Based only on the five papers you supplied (with DOIs and extracted-method/limitation notes), Junfeng Hao’s work appears to span metabolic regulation (review synthesis), neural decoding with patient ECoG, immuno-oncology via TAM polarization, pharmacovigilance signal detection, and bacterial AMR plasmid genomics. Evidence strength looks moderate-to-strong for mechanistic lab work (Sirp-α; AMR plasmid study) and higher model/analysis sophistication for the ECoG decoding paper, while the FAERS pharmacovigilance and some review-level claims are inherently limited for causality. However, the dataset you provided does not conclusively disambiguate which “Junfeng Hao” the papers map to, so confidence is capped by possible author-name mixing and by missing full-text methods/figures.
Long Explanation
Author Review: Junfeng Hao
Scope note (critical): Your input includes five DOI-linked papers plus extracted-method/limitation summaries, but it does not provide a reliable mapping that the supplied DOIs correspond to the same person “Junfeng Hao” (e.g., author-name collisions are common in scholarly databases). Therefore, the evaluation below is limited to the scientific quality signals explicitly present in the provided extracts, not a definitive track-record across the author’s entire publication history.
1) Evidence-quality snapshot (from your extracted scoring + limitations)
Higher values below reflect internal extract scores you provided (e.g., “paper_scientific_quality_score”, “paper_reproducibility_score”), plus a qualitative audit of the stated limitations (e.g., causality limits in FAERS; cross-model variability in TFE3 review; patient pathology generalizability; small in vivo n).
2) Paper-by-paper scientific-strength audit (known vs uncertain vs limited)
2.1 Emerging roles of TFE3 in metabolic regulation (review)
What’s claimed in your extract: A mechanism-oriented synthesis describing how MiT family transcription factor TFE3 coordinates insulin signaling, glucose/lipid metabolism, and links to mitochondrial/autophagy and lysosomal biogenesis across tissues and in disease contexts (including kidney cancers and models), while noting conflicting findings (e.g., adipose browning) and unresolved questions.
Strengths: Review synthesis can unify scattered mechanistic evidence and generate testable predictions; your extract explicitly flags cross-tissue/model variability and limited direct clinical data.
Key limitations (causality + model bias): The extract highlights reliance on overexpression/knockout systems that may not reflect endogenous regulation; cross-species heterogeneity; and limited clinical datasets—these factors reduce certainty about human causal effect sizes.
Epistemic confidence: Mechanistic plausibility is moderate, but causal specificity in humans remains uncertain because the work is a synthesis and because model-system artifacts are explicitly acknowledged.
Source
2.2 Dissociable frequency regimes in human temporal cortex integrate facial and acoustic cues during natural speech
What’s claimed in your extract: High-density ECoG in temporal cortex (STG and MTG) during continuous natural Mandarin audiovisual speech, using time-frequency Temporal Response Field (TRF) modeling, “unique R²” variance partitioning, and cross-modality neural speech reconstruction; your extract reports STG auditory-dominant, feature-selective encoding and MTG as a beta1-dominated multisensory hub that depends strongly on visual cues.
Strengths: The extract mentions relatively rigorous analysis components: leave-one-trial-out cross-validation, cluster-based permutation tests with FDR correction, and explicit modular feature extraction (facial action units + articulatory kinematics) tied to audiovisual segments. It also states data volume and electrode counts (1408 total electrodes across participants), which supports statistical power—while still being limited by the small participant number.
Key limitations (generalizability + design constraints): Your extract notes inability to record STS in the setup; passive viewing/listening rather than active tasks; pathology-containing participants (glioma/language disorders) which can change cortical organization; and inter-subject variability from electrode coverage. It also states data privacy restrictions (no public dataset), which can hinder independent reanalysis.
Epistemic confidence:Moderate-to-high for the specific STG/MTG findings in the studied context, but limited for generalization to healthy populations and to other languages due to patient sampling and constrained recording coverage.
Source
2.3 Sirp-α antibody inhibits renal cell carcinoma progression via Akt1/Akt2 modulation in tumor-associated macrophages
What’s claimed in your extract: A mechanistic immunology/cancer study linking Sirp-α mAb effects to PI3K/Akt signaling isoform-specific regulation in macrophages within RCC contexts, with downstream impacts on TAM polarization (shift toward M1 markers/cytokines) and reduced RCC migration/invasion and metastasis.
