Test Hypothesis Evidence-qualified hypotheses with BGPT odds of true. Updated October 08, 2026.
Reduced hPepT2 uptake for FADDI-795 is directly reported with retained MICs and no observed nephrotoxicity; lipid A binding preservation is inferred only from MICs, and the small mouse nephrotoxicity groups (n=3) widen uncertainty. View evidence →
Selection against respiration-deficient high-heteroplasmy cells is well documented mechanistically and evolutionarily, but no direct quantification in human tissues exists, so the size of the temporal bias remains unmeasured. View evidence →
Genomic evidence from 84 individuals (clinal isolation-by-distance, no sympatry, ~83% morphospace overlap) supports a single species, but karyotypeβgenome integration and reproductive isolation data are missing. View evidence →
Manipulation is directly evidenced (injection success up to 0.96; residual attack surface ~12% under strong defenses), but the undetectability-via-stress-monitoring half is a plausible inference, not a tested result. View evidence →
No supplied record covers TX01 or Anomalops katoptron, so the genome claim is unverified; the estimate rests on Roseobacter trait priors and analogues like the lux-like mel2 locus used for ROS defense without light. View evidence →
Signal exclusivity is high (54.6% of signals in one database only; 0.1% in all four) and reporting infrastructure differs by region, but direct enrichment for region-specific drugs has not been tested. View evidence →
Dawn chorus mechanisms (light thresholds, melatonin rebound) are well supported in temperate and captive songbirds, but no study samples Arctic birds under continuous light, where the darkness-suppression trigger disappears. View evidence →
TOPBP1 condensates activate ATM signaling without DNA damage and depletion abrogates ATM recruitment, but AZD2858 is undocumented in the supplied evidence and no pharmacologic ATM-axis data exist. View evidence →
E505D supports the clamp (abolished light-induced HDX protection, slowed photoreduction), but E505Q and CTT-truncation tests are absent from the supplied evidence, capping confidence near a coin-flip. View evidence →
Each arm has strong support (TLR-ECSIT mtROS for immunity; reverse electron transport blockade preventing reperfusion injury), but no study demonstrates both simultaneously β no in-vivo immune-enhancing dose without damage exists. View evidence →
The ~7 ka North Asian maternal entry is directly supported by phylogenetics and ancient DNA, but the route inference and the absence-based transit argument rest on missing data β those regions were never sampled for C4a2c1. View evidence →
The mechanistic chain is strong at every intermediate link (substrate loss → SCFA-producer suppression → barrier disruption), but no study tests the chain with ultra-processed foods themselves β the UPF attribution rests on observational diet-pattern epidemiology plus borrowed mechanisms. View evidence →
Intrathecal enrichment is well documented and highly MS-specific (8.36% vs 0.64%, OR 14.1), but lytic reactivation within CNS-resident B cells driving local selection is untested and CNS EBV reservoirs remain contested. View evidence →
Cargo co-transfer is credible (Arc binds Mbnl1; mbnl1 mRNA appears in neuronal EVs non-encapsulated, conserved fly-to-mouse), but IRSp53-dependence of that loading and the AMPAR functional claim remain extrapolation. View evidence →
No evidence directly tests perturbation during emergence; indirect evidence supports thalamic-cortical connectivity being necessary for cognition but shows thalamic activity alone is insufficient for arousal. View evidence →
The PhoS→PhoP→matrix-gene arm is supported in B. velezensis FZB42 (261 DEGs, EPS changes), but the grazing pulse trigger was never tested β and PhoS is induced by phosphate scarcity, so a grazing pulse could suppress rather than induce the loop. View evidence →
The integration is methodologically coherent but currently untestable: only 11 individuals have ancestry estimates paired with GO-term data, far below power for modest effects β and ancestry is a poor proxy for individual mobility compared to isotope data. View evidence →
Elderly H3N2 vaccine effectiveness is often near zero even with matched strains, and no supplied data verify Brazilian strain composition, VE, or subtype-attributed mortality for those years β mismatch cannot be isolated from immunosenescence-limited baseline failure. View evidence →
No supplied record measures compaction kinetics or fate outcomes in this system; the estimate reflects genuine ignorance, weakly informed by evidence that p53 status and timing gate RAS-driven transformation outcomes. View evidence →
Human-mediated dispersal is confirmed (28/98 French sites share wild-cultivated clones), but the study never identifies a "lineage B" or tests trade routes against distance, and clonal sharing is directionally ambiguous. View evidence →
The actin arm and immune-signaling arm are each demonstrated separately, but the independence claim is untested and null tumorigenicity results in syngeneic models weakly oppose it. View evidence →
Tissue-specific GPX4/FSP1 redundancy is well supported (including a single F360 residue gating FSP1-inhibitor species specificity), but deferiprone is mechanistically distinct from ferroptosis-inhibitor classes and no trial outcome data link the two questions. View evidence →
