BGPT: Author Review: Jie Guo

Fuel Your Discoveries

Quick Explanation Copied

Author Review: Jie Guo (data-limited)

Based on the papers provided in the input, the work spans immunology, protein engineering/computation, plant molecular biology, microbial ecology/proteomics, and evolutionary/phylogenetic genomics. Several papers include strong mechanistic assays and deposited raw data (e.g., GEO), but at least one correction notice suggests that data reporting errors can occur in the broader publication set. Identity ambiguity remains: “Jie Guo” is not uniquely resolvable from the provided metadata.

Long Explanation

Author Review: Jie Guo

March 22, 2026 • Evidence-grounding: only the provided raw paper metadata/excerpts and DOIs

Critical limitation (must read)

“Jie Guo” identity is ambiguous in the supplied OpenAlex-like data: the “top_author” record shown is Yanli Zhao rather than “Jie Guo”, so bibliometric inferences for “Jie Guo” cannot be trusted from that section.
Therefore, this review is restricted to evaluating scientific strength of the provided paper set (by DOI) rather than concluding anything about Jie Guo’s complete career output.

1) Visual synthesis: quality scores across the provided papers

These are the per-paper rubric scores embedded in your input (scientific quality / novelty / usefulness / reproducibility etc.). No additional scoring is imposed by BGPT.

2) What the provided papers imply about scientific capability

2.1 Mechanistic & experimental grounding (strong signal)

TCR–treg stability model in mice: The immunology paper’s excerpt describes fate-mapping Foxp3 lineage, antigen-immunization (NP-OVA/alum), altered peptide ligands with varying affinity, tetramer staining, and scRNA-seq with deposited raw data in GEO. This combination of causal perturbation (affinity/signal strength) + lineage readout + data deposition is a strong reproducibility signal. Evidence: the excerpt explicitly names methods and GEO accessions.
Evolutionary receptor–ligand validation via biophysics-like readouts: Two related CXCL17–GPR25 papers (fish/amphibian) describe engineered receptor activation assays (NanoBiT β-arrestin recruitment and ligand binding), chemotaxis, structural modeling (AlphaFold3), and sulfation dependence (TPST1/TPST2 coexpression). The excerpt also states in vitro reliance and the need for broader in vivo validation, which is appropriate epistemic humility.
Protein engineering pipeline with computational + experimental coupling: The binder-design paper’s excerpt states an integrated design pipeline, with in silico screening (AF3Score) and experimental binding measurements (BLI). While the excerpt acknowledges limited validation breadth and reliance on scoring metrics, the presence of both computation and BLI is a positive rigor indicator.
In silico proteomics classifier with multi-center validation: The tongue-coating metaproteomics paper excerpt reports a 5-center design and independent multi-center validation of a microbial-protein classifier. It also explicitly contains conflict-of-interest statements related to instrumentation support and patent co-invention, which is important for skeptical evaluation.

2.2 Reproducibility & data availability (mixed, but several strong cases)

Raw data explicitly deposited is present in the immunology case (GEO) and in the binder-design case (code/data links).
Potential reproducibility constraints appear where studies rely on specialized hardware (CIM/Ising machine framework) or where only reasonable-request data sharing is offered.

3) Skeptical red flags from the provided set

Correction indicates prior statistical/reporting inaccuracies: the provided correction DOI explicitly states inaccuracies in the original study regarding a tRF biomarker. That doesn’t invalidate the entire author record, but it is a meaningful quality/control signal.
Bias risks specific to translational/clinical models: diagnostic classifiers and biomarker pipelines are vulnerable to confounding (center, diet, sampling), feature leakage/overfitting, and reporting bias. The metaproteomics excerpt itself acknowledges limited external non-Chinese validation and potential cohort-dependent signals.
In silico dependence: several papers include strong computational steps (AlphaFold3, docking, scoring). Without additional experimental controls, predicted interfaces/poses can be fragile.

4) Focused critique: which evidence types best support (or weaken) the author’s claims?

Strongest evidence patterns observed

Lineage- and identity-based readouts for causality (immunology fate mapping + scRNA-seq) in .
Orthology + functional receptor activation with quantitative EC50/Kd/IC50 style assays in CXCL17–GPR25 studies, combined with truncated motif tests in and .

Where confidence should be capped

Clinical diagnostic claims should remain tentative until prospective, cross-population, blinded external validation is performed—consistent with the metaproteomics limitations described in .
Computational-only generalizations need cautious interpretation; even when experimental validation exists, the scope of targets/contexts matters (explicitly noted in ).

5) Bibliometric/citation-metric note (handled skeptically)

Provided OpenAlex snippet does not match the queried author “Jie Guo”. The only explicit citation metrics shown (e.g., works_count, cited_by_count, h_index) belong to an OpenAlex author record displayed as Yanli Zhao, not Jie Guo. Therefore, citation metrics cannot be used responsibly here without correct author-ID disambiguation.

6) Practical next steps (what to verify to improve evidential strength)

Resolve author identity: confirm the exact ORCID(s) or OpenAlex IDs for “Jie Guo” to avoid misattribution.
Check raw-data access: for each key claim, verify GEO/OMIX/SRA/Zenodo links and whether metadata fully supports reanalysis (batch info, processing scripts, QC metrics).
Audit model leakage: for diagnostic/prognostic signatures, verify that feature selection and thresholding are nested within cross-validation and that validation sets are truly independent and temporally separated.

Note on the requested “inline citations” policy

This response cites only the DOIs/titles explicitly supplied in your input for evidence statements. Bibliometric claims are intentionally limited due to author-ID mismatch in the provided snippet.

Feedback:

Updated: March 22, 2026