interface staple pairs, especially the +15δ pair assigned to the 109.5° target. If the +15δ pair equilibrates near 99–100°, as the paper's own oxDNA data in SI Section V suggests, the 109.5° toroid bonds are not being programmed to their target values and the closed toroids demonstrate tolerance rather than precision. If the same pair equilibrates near 109.5°, the 7.18°/bp calibration is independently supported and the switch is justified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"SI Section IX states that all angle-domain lengths for the non-Platonic shells and toroid are derived from the Fig. 2L linear fit, 'except we use a slope of 7.18 degrees per base-pair for the positive binding angles.' No derivation or justification is given for 7.18, and it conflicts with the 6.65°/bp positive slope reported in the main text. This is not cosmetic: the toroid's largest positive binding angle is 109.5°, which at 6.65°/bp would require nδ ≈ 16.5 bp, beyond the stated maximum of 15 bp; at 7.18°/bp it falls at roughly the 15 bp maximum. The paper's own oxDNA simulation of a +15δ dimer (SI Section V) gives a mean angle of 99.2°, close to the 6.65°/bp prediction and ~10° below the 109.5° target. Thus the headline toroid appears to be nominally encodable only by abandoning the measured calibration, and the closed toroids may assemble despite, not because of, the programmed angles. The concern is amplified by the paper's own documented deviations from the rigid-rod model: extrusion-angle bias, insufficient two-thymine spacers, and overhang bending (SI Sections II and V). If the true positive-angle slope is near 6.65°/bp, then the +109.5° toroid bonds actually form near 100°, and toroid closure is achieved through the same flexibility that produces the observed polymorphism and 21% off-target 5-fold toroids. In that case, the observation of complete toroids does not validate 'precise and independent programming' of binding angles; it validates assembly tolerance to large angular errors.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper introduces a modular DNA origami building block: a hollow right equilateral triangular prism with a conserved scaffold routing and more than 70% shared core staples. Interaction specificity is programmed through bond-domain sequences in variable interface staples, and binding angles are programmed through the length difference of double-stranded angle domains. The authors characterize binding angles with cryo-EM dimer reconstructions, gel electrophoresis of vertices and rings interpreted with a thermodynamic model, and oxDNA simulations. They then demonstrate self-closing shells, including four non-Platonic shells with improved yields, and a 96-triangle toroid with 12 species, 36 interactions, and 7 programmed angles, which is presented as the key example of a structure with spatially varying Gaussian curvature.","tokens_in":32432,"tokens_out":7347,"duration_ms":74578,"significance":"If the central claim holds, the paper would be a valuable step for DNA origami: an economical, single-scaffold route to complex 2D manifolds with nonuniform curvature, with interaction and angle specification separated into independent overhang domains. The strengths include a parameter-free geometric model (Eq. 1), multiple complementary characterization methods, deposited cryo-EM maps (EMD-48565 to EMD-48569), and an unusually candid reporting of discrepancies, including the four-fold bending-modulus mismatch between oxDNA and cryo-EM and the low toroid yield. However, the toroid demonstration currently rests on an unexplained change in the angle calibration, so the precision of the headline result is not yet established.","major_comments":[{"comment":"The toroid's positive-angle encoding relies on a calibration slope that the main text does not report and the manuscript does not justify. SI Section IX states that all angle-domain lengths for the non-Platonic shells and toroids are derived from the linear fit 'except we use a slope of 7.18 degrees per base-pair for the positive binding angles,' while the main-text calibration (Fig. 2L) reports 6.65 deg/bp for positive angles. This is load-bearing: the toroid requires a positive binding angle of 109.5 deg (main text Section D), which at 6.65 deg/bp would require n_delta of about 16.5 bp, beyond the stated 0-15 bp maximum, whereas at 7.18 deg/bp it falls near the 15 bp maximum. The authors' own oxDNA simulation of a +15-delta dimer gives a mean angle