Cotranscriptional RNA strand exchange underlies the gene regulation mechanism in a purine-sensing transcriptional riboswitch.
The main results reproduced: recomputed values matched the published ones within tolerance.
- ✓Same input data as the authors
- ✓Reported values were directly comparable
- ✓No relevant deviation in data/preprocessing
- 🟡A deviation was attributed to the published material
- 🟡Reported values were not (fully) derivable from the shared data
- 🟡The deviation was non-trivial in magnitude
- 🟡The central claim did not (fully) hold under reproduction
- 🟡Overall, the reproduction showed a material discrepancy
A 0–100 reproducibility-quality score from the per-question grades, shown as a z-score: standard deviations above (+) or below (−) the mean of comparable assessments.
▸Reproduction agent’s raw note
REPRODUCED. Primary in-scope pipeline result (R1: Spats v2.0.5 FASTQ -> per-nt SHAPE-Seq reactivity) reproduced for all 12 deposited runs (SRA PRJNA776034) and compared to the authors' RMDB deposit (Supp Table S4, 12 rho vectors). Per-nucleotide Pearson correlation: median 0.955, range 0.698-0.997 (6 exact >=0.95, 5 within-tol >=0.85, 1 partial). Key reproduction details recovered from Supplementary Note S1 (dumbbell-based SHAPE-Seq: dumbbell=TGAACAGCGACTAGGCTCTTCA, linker=CTGACTCGGGCACCAAGGAC, count_mutations, minimum_target_match_length=26); the default Spats config maps 0 reads, and single_target_linker breaks cotrans mode -- the correct setting is cotrans=True with the default cotrans_linker and NO single_target_linker. Infrastructure note: the shared «our HPC» account is over quota on «infra» (100%) and /home (109%), so all compute ran on the /nfs/ssd2.0 SSD scratch with HOME redirected. NOT attempted: R2D2 structure-population modeling (stretch, stochastic, downstream), MD/REMD simulations and flow cytometry (out of scope: simulations/wet-lab). Grades are provisional (must_be_independently_checked); a human reviewer confirms. Note Pearson captures reactivity-profile PATTERN agreement; absolute scales differ (spats reactivity r vs RMDB rho normalization).
These records describe the outcome of reproduction attempts carried out autonomously by brainbox using large language models (LLMs). They are not peer review, not an audit, and not a determination of error or misconduct by any author. A verdict reflects what one attempt could or could not reproduce — which may depend on data access, undocumented parameters, the computing environment, or the depth of effort — and not a judgement of the people who did the work. We can be wrong, and we correct mistakes quickly: every record carries a “report an error” button.
Assessment versions
Every reproduction run is kept as an immutable version — anchored to the data as it stood, with a tamper-evident chain hash. A rerun (e.g. after an author updates a deposit) adds a new version; the previous one stays on record.
-
v1 current initial assessment Score 87assessed: 2026-06-19 ⛓ 92cae0f1067e
✎ I am an author of this paper
Updated or fixed a deposit, or is there an erratum? Ask us to re-run the metrics. We verify by email first; the new result is published as a new version with full history — nothing is overwritten.
Provenance — full disclosure
When this reproduction was carried out, which methodology version was used, and by whom — so the record can be audited and checked independently.
- Reproduced
- 2026-06-22
- Rubric version
- v1.0
- Assessed by
-
🤖 AI curator · claude (ai-curator room) · v1.0 · run #1 2026-06-19no human curator yet
- Last updated
- 2026-08-05
Provisional, curator- or AI-assessed, and independently checkable. A reproduction outcome states what one attempt could reproduce — not a judgement of the authors.
Deep full-text extraction
Model: sonnetThe yxjA purine riboswitch, whose aptamer and expression platform sequences do not fully overlap, undergoes a ligand-dependent RNA strand exchange (via a nascent central helix) that determines the riboswitch's terminated vs. anti-terminated fold and thus its 'OFF' gene regulatory outcome.
