Where the author or a verifier believed the claims file misread the paper, they asked, citing the paper, and a reader who had not seen the data answered from the paper alone, amending the claims file where it was wrong. The author then ran the claim as answered. Attempts before protocol pilot-11 also asked here for method details the claims file did not give.
The OD600 sheet holds seven time points per population, day 0 (the inoculation, OD near 0.06) and days 2, 4, 6, 8, 10 and 12, while the criterion's repeated-measures ANOVA has time as a factor with df 5, so six time points. Which six enter the ANOVA? The same applies to fig1b-phage-reduces-density and fig1b-phage-effect-diminishes, and to which days 'all time points' and 'after inoculation' mean for the readings of auc-phage-reduction-by-temperature. Tested here under (A) days 2 to 12, (B) days 0 to 10, and all seven days as a variant.
Passage. Methods, Ralstonia solanacearum-phage selection experiment, "All cultures were propagated for five serial transfers (total of 12 days...). Before every serial transfer, bacterial optical densities (OD600) were measured"; Table S1 legend, "Three-way repeated measures ANOVA explaining variation in R. solanacearum densities by temperature and phage treatments for 5 transfers."
Why it matters. Days 2 to 12 reproduce every sum of squares of Table S1a and give the temperature x phage x time interaction F(5, 180) = 3.870, p = 0.0024; days 0 to 10 give F = 1.464, p = 0.204, which would make fig1b-phage-effect-diminishes partially reproduced. The temperature and phage effects hold under both. For the area under the curve, the trapezoid over days 0 to 10 gives a 43.0% reduction at 35 C, outside 37.18 +- 3.72, while the claims file's readings mapped to days 0 to 12 and 2 to 12 all match.
The reader’s answer. Days 2 to 12, reading (A). The Methods say OD600 was measured "Before every serial transfer", that is at the end of each 48 h growth period, and the last sampled time point is "48-h growth after the fifth transfer", day 12; Figure 1 legend, "Bacterial and phage densities were recorded at every transfer (2 days)". Day 0 is the inoculation, set at 10^7 CFU/ml, not a measurement before a transfer. Days 0 to 10 would leave out the day 12 sample the paper uses for the end of the experiment, so it is not a reading the paper allows. The criterion of fig1b-temperature-density now names days 2 to 12, and fig1b-phage-reduces-density and fig1b-phage-effect-diminishes use its ANOVA. For the area under the curve the paper says only "for the whole duration of the experiment"; the criterion of auc-phage-reduction-by-temperature now names its readings by day, (a) trapezoid over days 2 to 12, (b) trapezoid over days 0 to 12, (c) sum over days 2 to 12. Statements and expected values unchanged.
Under reading (a), is the 0.24% the mean of the four replicate frequencies of the 25 C +P treatment or the share of the treatment's pooled colony counts? fig1d-phage-raises-scv-frequency lists both ways of taking a treatment frequency; this claim does not.
Passage. Results, paragraph 2, "(Figure 1D; hardly visible in the mean plot due to low frequency of SCVs: 0.24% of all colonies within this treatment)".
Why it matters. The mean of replicate frequencies gives 0.244% (within 0.24 +- 0.024) and pooled counts 0.561% (outside), so the outcome is reproduced under the first and not under the second.
The reader’s answer. The paper does not say. "0.24% of all colonies within this treatment" reads as the share of the treatment's pooled colonies, while "hardly visible in the mean plot" ties the number to a plotted mean of replicate frequencies; the Figure 1D legend shows each replicate's frequencies and gives no pooling. The criterion now names both under reading (a), (a1) the mean of replicate frequencies and (a2) pooled counts, and keeps its rule that the 0.24% match under one reading, as in fig1d-phage-raises-scv-frequency. Statement and expected value unchanged.
