Check Your Claims Before Reviewers Do

 Quick Explanation Copied

 Long Explanation

Who is this author? Disambiguation first

The name "Guoxin Zhang" resolves to at least ten distinct OpenAlex author profiles; the dominant profile (ORCID 0000-0002-2124-8502, h-index 49, 268 works, ~14,121 citations) is the relevant one for the reviewed papers, which share a Zhejiang/Central-China funding footprint and GitHub org (zjupgx) OpenAlex profile for Guoxin Zhang (A5100730832). Citation counts-by-year show peak impact in 2010–2011 (graphene photothermal therapy, saliency detection) and sustained recent output (29 works in 2022; 7 in 2026 to date).

Evidence quality across the reviewed corpus

The three recent life-science preprints are the scientific core. ORIGAMI (bioRxiv, June 2026) reports strong zero-shot predictive metrics — auROC 0.84, auPRC 0.82, expression r = 0.84 — but perturbation-validity correlations are moderate (CTCF-KO r = 0.45; eQTL slope r ≈ 0.56), and the authors honestly flag reliance on noisy Hi-C priors and bulk-averaged data ORIGAMI cis-regulatory network reconstruction preprint. CHORD (April 2026) integrates 305,550-cell and 160,306-cell cross-species atlases with documented limitations (quadratic scaling, unstable broad splits) and an explicit inventor patent conflict-of-interest — a transparency positive but a bias consideration CHORD cross-species scRNA-seq integration preprint. The gliomagenesis study is methodologically the strongest: mosaic MADM genetics, phosphoproteomics (PXD068601), graft and human-sample triangulation, with mTORC1 blockade as a falsifiable mechanistic node OPC cell competition drives gliomagenesis preprint.

Breadth, red flags, and blind spots

The corpus spans computational genomics, microbiome diagnostics (CRC 14-species panel, validation AUC 0.784–0.790), gastric-cancer epidemiology, materials chemistry, and dam engineering — breadth that invites over-generalization risk and makes single-author attribution noisy. Older applied papers score lower on BGPT metrics (cooling-pipe FEM: reproducibility 4/9; uranium extraction: no spectroscopic confirmation of proposed UO2A2·HA stoichiometry) Cooling pipe thin-walled concrete FEM study. All three 2026 preprints currently have zero incoming citations — too early to judge community uptake. A separate, similarly-named profile (h-index 11) authors the gastric-cancer and CRC microbiome work; identity conflation is a genuine known unknown. What would change this assessment: replication of ORIGAMI perturbation predictions prospectively, independent validation of the 14-species CRC panel, and human-cohort confirmation of OPC competition.

 BGPT Author Review

Scientific Quality

h-index 49 with 14k+ citations and first-authorship on a 2019 EES methods paper indicate genuine methodological leadership; the 2026 preprints show multi-modal rigor (MADM genetics plus phosphoproteomics plus human cohorts). Deductions: zero incoming citations on recent work, moderate perturbation-validation correlations, and the patent COI on CHORD.

Communication Quality

Papers report clear quantitative metrics, explicit limitations sections, public GitHub code and PRIDE deposits — above-average transparency. Some legacy engineering papers lack data availability and software detail, holding the score down.

Author Novelty

ORIGAMI's joint graph-reconstruction-plus-expression framework and the OPC-competition gliomagenesis mechanism are genuinely new framings; CHORD's tri-level embedding is incrementally novel within a crowded integration field.

Scientific Rigor

Cross-validation strategies, falsification criteria, and proteomics deposits are rigorous. Weak points: duplicates-only uranium extraction, single-case FEM study, and reliance on slope analysis without spectroscopic stoichiometry confirmation.