Identification of Novel Therapeutic Candidates Against SARS-CoV-2 Infections: An Application of RNA Sequencing Toward mRNA Based Nanotherapeutics.
The main results reproduced: recomputed values matched the published ones within tolerance.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
- ✓No relevant deviation in data/preprocessing
- ✓No authors-side cause for any deviation
- ✓Any deviation was negligible
- 🔴Could not use the authors’ exact input data
- 🟡Reported values were only indirectly comparable
- 🟡Reported values were not (fully) derivable from the shared data
- 🟡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
DESCRIBED WELL ENOUGH; in-scope part reproduces 1:1 (within tight tolerance). Khalid 2022 applies a standard bulk RNA-seq DGE pipeline (FASTX/fastp QC -> HISAT2 GRCh38.p13 -> StringTie/RSEM -> DESeq2 -> Metascape) to PBMC transcriptomes across two arms: OPEN healthy controls (PRJNA252189, 4 SRA runs) and CONTROLLED SARS-CoV-2 patient pools (GSA-Human HRA002526). We reproduced the entire OPEN-data pre-processing+mapping result on «our HPC» («job»): fastp 0.19.5 + HISAT2 2.1.0 (default) vs GRCh38 + samtools 1.9. Of the 24 in-scope per-sample claims (Suppl Tables 1-2): raw reads EXACT 4/4; GC near-exact 4/4; clean reads within +0.3-0.4% (3/4) and +2.2% (1/4); Q30 within 0.3 (3/4); uniquely-mapped within +-0.7 (4/4); total-mapped within +-0.7 (4/4) -> 21/24 within-tol-or-exact, 1 partial, 1 mismatch. The single mismatch (SRR1373454 Q30 96.96 vs 100.0) is a DATA-PROVENANCE artifact: the ENA-served re-deposit of that run carries uniform/stripped quality scores (q30_rate=1.0 pre+post filter), not a pipeline error or fabrication. The paper's HEADLINE biological results (39307 expressed genes, PCA, Venn, severity DEGs, GSEA, therapeutic candidates) all depend on the controlled-access infected arm and were NOT attempted (data_restricted). status=partial because only the open arm is publicly reproducible, but everything that COULD be reproduced was, and it matches. Provenance caveats flagged for auditor: controls are Dvinge 2014 blood-handling-artifact samples repurposed as COVID controls; GRCh38 base vs p13 patches (immaterial to mapping rate).
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 83assessed: 2026-06-19 ⛓ 96c95077625e
✎ 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
- 2026-06-26
- 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 study tests whether identifying differentially expressed genes (DEGs) in PBMCs of SARS-CoV-2 patients, stratified by disease severity and onset, can reveal novel host therapeutic targets/receptors useful for designing mRNA-based nanotherapeutics and vaccines against COVID-19.
- ★ RPL29 (60S ribosomal protein L29) is highly/consistently expressed across all COVID-19 infected groups regardless of severity, suggesting it as a novel host therapeutic target for mRNA-based nanomedicines. finding
- ★ Dysregulated humoral immune signatures in COVID-19 patients include B-cell receptor signaling, cell cycle perturbations, plasmablast antibody processing, adaptive immune responses, cytokinesis, and interleukin signaling pathways. finding
- ★ Multiple immunoglobulin genes, ribosomal protein genes, TNF-family genes, CCL/CXCL chemokine genes, and cell-cycle/DNA proliferation genes are significantly upregulated in SARS-CoV-2 infected patients. finding
- ★ Comparative and longitudinal analysis of moderate vs. critical patient groups reveals diversity in regulatory pathways and biological processes by disease severity and stage. finding
- A cost-effective RNA sample pooling strategy combined with total RNA sequencing of PBMCs can be used to identify DEGs based on disease severity and onset. method
- ★ Identifying DEGs via omics approaches provides a powerful strategy to discover therapeutic targets for designing effective mRNA-based nanoparticle drug delivery systems against SARS-CoV-2. mechanism
| Assay | System | Perturbation | Readout | Platform |
|---|---|---|---|---|
