Check Claims Before You Publish

 Long Explanation Copied

Portfolio breadth and quality — with a major identity caveat

The name "Zhe Su" maps to multiple distinct researchers. OpenAlex lists at least three distinct author profiles (ORCID 0000-0001-8683-5700 with 69 works, h-index 22, ~10,653 citations; ORCID 0000-0002-0889-4198 with 96 works, h-index 34; ORCID 0000-0003-2883-6996 with 44 works, h-index 9), while the supplied citation record shows only 6 papers, h-index 4, and 116 citations — on hydrogel sensors, TADF OLEDs, shale permeability, and H2S probes. This fragmentation means any single-author assessment is inherently uncertain; name disambiguation is the dominant blindspot here.

The strongest work is the UT-AIM250 ancestry panel (quality 8, reproducibility 9, with public code and validation across 1000 Genomes AMR, TCGA-LIHC, and an HCC cohort of 26 patients), and the anti-windup wide-area damping controller paper (quality 8, damping ratio improved from 0.0172 to 0.3172, validated by Opal-RT RTDS hardware-in-the-loop) UT-AIM250 ancestry panel paper Anti-windup wide-area damping controller paper.

Methodological patterns and weaknesses

The orthostatic hypotension study (n=282, 69.86% incidence) is solidly powered but retrospective and single-center, with OH detection dependent on clinician assessment and unmeasured confounders such as hydration and analgesia Postoperative orthostatic hypotension study. The BDNF-ketamine trial is a notable negative result (BDNF changes showed no correlation with antisuicidal response, all |r| < 0.17, all p > 0.28) — publishing a null finding against prevailing biomarker enthusiasm is epistemically commendable — but it is single-arm, open-label, and uses peripheral BDNF as an imperfect proxy for central levels BDNF-ketamine antisuicidal trial. The Ma'erkang seismology paper is competent sequence analysis (h≈1.09, p≈1.02, b=0.95) but relies on evolving station coverage and small numbers of ML≥4 events Ma.

Citation metrics

The citation record shows an h-index of 4 with 116 total citations across 6 papers — a modest early-career profile. The OpenAlex ambiguity (h-indices of 22, 34, and 9 attached to different "Zhe Su" identities) means metrics cannot be reliably attributed without manual ORCID verification. The portfolio spans orthopedics, psychiatry, seismology, power engineering, and genomics — unusual breadth that suggests either a shared-name conflation or a highly mobile collaborator rather than deep specialization in one field. Counterpoint: given the disambiguation problem, any low score reflects metadata uncertainty as much as author capability.

Bottom line: Evidence quality ranges from 4 (a critique-only commentary with no new data) to 8 (two well-validated, reproducible studies). Best falsification test: confirm ORCID 0000-0001-8683-5700 to separate the true author's work and re-score.

 BGPT Author Review

Scientific Quality

Portfolio spans five distinct fields, which complicates assessment. Strongest work (UT-AIM250 genomics panel, anti-windup power control) reaches quality 8 with reproducibility 9 and public code. Weakest is a critique-only commentary (quality 4) and a single-arm ketamine trial (quality 5). Early-career metrics (h-index 4, 116 citations) are modest but confounded by name disambiguation across OpenAlex profiles with h-indices up to 34. Without confirmed identity, score carries high uncertainty.

Communication Quality

Papers are clearly structured with explicit limitations sections (a strength). The seismology and genomics papers communicate methods transparently. However, the BDNF null-result paper buries key findings in correlation tables, and the orthostatic hypotension paper reports relative risks without confidence intervals in extracted data. Multi-field writing suggests adaptability but no single distinctive voice.

Author Novelty

UT-AIM250 is genuinely novel as an exonic-only AIM panel for Hispanic admixture estimation. The anti-windup controller applies existing LMI methods to a known problem. The orthostatic hypotension and earthquake papers are incremental. The PDAC commentary is derivative critique. Overall mixed novelty with one standout contribution.

Scientific Rigor

Strengths: public code (UT-AIM250), hardware-in-the-loop validation (RTDS), explicit falsification criteria, declared conflicts of interest, and honest reporting of null results (BDNF-ketamine). Weaknesses: retrospective designs without controls (orthostatic hypotension), single-arm open-label trial (ketamine), evolving detection thresholds in seismology (Mc from 1.6 to 1.2), and no confidence intervals reported for logistic regression ORs. Rigor varies significantly across papers.