{"id":"edd6af3a-5078-4023-bcd6-ea5f3e94ecd6","arxiv_id":"2607.13836","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The partial-structure R1 function can assemble small-molecule crystal structures from fragments, but success is limited to favorable cases and light-atom-only fragments require residual intensities after removing heavy-atom contributions.","lead":"This paper demonstrates how the partial-structure R1 function can be used to assemble small-molecule crystal structures from fragments, reporting success on three example datasets. A crystallographer might read it if they want an alternative to traditional phasing when the structure can be built from known fragments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Examples are validated only by 'no fragments fall apart' and visual inspection; no final R-values or coordinate comparisons are reported, and Example 3 involves manual retries, so the claim that pR1 assembles correct structures is not quantitatively established.","rationale":"The manuscript is best read as a demonstration of a method, not a full methods paper. The pR1 function is a known quantity from prior work, so omitting its definition is defensible; however, the demonstration's success claims rest on visual inspection. The strongest claim says 'useful tool'; for a crystallographic method, usefulness is normally established by producing a structure that agrees quantitatively with diffraction data. Since pR1 is literally an R1-based objective, the absence of any final R1 is striking. The manual retries in S5 are especially problematic: they show the orientation search is not reliably returning the correct minimum, which is exactly the kind of qualitative behavior that needs quantitative characterization. My proposed test would settle whether the accepted models are actually the best under the objective and whether the visual success criterion is sufficient. This supports the reader's conditional verdict rather than overturning it; I do not see a reason to reject the work, but the evidence base is too thin for acceptance.","tokens_in":6471,"tokens_out":4887,"duration_ms":61930,"concrete_test":"Using the public GitHub code and the JPD1252 reflection data, reproduce the assembly of Sections 3-8 exactly. After the final tweak in Section 8, compute the conventional R1 = Σ||Fo|-|Fc||/Σ|Fo| for the full model and compare it with the R1 values obtained by re-running the final placement with the next-best pR1 orientations (e.g., orientations 2/3 for SiPh2tBuMoO4 and orientations 4/5 for NnPr4) and re-tweaking. If a wrong-orientation model yields an R1 within a few percent of the claimed solution, or if the claimed solution's R1 is above ~20%, then the qualitative success criterion is insufficient; if the claimed solution is unambiguously lowest and R1 is low, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that pR1 can assemble correct crystal structures in favorable cases. What would have to be true is that the assembled models are demonstrably the correct structure and that the orientations selected as '0,1,...' are systematically meaningful. The paper's evidence falls short on both because success is asserted qualitatively. Section 6 says the partial model is correct because 'no fragments fall apart after tweak'; Section 8 repeats this as the criterion for the whole structure. No conventional R1, Rfree, coordinate r.m.s.d., or comparison to the deposited/known structure is reported for any of the three examples, even though pR1 is an R1-derived quantity. In Supporting Information S5 (JPD1249), the narrative shows repeated failed attempts before success: for OC8, orientations 0/1 fail and 2/3 succeed; for a second OC4, orientation 0 fails, orientation 3 is 'not acceptable', and only after recalculating residuals does a 'new orientation 1' work. With no quantitative score for the accepted vs rejected models, the reported successes could reflect post-hoc selection and the 'no fall apart' criterion could accept a locally stable but incorrect packing. The second issue—pR1's mathematical definition is only cited, not reproduced—is secondary given the available code, but it reinforces the need for independent reproducibility.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents a showcase of the partial-structure R1 (pR1) method for assembling small-molecule crystal structures. The method, introduced in prior work by the same author, is an approximate R1 that supposedly depends only on the orientation and location of a fragment to be added to a partial model. The paper describes one main example (JPD1252, C2/c) in which a SiPh2tBuMoO4 fragment is built in situ, four such fragments are oriented and placed, and four NnPr4 fragments are added using residual reflection intensities to overcome heavy-atom interference. Two further examples (RAP119 and JPD1249) are given in the supporting information. The central claim is that pR1 is a useful tool for solving some small-molecule crystal structures.","tokens_in":6767,"tokens_out":5685,"duration_ms":47619,"significance":"If the claim is correct, pR1 would be a phasing-free alternative for structure assembly, particularly useful for structures containing heavy atoms where light-atom fragments are otherwise overwhelmed. A positive aspect is that the Python code is available on GitHub. However, the evidence presented is largely qualitative: the paper does not report any numerical R-factors, coordinate accuracies, or comparisons with independently determined structures, and the selection of fragment orientations appears to involve post-hoc trial and error in at least one example. Thus the significance is potentially real but not yet established.","major_comments":[{"comment":"The pR1 function is not defined in this paper. Section 1 states only that pR1 is defined 'via modifying the traditional R1 by removing its dependence on the location of these other atoms' and that it 'only depends on the orientation and location of the defining fragment.' This is the mathematical foundation for the 3-d orientational search used throughout. The paper should restate the definition from Zhang & Donahue (2024) or provide a self-contained derivation of the removal step. As written, a reader cannot check whether the approximation is valid for the examples or reproduce the method without