of 99.2 deg (SI Section V), close to the 6.65 deg/bp prediction and about 10 deg below the 109.5 deg target. No measurement, derivation, or error analysis is supplied for the 7.18 deg/bp value. Please either justify this slope with data, show that toroid closure is robust to the choice of slope, or revise the claim that the toroid validates precise angle programming.","section":"SI Section IX (Fig. S20 caption) and main text Section D"},{"comment":"The empirical angle calibration is model-dependent, and the manuscript does not propagate the resulting uncertainty into the toroid and shell designs. The inferred positive-angle slopes vary with the assumed bending modulus: from 5.70 to 7.78 deg/bp for vertices and from 6.23 to 7.45 deg/bp for rings, depending on whether B = 10 or 100 kT (Figs. S13E and S14E). The main text fixes B = 25 kT/rad^2 from cryo-EM, but SI Section V reports that oxDNA and cryo-EM disagree on the bending modulus by roughly a factor of four. Since the toroid's largest positive angle sits at the edge of the encodable range, this uncertainty is large enough to shift the required n_delta by several base pairs. Please report confidence intervals for the fitted slopes and show how the uncertainty affects the assigned angle-domain lengths in the toroid design.","section":"SI Section VI, Figs. S13E and S14E"},{"comment":"The toroid evidence for 'precise' geometric programming is qualified by the authors' own data: only a small fraction of fully assembled toroids is observed, 21% of closed toroids have the wrong 5-fold symmetry, and the bend angle measured from opened toroids is 104.2 deg rather than the designed 90 deg (Fig. S17G). The thermodynamic model that matches the 4-fold/5-fold selectivity uses this 104 deg bend angle, not the programmed value. Thus the assembled toroid appears to close through bend-angle flexibility and tolerance rather than through faithful reproduction of the programmed local angles. Please provide a quantitative yield, compare designed versus measured local angles in the toroid, or explicitly limit the claim to robust assembly under angular tolerance rather than precise angle encoding.","section":"SI Section VII, Fig. S17"}],"minor_comments":[{"comment":"The root-mean-square deviation formula contains a malformed radical expression ('/radicaltp/radicalvertex/radicalvertex√') that should be typeset as a standard square root.","section":"SI Eq. (10)"},{"comment":"The name 'Karfuscher et al.' should be 'Karfusehr et al.' to match reference [35].","section":"Introduction, paragraph 5"},{"comment":"The text discusses a +20-delta dimer, but Section B of the main text states that angle domains are tunable from 0 to 15 base pairs; please clarify whether +20-delta is a simulation-only probe beyond the design range.","section":"SI Section II"},{"comment":"The statement that the angle-domain lengths are derived from a linear fit shown in Fig. 2C is inconsistent with the main text, where the linear fit appears in Fig. 2L; the cross-reference should be corrected.","section":"SI Section IX caption"},{"comment":"Error bars are shown on several data points but the caption does not define what they represent (standard deviation, standard error, or fit uncertainty); please define them explicitly.","section":"Fig. 2L"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope for a soft-matter or DNA-nanotechnology journal and the modular design idea is appealing, but the unexplained switch to a 7.18 deg/bp slope in SI Section IX is a load-bearing issue. I do not see a novelty-disclosure problem; the authors cite related work from the same group. If the authors supply a proper justification or sensitivity analysis for the slope and revise the toroid claim, I would support publication; otherwise the headline demonstration of precise angle programming is not supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is a real advance in modular DNA origami: a single scaffold routing for a triangular prism, >70% staple reuse across designs, and separate bond/angle domains that let them program interaction specificity and binding angle independently. The toroid assembled from 12 triangle species is the first demonstration of nonuniform Gaussian curvature in a DNA origami assembly, and it is validated by cryo-EM, TEM tomography, gel electrophoresis, and oxDNA. Second, the headline 'precise and independent programming' is softer than the abstract implies, because the toroid design silently uses a different positive-angle slope (7.18°/bp in SI IX) than the one reported in the main text (6.65°/bp), with no derivation. Without that ad hoc change, the 109.5° target would require a 16.5 bp angle domain, beyond the stated 15 bp maximum. The paper's own oxDNA gives 99.2° for +15δ, close to the 6.65 line and ~10° below target. So the toroid may close because the system tolerates angle error, not because the angles are encoded as claimed. That is the main soft spot, and it is load-bearing for the strongest claim.