Strengths: Your extract describes a multi-level approach: public dataset bioinformatics (DESeq2/fgsea/clusterProfiler), human specimen and blood sample comparisons (20 tumor tissues, 20 adjacent; 40 blood samples), in vitro co-culture/conditioned-medium assays, and in vivo NSG metastasis and BALB/c subcutaneous models. It also includes pathway assays (qRT-PCR, Western blot, ELISA) and functional tests (Transwell, wound healing, macrophage killing assay).
Key limitations (causal direction + model constraints): Your extract explicitly notes relatively small in vivo group sizes (n=3/group for metastasis), reliance on THP-1–derived macrophages and cell-line models, incomplete clarity about whether Sirp-α acts directly vs macrophage-mediated, and potential off-target effects. Translation from murine/human systems to clinical practice remains to be validated.
Epistemic confidence: The mechanistic chain is moderately strong due to multi-modal evidence, but certainty about isoform-specific causal mechanisms and translational generality is limited by small in vivo cohorts and macrophage model substitution.
Source
2.4 Analysis of adverse event reporting with casimersen (FAERS pharmacovigilance)
What’s claimed in your extract: A disproportionality/signal-detection analysis of FAERS reports for casimersen, identifying PT-level safety signals across SOCs and describing time-to-onset patterns (mean/median; delayed onset) and demographic/reporting characteristics.
Strengths: The extract specifies dataset timeframe (2004 Q1–2024 Q3), deduplication, standardized coding (MedDRA v26.1), and multiple disproportionality statistics (ROR, PRR, BCPNN, EBGM) with explicit thresholds. It also includes time-to-onset analysis and age/sex subgroup comparisons.
Key limitations (spontaneous reporting epistemology): Your extract highlights classic FAERS limits: underreporting and missing data, lack of denominators (cannot estimate incidence/rates), inability to establish causality, possible confounding (comedications/comorbidities), and that signals require prospective validation. These limitations heavily constrain interpretability even when signals are statistically “significant”.
Epistemic confidence: Stronger for hypothesis generation about which adverse-event categories might deserve follow-up, weaker for causal claims about the drug’s direct effects.
Source
2.5 Co-existence of plasmids carrying tmexCD1-toprJ1 and mcr-8 in Klebsiella pneumoniae
What’s claimed in your extract: Two K. pneumoniae isolates from chicken manure carry mcr-8 on one plasmid and tmexCD1-toprJ1 on a second plasmid; genetic context suggests mobile/dispersing behavior (IS26-driven mobilization) and potential for co-transfer, emphasizing One Health relevance.
Strengths: The extract reports a full lab + genomic workflow: identification (MALDI-TOF), mcr detection (PCR + Sanger), plasmid mapping (S1-PFGE + Southern), conjugation with recipients (E. coli J53/EC600), susceptibility testing (microbroth/agar dilution with interpretive criteria), and WGS (hybrid assembly) with AMR/plasmid typing and IS element analysis. It also provides GenBank submission accession ranges.
Key limitations: Your extract emphasizes limited geographic scope and small number of mcr-8-positive isolates (2), plus functional validation gaps for actual plasmid fusion events and broader host-range/fitness effects. Inference of “co-transfer risk” is partly based on genetic contexts and in vitro transfer, not on demonstrated in vivo fusion dynamics across diverse ecological settings.
Epistemic confidence: High for the observed presence of gene/plasmid co-occurrence and conjugation evidence under tested conditions; moderate-to-low for generalizing fusion/co-transfer likelihood globally without broader sampling and functional fusion validation.
Source
3) Cross-paper patterns: what the provided evidence implies about scientific strengths
Multi-method triangulation appears in the lab/omics studies: the immunology RCC work combines human specimen context, macrophage functional polarization assays, and PI3K/Akt pathway readouts; the AMR plasmid work combines phenotypic susceptibility, conjugation experiments, and WGS plasmid/IS-element context.
Modeling/analysis sophistication is prominent in the ECoG audiovisual decoding work: explicit feature extraction pipelines and statistical safeguards (cluster permutation + FDR; cross-validated TRFs; unique R²) are described in your extract.
