Check Claims Before You Publish

 Long Explanation Copied

Identity Attribution: A Critical First Caveat

The query targets "Yuhan Fu," but the supplied citation metrics belong to "Yu-ting Fu" (h-index 2, 6 citations, 3 papers) whose titles (salted casings, wildlife conservation, liquid desiccant dehumidifiers) do not match any of the six reviewed papers — a clear author-disambiguation conflict. OpenAlex lists at least three distinct "Yuhan Fu" profiles (h-indices 4–5, 363, 131, and 97 citations) with no ORCID linkage to the papers below, so any single-profile metric assignment is uncertain and likely conflated. This review therefore judges the scientific work on its own merit, treating author-level metrics as unreliable provenance.

Paper-by-Paper Scientific Assessment

Tuberculosis genomics (2022–2023): Two regional Chinese studies anchor the biomedical work. The larger BMC Genomics analysis (424 WGS samples, 357 RR-TB) reports adjusted ORs of 1.332 (MIRU10), 1.257 (MIRU39), and 1.400 (QUB4156), all statistically significant, with open data (SRP134826, CRA000786) supporting reproducibility — though the observational design cannot establish causality and results may not generalize beyond Chinese CDC-sourced strains BMC Genomics MIRU-rpoB association study. The smaller Anqing study (194 isolates, only 19 RR-TB) reports an implausibly wide OR of 195.2 (95% CI 46.3–823.5) for rpoB mutation predicting resistance, and contains an internal inconsistency: 19 vs 27 RR-TB strains between text and table — a red flag for reporting rigor Anqing rpoB mutation characterization.

Cancer immunology (2026, JCI): The tescalcin study is the strongest work — a mechanistic chain (H3K4me3 → TESC → ER Ca²⁺ → CALR exposure → APC phagocytosis → CD8⁺ T-cell response) supported by a prospective trial cohort (NCT03869034; AUC=0.88 for predicting ICB response), multiple mouse models including a spontaneous hepatoma, rescue with the D128A mutant, and deposited multi-omics data JCI tescalcin phagocytic checkpoint paper. Limitations: correlative human data, mouse-to-human translational gaps, and untested inhibitor safety.

Chemistry/catalysis (2018–2024): The ChemCatChem DRM paper (quality 9, novelty 9) is a well-characterized mechanistic study (100 h time-on-stream, coke 3.0% vs 4.8–16.3% for controls) but lacks public raw data ChemCatChem bimodal Ni perovskite DRM catalyst. The SAC review is comprehensive (224 refs) but risks selective coverage of positive results Electrochemical Energy Reviews SAC survey.

Veterinary virology (2026 preprint): CPV-VLP work is methodologically sound for a preprint (23.5 nm assembly, HAI ~1:29) but suffers from n=5 groups, no canine challenge data, and no endotoxin assessment CPV/CDV VLP vaccine preprint.

Overall Judgment

The corpus shows genuine methodological breadth (bioinformatics pipelines, wet-lab immunology, catalysis, VLP engineering) and increasing sophistication over time, peaking with the JCI paper's multi-modal causal design. However, the TB work shows reporting inconsistencies and implausibly wide confidence intervals typical of small, single-region samples, and identity/metrics remain unverifiable. What would change this assessment: confirmation of which OpenAlex profile belongs to the author, independent replication of the MIRU–rpoB association in non-Chinese populations, and canine-host challenge data for the VLP claims.

 BGPT Author Review

Scientific Quality

The biomedical papers show competent bioinformatics pipelines and, in the JCI tescalcin work, genuinely strong mechanistic design (mutant rescue, spontaneous tumor models, prospective trial data). However, the TB studies are single-region, observational, with internal count inconsistencies (19 vs 27 RR-TB) and an implausibly wide OR (195.2, CI 46–824), signaling limited statistical care in the smaller study. Author-level metrics cannot be confidently assigned due to identity ambiguity (multiple OpenAlex "Yuhan Fu" profiles; supplied metrics belong to a differently spelled name). Scored mid-range: solid technical execution with significant rigor gaps in the weaker papers.

Communication Quality

Papers follow standard structured reporting with clear tables of mutation frequencies and regression results, and data/code links are provided for the WGS study. Deductions: the Anqing paper contains numeric inconsistencies between text and tables, and the VLP preprint reports titers with ambiguous notation ("1:29") and incomplete safety reporting.

Author Novelty

The MIRU–rpoB association is a moderately novel hypothesis linking VNTR typing to resistance mutation propensity; two novel rpoB mutations were reported. The tescalcin phagocytic-checkpoint concept is genuinely novel in the immunotherapy field, while the catalysis reviews and perovskite DRM work refine rather than open new fields.

Scientific Rigor

Rigor is uneven. Strengths: standardized pipelines (BWA/VarScan thresholds, HGDI metrics), deposited data, mutant-rescue controls in the JCI study. Weaknesses: small RR-TB sample (n=19) producing a 195-fold OR with huge CI, text-table contradictions, no MIC data, n=5 mouse groups, reliance on in silico epitope prediction without canine validation, and no public raw data for the catalysis work. Causal language in observational genomic work is also over-reached.