\n\nWhat the paper does well: the platform itself is clever and economical. The calibration from gel-based vertex/ring thermodynamics, cryo-EM dimer reconstructions, and oxDNA is internally consistent for the simple dimers and small assemblies, where the 6.65/-6.28 slopes are supported by multiple methods. The paper is unusually honest about discrepancies: the 4-fold bending modulus mismatch between oxDNA and cryo-EM, the extrusion-angle bias, the two-thymine spacer inadequacy, the 21% 5-fold toroid off-target, and the low toroid yield are all discussed in the SI. The symmetry-guided inverse design procedure is clearly explained.\n\nThe soft spots beyond the slope switch: yields are reported without error bars (gel band intensities, TEM counts); the thermodynamic model for vertex/ring angles uses fixed ΔE_B=-17 kT and bending modulus from cryo-EM, so the inferred angles are model-dependent; and the transfer of a two-species/small-cluster calibration to a 12-species toroid is a leap that the 7.18 slope is papering over. These are not fatal to the overall contribution, but they should be fixed or explicitly contextualized in revision. The central observation—that a toroid with 96 triangles assembled from a single scaffold platform—does not disappear; the interpretation of how precisely the angles were programmed does.\n\nWho this is for: DNA origami and programmable self-assembly researchers, and soft matter theorists working on design rules. It deserves a serious referee: the advance is real, the evidence is multi-method, and the calibration issue is exactly what peer review should catch. My recommendation: send it out, but flag the 7.18°/bp slope prominently and ask for either a justification, a re-fit to the +15δ oxDNA data in the design, or a tempered claim about precise angle control in the toroid.","headline":"Real modular DNA origami platform with a genuine toroid first, but the headline angle-programming claim runs through an unexplained 7.18°/bp slope that the paper's own oxDNA data undercut.","tokens_in":33079,"tokens_out":2669,"would_cite":true,"duration_ms":28128,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that a conserved triangular-prism DNA origami core, with only its overhangs varied, can encode both interactions and binding angles, enabling self-assembly of a 96-triangle toroid with locally varying curvature.","keywords":["DNA origami","modular design","programmable binding angle","self-assembly","Gaussian curvature","toroid","cryo-EM","coarse-grained simulation"],"falsifier":"Cryo-electron tomography of assembled toroids measuring the local dihedral angles at the eight junctions would settle the claim: if the measured angles deviate from the designed set (109.5 to -49.1 degrees) by more than the dimer-observed fluctuations, the calibration does not transfer, and toroid closure would be a fortuitous outcome rather than a programmed one.","tokens_in":31871,"feed_emoji":"🧬","tokens_out":10447,"duration_ms":92125,"temperature":0.7,"pith_summary":"The paper sets out to remove the main bottleneck in DNA-origami nanofabrication: the need to redesign the scaffold routing and most staples whenever a new block shape is wanted. It proposes a single modular core, a hollow right equilateral triangular prism, whose scaffold routing is completely conserved across designs and whose overhangs carry the information about which partners bind and at what angle. Because more than 70% of the staples stay identical, creating many distinct species costs only the synthesis of the changed overhang sequences. The authors validate the design rules with cryo-EM, gel electrophoresis, and coarse-grained simulations, then demonstrate self-assembly of Platonic and non-Platonic shells and a 96-triangle toroid with both positive and negative Gaussian curvature. The payoff, if correct, is that any triangulated nanoscale surface becomes accessible from a single core design at a fraction of the previous cost and