- ★ A nascent intermediate central helix forms in the yxjA riboswitch that is mutually exclusive with both the aptamer's P1 helix and the expression platform's intrinsic terminator hairpin. finding
- ★ Ligand binding stabilizes the aptamer P1 helix against strand invasion by the central helix, thereby allowing the terminator to form and shut off transcription. mechanism
- ★ Riboswitch regulatory outcome is determined by a competition between two strand exchange processes: central helix formation by unzipping the base of P1, versus terminator formation by unzipping the central helix. mechanism
- ★ Mutations disrupting central helix base pairing abolish riboswitch regulatory function, while compensatory rescue mutations restore it. finding
- ★ The yxjA riboswitch is a transcriptional 'OFF' riboswitch that downregulates expression of a nucleoside transporter (nupG) in response to excess guanine. finding
- yxjA and the adenine-responsive pbuE riboswitch share 75.4% sequence conservation and a similar predicted aptamer domain structure but differ in regulatory logic (OFF vs ON) and AD-EP sequence overlap. finding
- ★ Riboswitches lacking direct AD-EP sequence overlap may generally rely on ligand-dependent strand exchange, rather than mutually exclusive folding, to couple ligand binding to downstream expression platform folding. mechanism
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| In vivo flow cytometry fluorescence reporter assay | E. coli TG1 cells, yxjA riboswitch-SFGFP reporter plasmids (WT and mutants) | 2-aminopurine (2AP) ligand addition vs. no ligand; C79A/C79U mutations | SFGFP fluorescence (MEFL, calibrated) as proxy for gene expression | BD Accuri C6 Plus flow cytometer with CSampler Plus |
| In vivo bulk fluorescence reporter assay | E. coli, same riboswitch-SFGFP constructs | 2AP ligand vs. no ligand | Fluorescence normalized to OD600 | Biotek Synergy H1 plate reader |
| Cotranscriptional SHAPE-Seq (in vitro transcription with SHAPE chemical probing) | yxjA riboswitch RNA in stalled E. coli RNA polymerase transcription elongation complexes (two intermediate lengths and full-length constructs) | SHAPE reagent (BzCN) modification vs. DMSO control; transcription halted with EcoRI E111Q roadblock | Nucleotide-resolution SHAPE reactivity used to infer RNA structure | E. coli RNAP holoenzyme IVT system, high-throughput sequencing library prep |
| Equilibrium SHAPE-Seq structure probing | Refolded yxjA riboswitch RNA (intermediate and full-length transcripts) | SHAPE reagent (BzCN) modification under equilibrium refolded conditions | SHAPE reactivity restraints for structure modeling | — |
| Experimentally informed RNA secondary structure modeling/prediction | yxjA riboswitch sequence (computational) | none (uses SHAPE reactivity as restraints) | Predicted secondary structure models (aptamer, central helix, terminator conformations) | — |
| All-atom molecular dynamics (MD) simulation | yxjA riboswitch P1 helix/aptamer structural region (computational, restrained) | Ligand-bound vs. ligand-free state | Stability of P1 helix against strand invasion by the central helix | — |
- ▼ yxjA riboswitch dose-response assay shows fluorescence output changes with 2AP ligand concentration, consistent with OFF-switch regulation
- – SHAPE-Seq and structure modeling reveal a nascent intermediate central helix competing with P1 and the terminator
- – Mutants disrupting central helix base pairing lose riboswitch regulatory function; rescue mutants restore it
- ▲ Restrained all-atom MD simulations show ligand binding stabilizes P1 helix against strand invasion by the central helix
- other 75.4% sequence conservation (Sequence conservation between yxjA and pbuE riboswitch aptamer domains)
- count n = 9 (Dose-response fluorescence data points per condition (3 biological x 3 technical replicates))
- other 1 mM 2-aminopurine (2AP) (Ligand concentration used in 'with ligand' growth condition)
- other 40 mM BzCN final concentration (SHAPE reagent concentration used for chemical probing)
- other 5 mM MgCl2, 10 μg/ml rifampicin (Transcription initiation/roadblock conditions for cotranscriptional SHAPE-Seq)
Statistical methods review
Model: sonnetA neutral, descriptive read of the statistical approach — what was done, and (for shared learning, not as criticism) what could also have been done.