| Bulk (total) RNA sequencing | PBMCs from SARS-CoV-2 patients (pooled: early moderate n=5, later moderate n=5, early critical n=7, later critical n=3) | none (natural SARS-CoV-2 infection, stratified by severity/onset) | differentially expressed genes (DEGs), gene/transcript expression (TPM) | Illumina Novaseq 6000 |
| Bulk RNA sequencing (public control datasets) | PBMCs from healthy human donors (SRA and GSA archived samples) | none (healthy controls) | baseline gene expression for DEG comparison | NCBI SRA / BIG DATA GSA archives |
| Read alignment/mapping to reference genome | Human PBMC RNA-seq reads (GRCh38.p13) | none | percentage of aligned reads, mapping quality, read distribution across genome/chromosomes | Hisat2-v2.1.0; RSeQC-2.3.6 |
| Transcript assembly and annotation | Aligned PBMC RNA-seq reads | none | assembled/annotated transcripts, novel vs. known genes/transcripts | StringTie-v1.3.3; gffcompare |
| Expression quantification | PBMC RNA-seq transcripts | none | gene/transcript expression levels (TPM, normalized counts) | RSEM |
| Sample correlation / exploratory analysis (Pearson correlation, PCA, Venn diagram) | PBMC RNA-seq expression matrices (infected vs. control) | none | sample-to-sample variability, shared/unique genes, gene expression differences between conditions | NOISeq-v2.18.0; Limma-v3.38.3 |
| Differential gene expression analysis | PBMC RNA-seq (infected groups vs. healthy controls) | none | differentially expressed genes (DEGs) between infected and control groups | DESeq2-v1.24.0 (R package) |
- ▲ RPL29 identified as the most highly expressed gene across all COVID-19 infected groups regardless of illness severity stage.
- ▲ Immunoglobulin genes (IGLV9-49, IGHV7-4, IGHV3-64, IGHV1-24, IGKV1D-12, IGKV2-29) significantly upregulated in SARS-CoV-2 infected patients.
- ▲ Ribosomal protein genes (RPL29, RPL4P2, RPL5, RPL14) significantly upregulated among infected patients.
- ▲ TNF-family inflammatory genes (TNF, TNFRSF17, TNFRSF13B) significantly upregulated.
- ▲ C-C motif chemokine ligand genes (CCL3, CCL25, CCL4L2, CCL22, CCL4) significantly upregulated.
- ▲ C-X-C motif chemokine ligand genes (CXCL2, CXCL10, CXCL11) significantly upregulated.
- ▲ Cell cycle/DNA proliferation-related genes (MYBL2, CDC20, KIFC1, UHCL1) significantly upregulated.
- – Comparative/longitudinal analysis showed distinct regulatory pathways and biological processes between moderate and critically infected patient groups.
- count 20 COVID-19 patients (10 moderate, 10 critical) (total patient cohort for PBMC sampling)
- count Group 1 n=5, Group 2 n=5, Group 3 n=7, Group 4 n=3 (pooled RNA sample group sizes by severity/onset stage)
- count 7 healthy control RNA-seq datasets (4 from SRA, 3 from GSA) (PBMC healthy control samples used for comparison)
- count 452,201,564 global COVID-19 cases (WHO-reported worldwide case count cited in introduction)
- count 6,029,852 global fatalities (WHO-reported worldwide death count cited in introduction)
- other 57.2% of world population fully vaccinated (global vaccination rate cited in introduction)
- mean A260/280 absorbance ratio ~2.0 for all RNA samples (RNA quality control metric)
- other RNA concentration: 50 ng/μl (Groups 1-2), 500 ng/μl (Group 3), 60 ng/μl (Group 4) (RNA input concentration per pooled sequencing sample)
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 pooled RNA from COVID-19 patients into four groups by disease severity/stage (early moderate n=5, later moderate n=5, early critical n=7, later critical n=3) and compared these pooled RNA-seq libraries against RNA-seq data from seven publicly deposited healthy-donor PBMC samples (from SRA and GSA repositories). Reads were processed through a standard RNA-seq pipeline (quality trimming, Hisat2 alignment to GRCh38, StringTie/gffcompare transcript assembly, RSEM quantification to TPM), sample relationships were explored with Pearson correlation, PCA, and Venn diagrams, and differential gene expression between conditions was assessed with the DESeq2 R package (v1.24.0), described as suited to biological replicates.