Does the paper fix the unit of the Table S3a to S3c ANOVAs, reading (a) the mean of the clones of each colony type within each replicate population, or (b) each clone? The same question applies to fig2c-scv-lower-carrying-capacity, and to the other Figure 2A to 2C claims, where the readings agree. In the data, (a) gives 22 units, and with sequential sums of squares it reproduces every F, p and residual sum of squares of Tables S3a, S3b and S3c.
Passage. Figure 2 legend, "N = 4 for all treatments and boxplots show the interquartile range (25%-75% of the data), the median as lines and replicate means as dots (based on eight clones isolated per replicate for A-C...)"; Tables S3a to S3c, residual df 15.
Why it matters. For fig2a-temperature-resistance the temperature effect has p < 0.05 under three readings but p = 0.071 with clones and sequential sums of squares; for fig2c-scv-lower-carrying-capacity the colony type interactions are not significant with unit means but are with clones (phage x colony type p = 0.020, temperature x colony type p = 0.0027). Both claims are blocked as methods_underspecified until this is settled; under reading (a) both would be reproduced.
The reader’s answer. Reading (a), from the reported degrees of freedom. The models of Tables S3a to S3c have three main effects and three two-way interactions, seven df with the intercept, and "Residuals 15", so 22 values; 128 clones would leave 121 residual df. The Figure 2 legend plots "replicate means as dots (based on eight clones isolated per replicate for A–C", and the Methods isolate "four mucoid and four SCVs" from replicates with both types (Figure S1), so means by colony type within replicate are the paper's unit. The paper does not name the unit in words. The criterion of fig2a-phage-exposed-more-resistant now fixes (a) and shows each clone as a variant; the other Figure 2A to 2C claims, fig2c-scv-lower-carrying-capacity among them, take their unit from it. The sums of squares readings stay open. The claims file's header said 22 units fit neither reading; that was wrong and is corrected. Statements and expected values unchanged.
The claims file marks the sequencing and RNA-seq claims `data: elsewhere` (NCBI SRA). The dataset itself holds the processed results these claims rest on: the Snippy variant calls aggregated by group with gene names (snp_indel, gene_blast in uniprot), the DESeq2 table and the Figure 4B and 4C lists (transcriptome), and the KOBAS output (kegg enrichment). PRJNA1028400 holds no reads (28 BioSamples, 14 assemblies) and PRJNA1032726 holds the ten RNA-seq runs. Should these claims be `data: dataset`, tested on the deposited tables? Applies to sequencing-mutation-count, the three table1 claims and the five rnaseq claims.
Why it matters. This attempt tested them on the deposited tables. It changes how the claims are counted (with the dataset or apart from it), not their outcomes.
The reader’s answer. Yes, data dataset. The Data availability statement says "The datasets and code generated during the current study are available in the dryad digital repository" and that the "sequences of evolved R. solanacearum strains and transcriptomics raw data" went to SRA. The variant calls, the DESeq2 results and the KOBAS output are data generated in the study, and the paper's counts rest on them, so the paper places what these claims test in the Dryad dataset; SRA holds the reads they were made from. The nine claims, sequencing-mutation-count, the three table1 claims and the five rnaseq claims, are now data dataset with no where field, and the header says why. Statements, expected values and criteria unchanged.
The criterion leaves open whether the p < .05 of the KEGG enrichment is corrected, and blocks the claim when the two readings differ. Does the paper say which it is? In the dataset's KOBAS table the five named pathways are exactly the five with uncorrected p < 0.05, and no pathway has FDR < 0.05.
Passage. Results, "a total of five metabolic pathways were significantly enriched in phage compared to control treatment (p < .05, Supplementary Figure S5)"; Figure S5 legend, "The significances are shown with dot colors."
Why it matters. Under uncorrected p the claim is reproduced (the five named pathways, p 0.0024 to 0.0464); under FDR none is enriched. The verifier of round 1 read the match as pointing to uncorrected p. The claim is blocked as methods_underspecified until this is answered.
No reader has answered it yet.