Cryan & Dinan formalized the microbiota-gut-brain axis in a field-defining synthesis; causal claims remain largely rodent-based and correlational.

Single-cell TCR sequencing (27,524 sequences) shows EBNA1-primed CD4+ T cells cross-reacting with ANO2; raw sequencing data lack public accession numbers.

Single-cell quantification finds roughly 13.7 somatic L1 insertions per human hippocampal neuron, with corrected estimates spanning 9.9–55.8 depending on artifact handling.

Captive finch experiments support dawn song as a circadian-timed rebound from night-suppressed motivation; generalization to wild birds is proposed but untested.

Live Bacteroides gavage reduced gut LPS and plaque in ApoE−/− mice, with the strongest human evidence remaining observational and species-resolution limited.

First microscopy of Hodarchaeales cells reports large, eukaryote-like morphologies from just 40 cells at one site, without ultrastructural confirmation.

Microfluidics and genetics show flagellar wrapping causally enables confined translocation, with hook stiffness controlling the behavior across 13 species.

The caterpillar recruits Pantoea dispersa to detoxify maize benzoxazinoids via plasmid-borne genes, vertically transmitted through eggs with strain specificity unresolved.

A conceptual BJPsych Open paper proposes a vulnerability–engagement–sycophancy mechanism for chatbot-associated psychosis without any original data.

Machine-learning-designed rhodopsins absorbing near 410 nm enhanced bacterial growth under blue light, with proton pumping inferred but never directly measured.

Observational cancer-protection signals rest on AMPK data generated at metformin doses far above clinical plasma levels; randomized trials later sided against the epidemiology.

A UK cohort (N=8,913) shows a near-zero direct effect of social media use on distress, with significant indirect paths via distrust, sleep timing, and body image.

True ASD incidence is likely 30–60 per 10,000, with most of the rise attributed to ascertainment and broadened criteria; no epidemiological support for MMR or thimerosal.

Loneliness predicted cognitive decline and Alzheimer's risk over 65 months in 823 older adults, yet was unrelated to neuropathology at autopsy.

A proposed three-arm trial would test whether conscious psychedelic experience is necessary for antidepressant effects; it remains a protocol with n=9 planned and no data.

High-activity MAOA buffered court-documented maltreatment effects in white participants but not non-white participants, anticipating ancestry-dependent biomarker validity issues.

Cryptic Kota/Kurumba substructure and a rare North Asian mtDNA lineage are reported with no autosomal Northeast Asian affinity; no data or code availability statement exists.

A preprint reports SKN-1–dependent fat gain and shorter lifespan in worms, with non-monotonic doses and classic worm-assay confounds unaddressed.

A preprint proposes the histone mark H2Aub acts as a glucose-starvation glutamine reserve, supported only by pharmacology in one cell line with n=2 proteomics replicates.

Helminth therapy, maggot debridement, and leech therapy are synthesized with genuine molecular detail, but underlying evidence is almost entirely preclinical.

Oral globin digest promotes hypertrophy across zebrafish, mouse, and C2C12 models; the peptide-relay mechanism is inferential, with corporate conflicts of interest disclosed.

A CDC minireview finds H5N1 can reach mammalian gut tissue and raw-milk ingestion can establish infection in animals; human risk is judged low but theoretically possible.

Merchant's critique argues in vitro BNT162b2 reverse transcription in Huh7 cells is unlikely to translate in vivo, while offering no new experimental data itself.

Equilibrium modeling explains cytosine and thymine absence from meteorites via deamination and oxidation — a testable prediction for pristine Ryugu/Bennu samples.

Symbiont-expressed dsRNA reduces Nosema ceranae replication and improves honey bee survival in a high-novelty PNAS study with small forager sub-samples.

Influenza triggers COMB-mediated Complex III suppression in macrophages, cutting ROS and inflammation in mouse-derived cells; human macrophage confirmation is pending.

Dipeptide repeat proteins inhibit GEM body formation in ALS iPSC models, tying RAN-translation products to RNA granule pathology without in vivo validation.

Work spans DNA renaturation kinetics and a TMLE-based multicenter bladder cancer protocol; the clinical study exists only as a protocol, with no reported results.

PBD–ciprofloxacin hybrids show promising in vitro activity against M. tuberculosis, with Gyrase A inhibition inferred from modeling rather than proven biochemically.

Abrupt stiffness interfaces in GelMA hydrogels drive perpendicular fibroblast alignment and localized αSMA upregulation, but only in a single in vitro cell model.

A deep-learning framework demixes identically labeled organelles from single-channel images using nanoscale texture, reaching up to 0.96 SSIM with limits in heterogeneous environments.

High-LET iron radiation induces persistent DNA methylation changes tied to 3D genome organization; the lead author is a declared Genegoggle cofounder and shareholder.

A 2026 systematic review of 50 studies finds ~80% reporting significant links to depressive symptoms, while causality and dose-response remain uncertain due to unvalidated self-report.

Only one high-quality MMR–ASD case-control study is available in the record (OR 0.86, 95% CI 0.68–1.09); pooling beyond that would exceed the actual evidence base.

Consistent rodent wound-healing and cytoprotective effects via VEGFR2/ERK signaling coexist with fewer than 30 humans ever studied and zero randomized controlled trials.