Data supporting an auditable sparse-evidence workflow for phage receptor-binding-protein prioritization and Gp17-mediated magnetic capture of Yersinia enterocolitica O:3
Description
Version 4 adds an ELISA experimental-unit correction and reanalysis record for the Journal of Biotechnology manuscript. No experiments were added. Frozen 827-record scores, weights, rankings and seeds are unchanged. V3 and earlier files remain as historical provenance. The authors confirmed that each of the three stored values per condition in the binding and specificity workbooks is the mean of three technical wells from one independent experiment. Thus n = 3 independent experiments. Individual well-level values are unavailable and were not reconstructed. V3's description of "36 technical wells" and its single-assay interpretation are superseded for the current analysis. Separate binding fits give apparent parameters of 41.21, 46.37 and 45.67 micrograms/mL: 44.41 +/- 2.80 micrograms/mL (mean +/- between-experiment sample SD). This mass-concentration parameter concerns the tagged preparation, not intrinsic molar affinity. Historical 44.38 +/- 2.98 (estimate +/- SE; R-squared 0.9904) remains an earlier analysis; its SE must not be relabelled SD. Binding uses the unrounded mean of recorded negative values. Specificity uses matched experiment rows and each group's negative-control mean, with paired two-sided tests and Holm correction. Small n limits assumptions. The corrected outputs, source workbooks and analysis code were already published in v1.1.0, https://doi.org/10.5281/zenodo.23156488, under extension_v04/04_ELISA, extension_v04/01_source_snapshot/ELISA and extension_v04/analysis/analyze_elisa.py. The added JoB_reanalysis_correction_20261006.zip contains 09_JoB_reanalysis_20261006: corrected derived tables, summary, ELISA figure, source hashes and figure crosswalk. It does not replace code. Historical software v1.0.0 remains at https://doi.org/10.5281/zenodo.22625899. Software and third-party licences remain separate. The retrospective evidence extensions distinguish 827 accession records from 744 unique sequences and examine identity, missing evidence and receptor/structure input requirements. They do not establish independent predictive accuracy, complete external prediction or cross-target generalization. The functional case concerns tagged φYeO3-12 Gp17, distinct from prior AP5-derived Gp17 magnetic-capture work. The historical 87.2% count index lacks recoverable volume/dilution normalization and is not absolute recovery. Food-extract observations are not a surface-inoculation challenge or validated detection limit. Current JoB Figure 4 (ELISA) corresponds to historical V3 Figure 3. Consult the new correction README and figure crosswalk before combining versions. Original source and image-provenance limitations remain applicable.
Files
Steps to reproduce
1. Read 09_JoB_reanalysis_20261006/README_CORRECTION.txt inside JoB_reanalysis_correction_20261006.zip first. The addition clarifies experimental units and current figure numbering; no experiments or frozen scores changed. 2. Use archive v1.1.0 at https://doi.org/10.5281/zenodo.23156488 for the corrected ELISA analysis. Workbooks: extension_v04/01_source_snapshot/ELISA. Experiment means and corrected outputs: extension_v04/04_ELISA. Analysis: extension_v04/analysis/analyze_elisa.py. Use the archived release and documented dependencies. The addition's SOURCE_SHA256.json identifies the source files and script. 3. Each stored value is the mean of three technical wells from one independent experiment (author-confirmed). Analyze the three experiment means (n = 3); do not reconstruct unavailable well-level values. Binding uses the shared unrounded negative-control mean; specificity uses matched rows and group-specific negative means. Current binding summary: 44.41 +/- 2.80 micrograms/mL (mean +/- between-experiment sample SD), distinct from the historical estimate +/- SE. 4. Retain V3 files, including 04_wet_lab_source_data/ELISA/current_full_precision, as historical analyses. The description of 36 technical observations within one assay is superseded by the author-confirmed independent-experiment interpretation. Do not overwrite historical estimates or combine their uncertainty definitions with the corrected analysis. 5. Frozen computational reproduction uses v1.0.0 (https://doi.org/10.5281/zenodo.22625899) with archived inputs/environment. Subsequent evidence extensions are separately archived in v1.1.0. This correction changes no score, rank, weight or seed. Reproduction and missing-evidence analyses do not establish independent predictive accuracy or complete external prediction. 6. Consult the addition's FIGURE_CROSSWALK.txt: current JoB Figure 4 is corrected ELISA; historical V3 Figure 3 is its earlier counterpart. Retain 06_image_manifest/FIGURE_COVERAGE.csv and original mappings for archived Figures 1-7 and S1-S2. Old manifests cover only their stated scope; the new addition has separate source hashes. 7. Preserve limitations of primary images, identities, volume/dilution normalization and food-extract observations. The 87.2% historical count index is not absolute recovery. Do not infer a validated detection limit, diagnostic performance or independent replication from photographs or historical counts.
Institutions
- Jilin UniversityJilin, Changchun
Categories
Funders
- Ministry of Science and Technology of the People's Republic of ChinaBeijing, BeijingGrant ID: 2023YFF0611502