Check Claims Before You Publish

 Quick Explanation Copied

 Long Explanation

Identity Disambiguation: A Critical First Step

OpenAlex returns at least six distinct researchers named 'Zihao Zhao' (h-indices 12–25, works 43–142), and the supplied corpus confirms severe homonymy: the ten papers span CRC deep learning (SCMO CRC survival model), C. elegans synapse biology (C. elegans electrical synapse elimination), marine fungal ecology, optical physics, zeolite chemistry, gene delivery, and ultrasound neuromodulation (Piezo1 mediates ultrasound neuromodulation). These cannot plausibly be one scientist. Any score below therefore applies to the papers, not a single author.

The corpus average is ~7.6, anchored by a PNAS study with rigorous conditional knockouts, pharmacology (Yoda1, ruthenium red), region-specific expression correlations, and public data — the strongest evidence unit supplied.

Red Flags and Blind Spots

• Retraction: The Oroxylin A / PKM1-HNF4α hepatoma paper in Cell Death & Disease is explicitly retracted (Retracted Oroxylin A hepatoma paper). Whether the author listed here is the same Zihao Zhao is unknown — retraction databases do not map cleanly across homonyms.
• Reproducibility variability: The Pearcey beam paper scored 4/10 for reproducibility ('Research data are not shared'); the lncRNA insulin-resistance study validated only 7 of 232 DE lncRNAs and used n=5/group.
• Small-n neuroscience: The fUS decoding study relied on 2 mice for its headline decoding result (balanced accuracy 71.9% from 229 trials), though code and Zenodo data were released.
• Positives: The strongest papers (SCMO, Piezo1) released code/data publicly, used external validation cohorts, and addressed confounds (e.g., auditory control in the Piezo1 work).

Verdict and Falsifiability

The evidence indicates competent, multimethod, often data-transparent science with occasional rigor lapses and one retraction — a profile consistent with productive mid-career researchers (h-index 12–25 for the plausible OpenAlex matches). What would change this conclusion: ORCID-resolved authorship mapping showing a single author produced all ten papers (implausible), or retraction involvement attributable to the specific author reviewed here.

 BGPT Author Review

Scientific Quality

Aggregate of ten papers scoring 5–9 on extracted quality, anchored by a strong PNAS Piezo1 knockout study and a rigorous CRC deep-learning model with public code/data. However, authorship is confounded by at least six homonymous OpenAlex authors, one paper is retracted, and reproducibility ranges from 4–8/10. Cannot credit or penalize a single individual without ORCID resolution.

Communication Quality

Papers follow standard structured reporting with quantitative provenance; the SCMO and Piezo1 papers are notably transparent (GitHub/Zenodo/Figshare). Weaknesses: the Pearcey paper shares no data, some studies omit per-condition n values, and the gastric cancer review is narrative rather than systematic.

Author Novelty

Corpus novelty scores 6–9; highlights include first quantification of fungal CAZyme contribution across global ocean depths, novel UNC-51/UNC-76 electrical-synapse elimination pathway, and tunable ring Pearcey beams. The gastric cancer review and lncRNA catalog study are more incremental.

Scientific Rigor

Strong rigor in Piezo1 (conditional KOs, pharmacological controls, auditory confound control, Holm correction in fUS work) and SCMO (five-fold CV, external cohorts). Countervailing red flags: one retraction, n=2 mice in the fUS headline decoding claim, n=5/group mouse transcriptomics with 7/232 qPCR validation, and non-shared optics data.