The study used in vivo flow cytometry fluorescence reporter assays (n = 9 per condition: 3 biological × 3 technical replicates) to characterize riboswitch dose-response behavior and mutant function, with results summarized as means ± SD. A linear regression (y-intercept fixed at zero) converted flow cytometer RFU to standardized MEFL units using bead calibration standards. Structural characterization relied on cotranscriptional SHAPE-Seq chemical probing (reactivities used as restraints for minimum-free-energy secondary structure prediction) and restrained all-atom MD simulations; statistical details for these modalities are not described in the provided text excerpt, which is truncated at the library preparation section.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| Linear regression with y-intercept fixed at zero (calibration curve) | Conversion of flow cytometer RFU to MEFL values using Spherotech 8-peak rainbow bead standards | 8 bead-peak data points per calibration run | not stated |
| Descriptive means with SD; no named inferential test | Flow cytometry fluorescence comparisons of WT and mutant riboswitch constructs ± 2-aminopurine (2AP) ligand (dose-response curves, Supplementary Figures S2–S4) | n = 9 (3 biological replicates × 3 technical replicates per construct per condition) | not stated |
| OD600 normalization followed by mean/SD summary; no named inferential test | Bulk plate-reader fluorescence assay (Supplementary Figures S3 and S4) | same replicate structure as flow cytometry assay; exact n not separately stated | not stated |
-
Technical replicates (n = 3) and biological replicates (n = 3) are pooled to yield n = 9 data points per condition for SD calculation and error bars↳ Could also: Technical replicates could be averaged within each biological replicate first, then inferential statistics computed across the n = 3 biological units as the true unit of analysis — Using the biological replicate as the unit of analysis avoids pseudoreplication; n = 9 implicitly treats technical measurements as independent observations, which inflates the effective sample size and the precision of the mean estimate
-
Comparisons between WT and each mutant construct are presented descriptively (mean ± SD) without a named inferential test↳ Could also: A one-way ANOVA with a post-hoc correction (e.g., Dunnett's test comparing each mutant to WT, or Tukey HSD for all pairwise comparisons) could also be applied — A formal test with a multiplicity-adjusted post-hoc procedure would quantify the probability of observed differences under the null and control the family-wise error rate across the set of mutant comparisons
-
The +ligand and −ligand conditions appear to derive from the same overnight culture split into parallel wells (a matched design), but no paired analysis is described↳ Could also: A paired t-test or Wilcoxon signed-rank test on the within-replicate ligand effect, or a mixed-effects model with replicate as a random effect, could also be used — Accounting for the paired structure of the data (same biological replicate appearing in both conditions) can increase statistical power and better reflect the actual experimental design
-
Dose-response curves are plotted as connected mean data points without fitting a parametric curve or reporting curve parameters↳ Could also: A four-parameter logistic (4PL) or Hill-equation sigmoidal model could also be fitted to the dose-response data — Fitted parameters (EC50, Hill coefficient, top/bottom plateaus) with their confidence intervals would provide quantitative, construct-comparable measures of ligand sensitivity, making it easier to assess differences across riboswitch variants
-
Error bars represent SD of n = 9 measurements↳ Could also: 95% confidence intervals (or SEM) could also be reported to convey uncertainty in the estimated group mean — SD describes the spread of individual measurements around the mean, whereas 95% CIs make the inferential precision of the mean estimate explicit; for small n the distinction is particularly pronounced and CIs are widely recommended for communicating estimation uncertainty
-
The bead calibration regression is constrained to pass through the origin (y-intercept = 0)↳ Could also: An unconstrained ordinary least-squares regression allowing a free intercept could also be fitted as a goodness-of-fit check — Fixing the intercept at zero is a defensible and common approach for bead-based calibration, but fitting an unconstrained model alongside it would verify that any detector baseline offset is negligible across the bead dynamic range
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — pmid-35348734 (R2D2 / yxjA purine riboswitch, NAR 2022 gkac102)
Paper: "Cotranscriptional RNA strand exchange underlies the gene regulation mechanism in a purine-sensing transcriptional riboswitch." Yu, Lucks et al., Nucleic Acids Research 2022, 50(21):12001. PMID 35348734 · PMCID PMC9756952 · DOI 10.1093/nar/gkac102.
Organism/system: Bacillus subtilis yxjA (nupG) purine-sensing transcriptional riboswitch. (pbuE adenine riboswitch used as a homology reference only.)
What kind of results does the paper report?