effort.","feed_headline":"Modular DNA origami core builds a 96-triangle toroid","feed_subtitle":"Most staples are reused between designs, so curved nanostructures become faster and cheaper to build.","key_machinery":"The load-bearing object is the modular triangular prism itself: a hollow right equilateral triangular prism whose 148 core staples and full scaffold routing are frozen across every design, with only the 60 interface staples allowed to vary. Each interface staple carries two independently tunable parts—a five-nucleotide bond domain that encodes interaction specificity (which sides bind) and a hybridized angle domain whose length difference $n\\delta$ between two rows of overhangs sets the binding angle $\\theta$. The geometric model $\\theta = 2\\arcsin(0.34\\,n\\delta/(2d))$ with interhelical spacing $d = 2.6$ nm gives the design intuition, but the paper's actual design rule is the empirical linear calibration measured from self-closing assemblies. This calibration—roughly $6.65^\\circ$/bp for positive and $-6.28^\\circ$/bp for negative angles—is what makes it possible to prescribe the seven distinct angles needed for the toroid. The symmetry-guided inverse-design step then assigns species and interactions by the 422 symmetry group of the target toroid, reducing the 96 triangles to 12 unique species and 36 unique interactions.","core_discovery":"The paper's central claim is that interaction specificity and binding angle can be programmed independently on a single conserved DNA-origami block, and that this decoupling is enough to assemble complex, self-limiting two-dimensional manifolds. The block is a hollow right equilateral triangular prism folded from an 8064-nucleotide scaffold; 148 core staples are identical across all designs, while 60 interface staples, 20 per side, carry the program. Each interface staple ends in a five-nucleotide bond domain that sets which sides bind, and adjacent interface strands form a double-stranded angle domain whose length difference $n\\delta$ sets the binding angle $\\theta$ between neighbors. From dimer cryo-EM, coarse-grained simulations, and gel-electrophoresis measurements of self-closing vertices and rings, the paper derives a linear design rule of roughly $6.65^\\circ$ per base pair for positive angles and $-6.28^\\circ$ per base pair for negative angles. Using this rule with a symmetry-guided inverse-design algorithm, it assembles non-Platonic shells with substantially improved yields and a toroid of 96 triangles built from 12 unique species, 36 specific interactions, and 7 binding angles spanning $109.5^\\circ$ to $-49.1^\\circ$, demonstrating a DNA-origami structure whose Gaussian curvature varies globally.","pith_inferences":["The conserved-core principle should transfer to other rigid origami polyhedra, such as square- or honeycomb-lattice blocks, since the decoupling of interaction and angle programming does not rely on triangular geometry; the authors gesture at this in their conclusion but do not demonstrate it.","A systematic map of inner-versus-outer junction bond-length imbalance versus toroid bend angle would convert the paper's observation that the measured bend angle is about $104^\\circ$ rather than the designed $109.7^\\circ$ into a quantitative design rule for suppressing 5-fold toroid byproducts.","The roughly four-fold discrepancy between coarse-grained simulations and cryo-EM bending moduli implies the simulation model underestimates single-bond fluctuations; if resolved, simulations could predict angle distributions ahead of experiment rather than merely confirming them.","The sign-dependent calibration (6.65 vs -6.28 degrees per bp) and the separate 7.18 degrees per bp used for positive angles in the toroid design suggest that a single global calibration may be insufficient at extreme curvatures or helix geometries; a context-dependent calibration table would be a natural testable extension."],"forward_implications":["Any triangulated two-dimensional manifold becomes a candidate target with one fixed scaffold routing, since only the overhangs change between designs.","The cost of a new assembly drops to roughly the synthesis cost of the changed overhangs; the paper quotes about $8,000 for the toroid's 520 unique staples versus about $36,000 for 12 fully bespoke triangle designs.","A validated library of bond domains and angle domains becomes reusable across shapes, so strands ordered for one structure can be repurposed for another at no