Check Claims Before You Publish

Structural Biology & Biophysics

Tom Blundell — MODELLER and pkCSM creator

OpenAlex h-index 114 (~63k citations) reflects field-defining computational tools, though a conflicting low-metric record is almost certainly a disambiguation error.

Achilleas S. Frangakis — cellular cryo-electron tomography

H-index 48 with a record of method-building and falsifying textbook models, but the flagship bacterial DNA-conduit claims rest on correlative cryo-ET evidence without a genetic determinant.

Matthias Geyer — GSDMD nanobody inhibitors

H-index 73 with rigorous multi-modal validation of nanobody pore-blocking, though efficacy data are in vitro-only and a declared commercial conflict adds therapeutic-claim uncertainty.

Francesco Saccon — peptide ML benchmarking

A 45-descriptor XGBoost baseline matched billion-parameter protein language models on most peptide tasks, with the author's own model's molecular-weight shortcut (ω=0.72) honestly quantified; impact rests mostly on collaborative work.

Yuki Kagaya — AI cryo-EM structural modeling

H-index 13 with CryoZeta showing strong gains on 200 cryo-EM maps, but PathFold validation is simulation-only and the queried Semantic Scholar profile reflects mispairing, not limited science.

Neuroscience

Gina R. Poe — REM sleep and memory

H-index 27 with solid empirical work, though citation impact is review-driven and key correlational claims rest on n=8 subsamples with a non-REM-specific drug manipulation.

Ichiro Kanazawa — neurogenetics and neuroimmunology

H-index 82–90 with foundational discoveries (DRPLA, anti-GQ1b) and rare published null trials, tempered by heavy reliance on large-collaboration authorship and unexplained findings in the rTMS trial.

Bibiana Bielekova — multiple sclerosis immunology

H-index 47 with a landmark CSF proteomics study and open code/data, but recent proteomic associations are correlational and industry CRADA entanglement raises sponsor-bias risk.

Javier Rueda-Carrasco — microglia and Alzheimer's

H-index 8 with strong first-author work reframing synapse engulfment as potentially adaptive, though a 2026 token-funded longevity screen shows small samples and sponsor conflicts.

Farr Niere — DJ-1 and synaptic signaling

A classic one-strong-first-author-paper citation pattern, with the most novel DJ-1/Alzheimer's hypothesis still abstract-only and unreplicated.

Moustafa Algamal — anti-Aβ immunotherapy neuroimmunology

Rare full-stack methodological breadth (NMR to intravital imaging) in a rigorous 2026 ARIA-mechanism study, capped by small group sizes (n=3–6) and a single confounded human case.

Immunology & Glycobiology

Yvette van Kooyk — DC-SIGN and glyco-immunology

A single receptor discovery seeded an entire field (h-index 101), but the recurring blindspot is extrapolating in vitro human dendritic-cell biology to in vivo pathogenesis.

Wilson W. Wong — synthetic biology circuit design

Repeated firsts in CAR logic control and saRNA therapeutics with strong preclinical controls, limited by small NSG cohorts, an unequal CAR-dose comparison, and equity conflicts without independent replication.

Microbiology & Genomics

Klaas M. Pos — AcrB efflux and DNA conduits

H-index 43 with paradigm-setting 2006 AcrB work and continued late-career novelty, though conduit claims rest on a single strain pair with no biogenesis gene identified.

Inga Hänelt — membrane transport and bacterial conduits

H-index 23 with multi-technique triangulation and honest confounder reporting; the conduit-discovery claim is mechanistically incomplete pending genetic tests of conduit disruption.

Kerwyn Casey Huang — quantitative microbiology

H-index 73 with large-scale, open-data studies across scales (2,876 fecal samples; >1M mutants), demonstrating a transferable falsifiable framework rather than a single discovery.

Brett J. Baker — archaeal genomics and tree of life

H-index 57 with co-definition of CPR/DPANN and Asgard archaea, though several flagship works are consortium papers where his specific conceptual contribution is hard to isolate.

Shanfeng Zhu — metagenomic binning and alignment tools

H-index 40 with community-benchmark authorship (CAFA, CAMI) and well-benchmarked 2026 tools, with the caveat that SCGBinner's flagship claim rests on a single soil sample.

Sonny T. M. Lee — microbiome and tumor immunity

Citation profiles diverge >30-fold between databases for the same ORCID person, showing that database provenance is itself a scientific variable when judging an author.

Ecology & Evolution

J. G. M. Thewissen — whale origins paleontology

H-index 46 anchored by ~5 decisive transitional-fossil papers, illustrating how paleontology rewards rare decisive specimens over publication volume.

Peter Bridgewater — novel ecosystems and conservation governance

H-index 30 with field-defining framing papers, but conceptual governance science is difficult to falsify and the novel-ecosystems debate remains philosophically rather than empirically resolved.

Alexandros Bousios — transposable elements and centromeres

H-index 15 with explicit statistical testing and open code, though flagship 2026 outputs are computational without direct functional validation of proposed mechanisms.