going to the cited paper.","section":"§1"},{"comment":"The success criterion is entirely qualitative: 'no fragments fall apart after tweak' (Sections 6 and 8), with final models shown visually. No final R1, aR1, Rfree, coordinate r.m.s.d., or comparison to a deposited/known structure is reported for any example, even though pR1 is an R1-derived quantity. Because tweaking can stabilize an incorrect but locally stable packing, this criterion does not establish that the assembled model is the correct crystal structure. The paper should report, at minimum, the final aR1/R1 for the accepted model and, where the structure is known, the coordinate deviation from the reference.","section":"§6, §8, SI S5"},{"comment":"The selection of orientations appears post hoc. In SI S5, OC8 orientations 0 and 1 are tried and fail; only orientations 2 and 3 'are correct'. For OC4, orientation 0 works for the first fragment, orientation 3 is 'not acceptable', and after recalculating residuals a 'new orientation 1' works. No pR1 minima or any score are given for tried and rejected orientations. Likewise §5 chooses 'orientations 0 and 1' as a 'best guess' simply because four fragments are needed. To substantiate that pR1 systematically identifies correct orientations, the paper must give the full list of candidate orientations with their pR1 values and an a priori selection rule.","section":"§5, SI S5"}],"minor_comments":[{"comment":"The synopsis says 'one concrete example' but the manuscript plus supporting information give three examples; align the summary.","section":"Synopsis"},{"comment":"Figures 1 and 2 are referenced in the text but do not appear to be embedded in the manuscript; only captions are present. Please include the actual figure panels.","section":"Figures 1 and 2"},{"comment":"'For examples' should be 'For example'.","section":"§1"},{"comment":"The abbreviation 'NnPr4' is unusual; define it as N(n-Pr)4 or tetra-n-propylammonium at first use.","section":"§2"},{"comment":"The discussion of aR1/sR1/pR1 naming is confusing: 'if the partial model contains an undetermined fragment, the aR1 is called the pR1' seems to contradict the earlier definition of pR1 as a function of a fragment to be added. Please clarify the terminology.","section":"§6"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is more of a technical note than a full research article, and the main text lacks embedded figures and quantitative validation. The author's prior work presumably defines pR1, but the showcase should be self-contained and report numerical scores for accepted and rejected candidate models. The open-source code is a strength, but the current evidence does not yet support the strength of the central claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper does what it says — it shows pR1 assembling small-molecule structures in three examples — but the evidence is mostly visual and the method's definition lives in the prior paper. I'd send it to referees, but I'd expect major revision.\n\nThe genuinely new piece is the residual-intensity workaround for light-atom-only fragments, and Example 1 (JPD1252) is a nice walk-through of building a fragment in situ. The paper is honest about what pR1 can and cannot do right now ('only some favorable structures'), and the GitHub code is a real plus: someone could rerun the calculations without reverse-engineering the method. That is more than many crystallography papers offer.\n\nThe soft spots are the ones you'd expect. The pR1 function is never defined in this paper; the reader is pointed to Zhang & Donahue (2024). For a 'showcase' paper, that's awkward but not disqualifying — the prior work is citable, and the code ties the definition down. More serious: every success is verified by 'no fragments fall apart after tweak' and visual inspection. No final R1, no Rfree, no r.m.s.d. to the known structures, even though the heavy-atom example has a known answer. Example 3 (JPD1249) is transparently a trial-and-error story — orientations 0/1 fail, then 2/3 work; later a second OC4 fails, and only after recalculating residuals does 'new orientation 1' work. With no score attached to accepted vs. rejected orientations, the reader cannot tell whether the reported successes are the result of a meaningful ranking or post-hoc selection. The 'no fall apart' criterion is also too weak: a locally stable but wrong packing could pass it.\n\nI don't think the central claim is circular. The method is not fitting the answer; the fragments are assembled using pR1 minima, and the residual-intensity trick is a sensible adaptation. But the lack of quantitative diagnostics makes the claim weaker than it needs to be.\n\nWhom is this for? Practicing small-molecule crystallographers curious about phasing alternatives. It deserves a serious referee, not a desk reject, but the authors should be pushed to define pR1 inline, report conventional R-factors for the final models, compare coordinates to deposited structures where available, and tabulate pR1 values for accepted and rejected orientations. If that revision lands, this becomes a solid, citable application note.","headline":"A genuine but mostly qualitative showcase of the pR1 assembly method; worth refereeing after the authors add a self-contained definition and quantitative validation.","tokens_in":7219,"tokens_out":1997,"would_cite":false,"duration_ms":19782,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The partial-structure R1 method can assemble small-molecule crystal structures fragment by fragment, without traditional phasing.","keywords":["partial-structure R1","pR1","crystal structure solution","fragment assembly","residual reflection intensities","phasing-free method","small-molecule crystallography"],"falsifier":"Take a crystal structure whose correct solution is already known and that contains a light-atom-only fragment plus heavy atoms. Run the pR1 calculation with the heavy-atom partial model subtracted, and check whether the correct orientation of the light-atom fragment appears among the low pR1 minima and leads to a model that does not fall apart on tweaking. If the correct orientation is absent from the ranked minima, the method's claim of usefulness in that case is falsified; more decisively, a direct computation of pR1 with and without the 'removal' of other