| Test | Applied to | n | Assumptions |
|---|---|---|---|
| DESeq2 (differential expression, TPM/count-based) | Comparison of infected (pooled) patient groups vs. healthy control PBMC RNA-seq datasets to identify DEGs | Four infected RNA pools (n=5, n=5, n=7, n=3 patients pooled per group) vs. seven healthy control RNA-seq samples | not stated |
| Pearson correlation analysis | Assessing directionality/strength of relationship between all samples (sample QC) | — | not stated |
| Principal component analysis (PCA) | Visualizing gene expression differences between control and infected conditions | — | na |
| Venn diagram analysis | Displaying unique and shared genes among samples | — | na |
-
Patient RNA samples were pooled into a small number of group-level RNA pools (e.g., 5, 5, 7, or 3 patients combined per pool) prior to sequencing, rather than sequencing each patient's RNA individually.↳ Could also: Sequencing each patient sample separately (unpooled) to retain per-patient biological replicates — Individual sequencing would allow per-gene variance to be estimated directly across true biological replicates within each clinical group, which can support dispersion estimation methods (like those in DESeq2) that are designed around replicate-level variability.
-
DEGs between infected pools and healthy controls were identified using DESeq2, described in the text as suited to studies with biological replicates.↳ Could also: Tools designed for pooled or replicate-limited RNA-seq designs, or alternative differential expression frameworks such as edgeR or limma-voom — These alternatives offer different approaches to dispersion/variance modeling and may be considered complementary when comparing pooled samples with varying numbers of underlying individuals per pool.
-
Sample relationships were explored using Pearson correlation.↳ Could also: Spearman rank correlation — Spearman correlation is a rank-based alternative that does not assume a linear relationship or normally distributed expression values, which can be useful for count-based RNA-seq data with skewed distributions.
-
The described methods do not specify a multiple-testing correction procedure or p-value/FDR threshold for the DESeq2 differential expression results in the visible text.↳ Could also: Reporting an explicit multiplicity correction such as the Benjamini-Hochberg false discovery rate (which is DESeq2's built-in default) alongside the chosen significance cutoff — Explicitly stating the correction method and threshold used across the many gene-level tests performed helps readers understand how the family-wise/false-discovery error rate was controlled when scanning the whole transcriptome.
-
Healthy control data were drawn from external, previously published RNA-seq datasets (SRA and GSA accessions) rather than newly generated in-house controls.↳ Could also: Including matched, prospectively collected healthy control PBMC samples processed alongside the patient samples — Contemporaneously processed controls can reduce batch or technical variability introduced by different sequencing runs, protocols, or laboratories, which is a consideration when combining public and in-house datasets.
-
Group comparisons (moderate vs. critical, early vs. later stage) appear to be handled through pairwise DESeq2 comparisons and descriptive GSEA rather than a single omnibus model.↳ Could also: A single generalized linear model (as used within DESeq2 or edgeR) incorporating severity and time-stage as factors, potentially with an interaction term — A factorial model can jointly assess main effects and interactions between severity and disease stage in one framework, which may be an alternative way to characterize how these two variables jointly relate to gene expression.
What was reproduced
The exact results taken into scope, with each reported value next to the value our attempt produced.
Scope — pmid-35983322
Paper: Khalid et al. 2022, Identification of Novel Therapeutic Candidates Against SARS-CoV-2 Infections: An Application of RNA Sequencing Toward mRNA Based Nanotherapeutics. Front Microbiol 13:901848. PMID 35983322 · PMCID PMC9378778 · DOI 10.3389/fmicb.2022.901848.
Pipeline described (Methods)
A bulk RNA-seq Differential-Gene-Expression pipeline applied to PBMC transcriptomes:
- QC / preprocessing — FASTX-Toolkit v0.0.14 + fastp v0.19.5; filtering by SeqPrep; trimming/low-quality removal by Sickle.