added synthesis cost.","Programming both type specificity and binding angles suppresses off-target polymorphism; yields for the triangular bipyramid, for example, rise from about 12% to 54% when both channels are programmed.","Self-limiting structures with nonuniform Gaussian curvature, such as the toroid, are assemblable in one pot, although current yields are limited by kinetic deadlock and by misassembled 5-fold toroids."],"supporting_citations":[{"why":"Supplies the square-lattice multilayer DNA-origami design on which the triangular block is based.","marker":"[16]"},{"why":"Provides the thermodynamic model of self-closing polymorph selection used to infer binding angles from gel yields.","marker":"[20]"},{"why":"Earlier triangular-prism tubule assembly with shape-complementary bonds; the approach this work replaces, and the source of the zigzag tubule geometry used for ring angles.","marker":"[22]"},{"why":"Establishes the economical overhang interaction protocol and the self-closing structure assays that this paper extends.","marker":"[23]"},{"why":"Contributes the symmetry-guided inverse-design algorithm used to assign the toroid's 12 species and 36 unique interactions.","marker":"[24]"},{"why":"Prior modular subunit design using single-stranded spacers that this approach extends to double-stranded angle domains.","marker":"[34]"},{"why":"Provides the coarse-grained DNA model used for binding-angle simulations.","marker":"[39]"},{"why":"Provides the multibody refinement method used to extract binding-angle distributions from cryo-EM data.","marker":"[41]"}],"fun_headline_variants":["DNA origami module builds curved shells and toroids","Reusable DNA blocks assemble complex curved structures","Modular DNA origami makes curved nanostructures affordable","One core, many angles: DNA origami toroid from 96 triangles","DNA origami design rule cuts cost for curved shapes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes that the angle calibration measured on dimers, vertices, and rings transfers unchanged to the 12-species toroid mixture, where closure depends on the prescribed angles summing correctly.","fun_headline_variants_meta":{"raw":{"variants":["DNA origami module builds curved shells and toroids","Reusable DNA blocks assemble complex curved structures","Modular DNA origami makes curved nanostructures affordable","One core, many angles: DNA origami toroid from 96 triangles","DNA origami design rule cuts cost for curved shapes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000438,"raw_usage":{"total_tokens":2230,"prompt_tokens":957,"completion_tokens":1273,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":573,"completion_tokens_details":{"reasoning_tokens":1194}},"tokens_in":573,"tokens_out":1273,"duration_ms":8635,"temperature":1.0,"reasoning_tokens":1194,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T19:31:59.149113+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Cryo-electron tomography of assembled toroids measuring the local dihedral angles at the eight junctions would settle the claim: if the measured angles deviate from the designed set (109.5 to -49.1 degrees) by more than the dimer-observed fluctuations, the calibration does not transfer, and toroid closure would be a fortuitous outcome rather than a programmed one.","supporting_citations":[{"cited_title":"Bednar, P","cited_arxiv_id":null,"evidence_quote":"Supplies the square-lattice multilayer DNA-origami design on which the triangular block is based."},{"cited_title":"Pazdernik and N","cited_arxiv_id":null,"evidence_quote":"Provides the thermodynamic model of self-closing polymorph selection used to infer binding angles from gel yields."},{"cited_title":"Hayakawa, T","cited_arxiv_id":null,"evidence_quote":"Earlier triangular-prism tubule assembly with shape-complementary bonds; the approach this work replaces, and the source of the zigzag tubule geometry used for ring angles."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the economical overhang interaction protocol and the self-closing structure assays that this paper extends."},{"cited_title":"Hayakawa, T","cited_arxiv_id":null,"evidence_quote":"Contributes the symmetry-guided inverse-design algorithm used to assign the toroid's 12 species and 36 unique interactions."},{"cited_title":"Nakane, D","cited_arxiv_id":null,"evidence_quote":"Provides the multibody refinement method used to extract binding-angle distributions from cryo-EM data."}],"review_version":1}