The paper mixes wet-lab assays, one sequencing-data pipeline, a structure- modeling algorithm, and large MD/REMD simulations. Classifying each by whether it is a bioinformatic-pipeline-derived result that we can re-run from the deposited data:
| # | Reported result | Pipeline / origin | In scope? |
|---|---|---|---|
| R1 | Per-nucleotide SHAPE reactivity profiles for yxjA at 135/160/220 nt, cotrans + equilibrium, 2 reps each (Supp Table S4 → RMDB) | Spats v2.0.5 on the deposited FASTQ (SRA PRJNA776034) | YES — primary target |
| R2 | Replicate consistency of reactivity patterns (Supp Figs S8/S9) | derived from R1 (Spats) | YES (derived) |
| R3 | R2D2 experimentally-informed structure populations: "100 structures most consistent with the data" per length; RNAbow plots + consensus secondary structures (Fig 3, S10) | R2D2 algorithm on R1 reactivities | STRETCH (stochastic; needs R1 first; env-heavy) |
| R4 | Aptamer homology models / X-ray template fitting | manual structural modeling | NO (manual/wet-structural) |
| R5 | MD equilibration (50 ns), 2D-REMD strand-exchange simulations (~1.6 µs total), RMSF (Table 1, Fig 6) | GROMACS MD / REMD | NO (out of scope: not a data-pipeline reproduction; ~µs of REMD is infeasible and is a simulation, not a re-analysis of deposited data) |
| R6 | Flow-cytometry gene-expression / dose curves, mutant screens (Figs 1B,2,4,5; FlowRepository FR-FCM-Z4V9…Z4VF) | wet-lab BD Accuri C6 | NO (wet-lab) |
In-scope plan
Primary (R1): Re-run Spats v2.0.5 (Python 2.7) on the 12 deposited paired-end FASTQ runs (PRJNA776034) to regenerate per-nucleotide reactivity profiles, and compare against the values the authors deposited on RMDB (Supp Table S4). Grade per-nucleotide reactivity agreement (correlation / within-tolerance), not just a single scalar.
Derived (R2): Quantify replicate-vs-replicate reactivity correlation at each length — the paper claims "general consistencies between replicates."
Stretch (R3): If R1 succeeds and the R2D2 env builds, run R2D2 on the reproduced reactivities to regenerate the 100-structure populations and compare consensus base-pairing (P1 / central helix / terminator) qualitatively. R2D2 is stochastic and selection-based, so this is a qualitative/structural comparison, not an exact-value one.
Known gaps to resolve at run time
- Spats exact options live in Supplementary Note S1 (not in the main text); the Spats cotrans-SHAPE-Seq workflow + target/adapter handling is documented at https://spats.readthedocs.io/en/master/. Must pin these before grading R1 as exact vs within-tol.
- RMDB accession IDs (the reported reactivity values) are listed in Supplementary Table S4 — needed as the comparison ground truth. Retrieve from RMDB (rmdb.stanford.edu) at compare time.
Out of scope (explicitly not attempted)
MD/REMD simulations (R5), flow cytometry (R6), manual structural modeling (R4). These are either wet-lab, manual, or multi-µs simulations — none is a re-analysis of deposited sequencing data.
Compute placement
All heavy compute (Spats, R2D2) runs on «our HPC»/SLURM reading FASTQ from «infra»; downloads on front1. «host» holds only small results. As of first pass the «our HPC» VPN tunnel is intermittently down (ssh to «host».«infra» timed out) — prep done locally; compute deferred until the central tunnel is back (per BRIEF rule 1d, the room never touches the VPN).
Assessments & scoring basis
Each contributor’s verdict, the per-question basis, and the auditable, itemised worksheet behind it.
An automated assessment. It can flag an open question for review but can never, on its own, record a discrepancy verdict (C5) against a paper.
Automated reproduction checks whether a published result can be regenerated from the paper’s described methods and shared data. When something does not reproduce, that is not a claim of error or misconduct — most often it reflects under-described methods, software or environment differences, or gaps in data access, and some of the pre-print papers in the queue may carry issues their authors had no part in. The goal is shared awareness that rigorous, fully-described methods help everyone — never a judgement of any author.
Are you an author? We would genuinely like to hear from you — to clarify the record, add data or code, re-run the pipeline after an accession update, and publish your response right next to the assessment. Everything here is open and auditable.
🚩 Report an error in this record
Spotted something wrong — a verdict you’d contest, a data or value error, or a private detail that slipped through? Tell us, with a short justification. Authors and readers are equally welcome to write in; we review every report.
Prefer email, or the form below not working? Contact us at support@doesitreproduce.com.
Reproduction footprint
claude-opus-4-8Measured resources invested to assess this paper — sanitised (machine class only, no job ids/paths). Compute = HPC accounting (SLURM); tokens = the AI agent's session.