atoms would show whether the approx","tokens_in":6348,"feed_emoji":"🧩","tokens_out":3709,"duration_ms":30007,"temperature":0.7,"pith_summary":"The paper claims that a function called the partial-structure R1 (pR1) can assemble a small-molecule crystal structure from a few idealized molecular fragments, bypassing traditional phasing. It works by first orienting and placing large fragments or fragments containing heavy atoms, then attaching smaller fragments to the growing model by optimizing only their orientation. When the remaining fragments contain only light atoms, the paper shows they cannot be found directly because heavy-atom scattering overwhelms them; instead, the heavy-atom contribution must be subtracted from the reflection data to make residual intensities. Demonstrated on one main example and two supporting ones, the claim is that in favorable situations pR1 is a useful, if not universal, route to structure solution.","feed_headline":"pR1 method assembles crystals fragment by fragment","feed_subtitle":"A single scoring function orients and places each piece; heavy atoms give way to residual intensities for light-atom fragments.","key_machinery":"The pR1 function is an approximate R1 factor defined by taking the traditional crystallographic R1 and removing its dependence on the locations of atoms not yet in the model, so that it depends only on the orientation and position of the fragment being added. In practice a full six-dimensional search is split into two three-dimensional searches: orientation first, using a 'free-standing' fragment whose position in the cell is irrelevant, then translation. Attaching a fragment to an existing partial model reduces the search to orientation alone. Residual reflection intensities, formed by subtracting the calculated contribution of the known heavy-atom partial model from observed intensities, p","core_discovery":"On the paper's own terms, the central discovery is a set of practical observations about how to use the pR1 function to solve crystal structures: start with large fragments or fragments with heavy atoms, whose correct orientations appear as low-lying local minima of the pR1 surface; attach new fragments to the known partial model so that only a three-dimensional orientation search remains; and, for light-atom-only fragments in heavy-atom structures, search residual reflection intensities after removing the heavy-atom model. The author states that these observations indicate pR1 is a useful tool for solving some small-molecule crystal structures.","pith_inferences":["The residual-intensity trick may generalize to any weak structural feature hidden behind a known strong partial model, such as locating disordered solvent or guest molecules after subtracting the ordered framework.","The ordering of pR1 minima behaves like a scoring function; one could automate fragment assembly by greedily trying top-ranked orientations and validating with a tweak step, which the paper does by hand.","A testable extension is to benchmark pR1 assembly on a library of known small-molecule structures, measuring how often the correct orientation appears within the top few pR1 minima; this would map the boundary of 'favorable situations.'","For fragments that are small and light, the paper's own example 3 shows residual intensities may still rank wrong orientations first, implying the method's reliability depends on fragment size and scattering power; a quantitative criterion for when the method is safe remains open."],"forward_implications":["Solving a small-molecule structure can proceed by assembling idealized fragments, so prior chemical knowledge of fragment geometry replaces the need for direct phasing.","pR1 orientation ranking provides a natural way to order candidate placements; low pR1 values indicate correct orientation in favorable cases.","For heavy-atom structures, a two-stage workflow—build and tweak the heavy-atom partial model, then subtract it to generate residual intensities—makes light-atom fragments findable.","The method reduces computational cost when fragments are attached: only orientation needs optimizing, a 3-dimensional search instead of a 6-dimensional one.","If a full 6-dimensional search became feasible, the author suggests pR1 could become a general tool, not just for favorable cases."],"fun_headline_variants":["pR1 orients heavy fragments to assemble crystal structures","Fragment-by-fragment crystal assembly using pR1 scoring","Heavy atoms first: pR1 trick for light-fragment orientation","pR1: orient, attach, and solve small-molecule crystal structures"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The pR1 function is assumed to genuinely depend only on the fragment's orientation and location once the other undetermined atoms are 'removed'; this paper relies on that definition from prior work and does not reproduce the proof, yet the entire search strategy stands or falls on it.","fun_headline_variants_meta":{"raw":{"variants":["pR1 orients heavy fragments to assemble crystal structures","Fragment-by-fragment crystal assembly using pR1 scoring","Heavy atoms first: pR1 trick for light-fragment orientation","pR1: orient, attach, and solve small-molecule crystal structures"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000159,"raw_usage":{"total_tokens":1011,"prompt_tokens":637,"completion_tokens":374,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":381,"completion_tokens_details":{"reasoning_tokens":300}},"tokens_in":381,"tokens_out":374,"duration_ms":11719,"temperature":1.0,"reasoning_tokens":300,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T03:32:50.690878+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a crystal structure whose correct solution is already known and that contains a light-atom-only fragment plus heavy atoms. Run the pR1 calculation with the heavy-atom partial model subtracted, and check whether the correct orientation of the light-atom fragment appears among the low pR1 minima and leads to a model that does not fall apart on tweaking. If the correct orientation is absent from the ranked minima, the method's claim of usefulness in that case is falsified; more decisively, a direct computation of pR1 with and without the 'removal' of other atoms would show whether the approx","supporting_citations":[],"review_version":1}