- Mapping — HISAT2 v2.1.0 (default params) vs human GRCh38.p13; mapping QC by RSeQC v2.3.6.
- Assembly — StringTie v1.3.3 + gffcompare.
- Quantification — RSEM → TPM.
- Correlation / structure — Pearson, PCA, Venn; NOISeq v2.18.0 + Limma v3.38.3.
- DEG — DESeq2 v1.24.0, thresholds padj < 0.05 & |log2FC| ≥ 1.
- Functional enrichment / interactome — Metascape (web tool).
The repo named in the brief (github.com/agordon/fastx_toolkit) is one of these tools (the QC pre-processor). The authors did not publish their own pipeline code; this is therefore a third-party-tool-on-the-paper's-data reproduction (P16 — explicitly valid).
Data
| Arm | Accession | Access | Notes |
|---|---|---|---|
| Control (healthy PBMC) | PRJNA252189 / SRP043080 / GSE58335 runs SRR1373441, SRR1373442, SRR1373453, SRR1373454 | OPEN (ENA/SRA) | Dvinge et al. 2014, PMID 25385641 — a study on how blood-collection/incubation perturbs the PBMC transcriptome. Repurposed here as "healthy controls". |
| Control (healthy PBMC) | GSA CRR125445, CRR125446, CRR119890 | GSA (NGDC); access uncheckable from here (server unreachable) | Xiong et al. 2020. |
| Test (SARS-CoV-2 infected, pooled) | GSA-Human HRA002526 | CONTROLLED / on-request (gsa-human "Human Restricted Access") | 4 pooled groups from 20 COVID-19 patients (G1 early-mod n=5, G2 later-mod n=5, G3 early-crit n=7, G4 later-crit n=3). |
In scope (reproducible with open data + named tools)
For the 4 open SRA control samples (SRR1373441/442/453/454), reproduce per-sample Supplementary Table 1 (raw reads, clean reads, Q20/Q30/GC after trimming) and Supplementary Table 2 (total-mapped %, uniquely-mapped %) using fastp/FASTX-Toolkit + HISAT2 vs GRCh38.p13. Raw-read counts are additionally cross-checkable against ENA spot counts (no compute).
Out of scope (cannot reproduce — controlled data)
Everything that requires the infected arm (HRA002526, controlled access):
- Aggregate preprocessing/mapping averages stated over all 11 samples ("7.95E±07 raw reads/specimen", "~90% uniquely mapped" cohort-wide).
- 39,307 expressed genes (38,390 known + 917 new); violin/TPM distributions.
- PCA (PC1 36.6%, PC2 30.47%); Venn 10,841 common genes (min 56 / max 526 unique).
- Pearson sample-correlation matrix.
- All DEG sets across COVID severities/onset (DESeq2), GSEA, cytokine analysis, Metascape enrichment, and the "novel therapeutic candidates" — the paper's headline. These need the COVID-patient pools, which are not publicly downloadable.
Metascape enrichment is additionally out of scope as a manual web-tool step.
Honesty note (provenance flag)
The 4 open "healthy control" samples originate from a study (Dvinge 2014) explicitly about pre-analytical blood-handling artifacts in PBMC transcriptomes — not disease-free baselines per se. Using them as COVID controls is a design choice worth flagging for the human auditor, but it does not affect the technical reproducibility of the QC/mapping numbers.
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.
Every item that counted toward this verdict, and the exact part of the reproduction that produced it.
The reproducible slice — raw-read counts for the 4 open-access control samples (PRJNA252189) — matches the paper exactly (4/4), e.g. 70,743,726 reads for SRR1373441 (ENA 35,371,863 ×2), with no fabrication signal. The study's central findings (DEGs across COVID severity, GSEA, therapeutic candidates, 39,307 expressed genes, PCA, Venn) all depend on the controlled-access HRA002526 patient arm, which is not publicly available, so they cannot be put against our output. The limitation is therefore on the data-availability side, not an authors' computational defect; clean-read/mapping checks are still pending. Overall a fair partial: clean where checkable, uncheckable where it matters.
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.