Genetic architecture is strong (LOXL1 protective rare variant OR 0.04), but no longitudinal methylation-mediation analysis in glaucoma exists β and the framework hinges on blood methylation velocity being a valid surrogate for ocular epigenetic state, an assumption almost never validated. View evidence →
The monomer-shift premise has in vitro and MD support, but live-neuron smFRET has zero direct evidence and major technical barriers (labeling, dilution, oligomer confounds) β and hydrotrope "monomer shifts" and cluster-size effects may be the same phenomenon at different scales. View evidence →
No 5-HT2CR data exist in the supplied records, and the IHCH-2330 study itself shows comparable ligand poses producing opposite pharmacology via receptor-intrinsic dynamics, cutting against sequence-determinism. View evidence →
Modeling supports the physical premise (required deceleration rises from ~2.6 g at 20 m to ~4.6 g at 80 m), but pitch-up timing behavior is untested and the only field data (colugo landing forces) run counter to naive model predictions. View evidence →
Reduced MSE appeared across rest and movie conditions with no significant Group × Condition interaction (single study, n=87), which weakens a stimulus-coupling explanation; no ASD hyperscanning or stimulus-entropy data exist. View evidence →
HERV stratification of ME/CFS is moderately supported in one small cohort and complement activation is independently observed post-exercise, but no study jointly measures HERV expression and microclot/complement outcomes β the predictive claim is untested. View evidence →
Genotype-dependent synchrony changes are supported in Cntnap2 mouse models, but no evidence supports any effect of Faraday shielding on infant neural development β the environmental interaction premise is entirely unsourced. View evidence →
The in vitro mechanism is robust (5 orthogonal evidence units), but zero in vivo or behavioral data exists β and the brain's confined extracellular space (~20 nm clefts, ~20% ECS volume) is the single most important untested variable. View evidence →
Precedent exists for EM fields altering expression without cytotoxicity and for PEMF altering biofilm taxa, but no study has paired metatranscriptomics with 16S under realistic smartphone exposures β the hypothesis cannot be falsified by existing literature. View evidence →
The biomarker claim (decreased CSF sCD30 tracking progression across two cohorts) is strongly replicated, but the NF-kB attribution is inferred and untested β the decline may indicate immune exhaustion rather than simple pathway failure. View evidence →
The sighted-side mechanism is supported by one sEEG study of semantic search, but the blind mental-number-line claim has zero direct evidence, and cross-modal plasticity could push hippocampal recruitment in either direction. View evidence →
Behavioral biometrics are reliably identifying (fusion EER ~2.35%), which keeps GDPR engaged; no legal source supports a workaround, though genuine scope divergence between the regimes leaves residual uncertainty. View evidence →
FSP1 targeting in KRAS-driven lung tumors is strongly evidenced, but CPT1A co-inhibition, the cancer-stem-cell context, and TAM-derived carnitine regulation are unevidenced β and FSP1 stability is better explained by FAD metabolism and RNF8 turnover. View evidence →
FDX1-lipoylation-gated cuproptosis is well supported, but the COF-Tpy-Se-Cu material, adenovirus scheduling, and radiation-fraction dependence have zero direct evidence β plausibility is mechanistic only. View evidence →
CD24 as a macrophage checkpoint and the stage-dependent reversal are evidenced, but compensatory sialylation and the Siglec-10→STAB1 redirect have no direct supporting data in any supplied record. View evidence →
No primary data, studies, or field reports about Bayug Island's warning hierarchy or mobility-constrained households exist either way; a neutral prior is assigned with wide uncertainty, and the two mechanisms (warning reach vs evacuation capacity) predict different interventions. View evidence →
No supplied record contains data on NiV minibinders, R242/R442, or the ephrin-B2 hydrogen-bond network β and the question conflates viral with receptor-pocket numbering, an ambiguity that must be resolved before any structural argument. View evidence →
The hypothesis is speculative and untested by any supplied evidence β TERT literature focuses on cancer regulation, leaving its deep viral evolutionary origin empirically uncharted; the low value reflects absence of data, not falsification. View evidence →
Inferred from evidence: satellite identity half-life (~12 My) predicts ~89% identity after 2 My, yet G. urbanum shows complete satellite-to-Athila replacement with 57→18 Mb centromere contraction β a displacement rate ~6× passive decay, suggesting transposons act as active agents. View evidence →
ABCG2 gating is well supported for substrate drugs (mitoxantrone ~435-fold resistance), but no evidence shows CDK inhibitors are ABCG2 substrates in TNBC β and ABCG2 can mediate import (ruxolitinib) as well as efflux, so stratification direction is drug-specific. View evidence →
Know what changed, what holds up, and what remains uncertain. Every Friday. No ads.