Mansi Srivastava — acoel regeneration and the tiled brain

H-index 20 with a bespoke model organism (Hofstenia), open data, and strong statistics; the tiled-brain hypothesis is a falsifiable alternative to nerve-net narratives but remains partially unproven.

Virginia M. Weis — coral symbiosis and bleaching

H-index 57 with field-defining mechanistic work and a rare self-critique quantifying how culture-media nutrient excess confounds symbiosis comparisons.

Metabolism & Molecular Biology

Bruce M. Spiegelman — metabolic master regulators

H-index 186 with repeated field-opening discoveries (PPARγ, PGC-1α, irisin), tempered by translational gaps: TNF-α in obesity and irisin in humans have not translated clinically.

Susanne M. Clee — ABCA1 and diabetes genetics

Roughly 40% of lifetime citations trace to a single 1999 gene-discovery paper, showing how raw citation counts can obscure where authorship leadership actually lies.

Rafael Loch Batista — DSD endocrine genetics

Genuine clinical expertise in rare-disease cohorts, though most outputs are narrative reviews without evidence grading and rare-disease designs carry inherent ascertainment bias.

Lance A. Johnson — APOE glial immunometabolism

H-index 33 with a coherent lipid-droplet/inflammation program for APOE4 risk, but mechanistic claims rest mostly on immortalized cell lines and mouse models with limited human validation.

Juan Manuel González-Rosa — zebrafish cardiac regeneration

H-index 21 with a field-standard cryoinjury model and Tnni3k work that empirically falsifies the lab's own earlier polyploidization narrative for that locus.

Florian Wilfling — in situ autophagy structure

A method-building PI converting cell-biology observations into quantitative physical mechanisms, with the caveat that the 2026 flagship is a preprint with small tomogram counts in yeast only.

Methods, Computation & AI

Ralf Jungmann — DNA-PAINT super-resolution

H-index 58 with sustained invention of multiplexing paradigms (DNA-PAINT through Combi-PAINT), though some claims such as >1300-channel potential are projections not yet empirically demonstrated.

Nina Miolane — geometric statistics and organoid topology

A rising, technically excellent mid-career scientist whose citation lag (2024–2025 works uncited) is typical of infrastructure scientists whose impact compounds later.

Grace Shin — FALCON lipid nanoparticle optimization

H-index 9 with an innovative closed-loop ML approach showing 84-fold cell selectivity gains, but only mouse-model validation and OpenAlex entity ambiguity moderate confidence.

Yao-Qing Chen — AMPBAN antimicrobial peptide prediction

Solid data curation and leakage-controlled benchmarking (MCC 0.86 benchmark vs 0.69 independent), reflecting a field-wide database-bias problem rather than an author-specific failure.

Clinical & Translational Science

Blood-brain barrier — see William A. Banks

H-index ~126 with a program that inverted the "BBB as wall" paradigm into "BBB as controlled interface," though some HPLC timepoints used 1–3 animals and data-sharing statements are absent.

APOE4 and anti-amyloid risk — see Lance A. Johnson

The glial lipid-handling hypothesis predicts which patients will suffer ARIA from anti-amyloid immunotherapy, but human in vivo validation remains limited.

Merle-Marie Pittelkow — Bayesian evidence assessment

The BF_BMA framework exposed a ~40-fold spread in evidential strength across FDA-approved psychotropic classes, with h-index 9 indicating an emerging rather than transformative record.

Giuseppe Riva — VR for mental health

H-index 103 as the field's most prolific author, but the clinical evidence base includes n=1 designs without controls and prolific output with occasional thin empirical grounding.

Physics, Education & Metascience

Supriya Pan — interacting dark energy cosmology

H-index 51 with honest self-reporting that even successful accelerating attractors fail to solve the cosmic coincidence problem — a model of rigorous model-building over narrative-driven cosmology.

Xinjie Xie — LLM humor in AI tutoring

A well-powered 2×2 experiment showing large affective effects (η²≈0.14–0.16) but small retention effects (η²≈0.05), suggesting humor acts primarily as an engagement lever with cognitive benefits contingent.

Rita Sleimen-Malkoun — multiscale entropy and aging

~13% of lifetime citations flow from one methods paper, so the real asset is a measurement toolkit rather than a single empirical discovery; her 2026 gerophysics pivot remains mechanism-untested by her own account.

Shuang Li — materials and early-career breadth

H-index 3 across six papers spanning four unrelated topics, with an introspective finding that rating systems rewarding extractable numbers can conflate rigor with data density.

Cross-Cutting Limitations

Author identity in bibliometric databases is unreliable: several reviews above (Blundell, Kagaya, Sonny T. M. Lee, Thewissen, Batista, Bridgewater) flag disambiguation failure that can misjudge an author by an entire career stage.

Scores are grounded in the extracted record only; claims dependent on preprints, surrogate assumptions, single-strain scope, or missing full-text methods are stated as such and should not be treated as independently replicated.