Clear acknowledgement of inferential limits is present where appropriate (FAERS causality limits; review cross-model caveats; patient-pathology generalizability constraints). This is scientifically positive because it reduces overclaiming risk.
4) What would most improve confidence (what could falsify/alter the assessment)
Author disambiguation: verifying ORCID/affiliations tying “Junfeng Hao” to each DOI is essential; without it, the evaluation may incorrectly pool work from different individuals.
Reproducibility evidence: public code/data access was stated for only one of the five extracts (ECoG code/supplementary available; FAERS is publicly available but still bias-laden; AMR provides GenBank; TFE3 review has no primary dataset). Independent replication attempts and availability of analysis scripts would strengthen rigor.
Causal falsification for the immunology and pharmacovigilance contexts: for Sirp-α, stronger demonstration of macrophage-specific causal necessity/sufficiency (beyond THP-1 derived systems) and larger in vivo cohorts would be decisive. For FAERS, replication in other surveillance systems (or prospective cohort/registry validation) is required before any causal interpretation.
Most relevant “BGPT next steps” you can run
Use BGPT to verify whether “Junfeng Hao” is the correct author identity across these DOIs and to pull additional mechanistic details (full-text methods, figure-level evidence, and reproducibility artifacts) for each paper.
Feedback:
Updated: April 29, 2026
BGPT Author Review
Scientific Quality
60%
Strength signals in the provided extracts include: (i) multi-modal evidence and mechanistic readouts in the immunology and AMR plasmid studies, (ii) relatively rigorous modeling/validation elements in the ECoG decoding work, and (iii) explicit limitations acknowledged where inference is inherently constrained (FAERS/KO-overexpression/review synthesis/patient cohort generalizability). Major scientific red flags are not about technique, but about epistemic linkage: the dataset does not reliably prove that all supplied DOIs correspond to the same “Junfeng Hao” (author name collision risk). Also, reproducibility is partially limited by data privacy and by small in vivo group sizes described in the extract. Overall: moderate-to-solid competence, but limited confidence in overall authorship track record and causality strength.
Communication Quality
70%
The extracted summaries suggest clear articulation of methods, limitations, and statistical approaches (e.g., TRF modeling, permutation testing, multiple disproportionality methods, and explicit caveats). However, because this evaluation is based on provided extracts rather than full manuscripts, I cannot assess clarity of writing/figures directly; the communication score is therefore moderately high but not maximal.
Author Novelty
70%
Novelty appears strongest in the ECoG audiovisual frequency-region framing (per your extracted novelty score) and in the AMR plasmid co-existence emphasis with specific gene clusters and mobilization context. The TFE3 paper is a review (lower novelty by nature), and the FAERS paper is largely methodological/surveillance signal generation (moderate novelty).
Scientific Rigor
60%
Rigor is moderate-to-strong where the extract describes concrete analytical/statistical safeguards (ECoG) and comprehensive lab+genomic pipelines (AMR plasmid study). Rigor is reduced by limitations common to the supplied contexts: FAERS design cannot support causal inference; immunology in vivo cohort size is small; and the TFE3 work is review-only. Additionally, author disambiguation uncertainty undermines assignment confidence.
I will extract each paper’s structured claims (methods, limitations, scoring) into a table, compute cross-paper averages/variance, and generate publication-level evidence-strength plots for rapid comparison.
Get emailed when your analysis is done!
We'll email you the results when your analysis is finished.
Hypothesis Graveyard
The idea that FAERS disproportionality signals reflect direct drug causation is unlikely: the extract explicitly notes spontaneous-reporting bias, missing denominators, and confounding, which typically prevents causal attribution.
The idea that TFE3 knockout/overexpression models universally reflect endogenous TFE3 regulation across tissues is weakened by the extract’s stated reliance on context-dependent and cross-model variability; effects may be model-artifact sensitive.
Science Movie
Make a narrated HD Science movie for this answer ($32 per minute)
Quick Explanation
Long Explanation
Top Data Sources
ExportMCP
Analysis Wizard
Hypothesis Graveyard
Discussion
Fuel Your Discoveries
Ask a Follow-Up
Key Insight