Archives
How COMET Learns Lipid Nanoparticle Design
2026-11-14
A literature-focused analysis of COMET, a transformer-based model that integrates lipid identities, composition and formulation-wide features to predict lipid nanoparticle performance. The study shows strong ranking accuracy within its dataset, useful but uneven chemical and cell-line generalization, and important limits imposed by downstream reporter readouts and weak in vitro–in vivo correspondence.
What LNP Models Can—and Cannot—Show
2026-11-13
A transformer-based study shows that lipid nanoparticle performance depends on interacting components, composition and formulation context. Its findings clarify how to interpret luciferase-based delivery measurements, but they do not establish the activity or biological effect of 5-Methoxy-UTP, which was not reported as a tested nucleotide variable.
ACE2, Mitochondria and Septic Cardiomyopathy
2026-11-13
A mouse study examined whether pharmacological ACE2 modulation changes sepsis-associated cardiac injury and whether the MasR-Sirt1 pathway links ACE2 activity to mitochondrial biogenesis.
Hyperspectral Mapping of Infarcted Heart Structure
2026-11-12
A proof-of-concept study combined panoramic hyperspectral imaging with optical mapping to relate epicardial tissue composition to cardiac electrophysiology in infarcted rat hearts. The approach identified infarct and border-zone regions, spatially paired them with action-potential measurements, and revealed preliminary tissue-specific differences in action-potential duration.
What COMET Shows About LNP Design
2026-11-12
A critical interpretation of COMET’s reported performance, the LANCE dataset, formulation comparisons, and the limits of translating cell-based ranking into broader claims about lipid nanoparticle efficacy.
SCP4, H3T3 Dephosphorylation, and Chromosome Stability
2026-11-11
A paper-centred interpretation of evidence that the nuclear phosphatase SCP4 directly removes the H3T3 phosphate during early mitosis. The study connects this activity with chromosomal passenger complex recruitment, chromosome segregation, aneuploidy, and early embryonic cleavage, while retaining important limits around replication, model scope, and embryo-level attribution.
What COMET Shows About LNP Design
2026-11-11
A claim-focused synthesis of how COMET modeled lipid nanoparticle performance, what its reported correlations establish, and why formulation prediction should not be confused with causal or therapeutic validation.
How COMET Learns the Logic of LNP Formulations
2026-11-10
A literature-focused analysis of COMET, a transformer-based model that predicts lipid nanoparticle performance from molecular identity, composition and formulation-wide features, with attention to validation design, reported effect sizes and limits on transferability.
5-Methoxy-UTP and the Limits of LNP Evidence
2026-11-10
A transformer-based model showed that lipid nanoparticle performance depends on interacting components, composition and formulation context. The findings support formulation-specific ranking in defined assay systems, but they do not establish any performance, incorporation or biological effect for 5-Methoxy-UTP itself.
What COMET Shows About LNP Design
2026-11-09
A critical interpretation of COMET’s multi-component LNP predictions, including its strongest benchmarks, cell-context effects, mechanistic implications, and unresolved limits for cargo and therapeutic translation.
COMET and the Design of Multicomponent LNPs
2026-11-09
A literature-focused analysis of how COMET models lipid nanoparticle efficacy from component identities, composition and formulation-wide features. The study shows that transformer-based ranking can identify high-performing formulations across cell contexts, while also revealing important limits: luciferase expression is an integrated downstream readout, in vitro performance does not reliably predict in vivo behavior, and model discrimination weakens near already-strong formulations.
Mechanistic Reading of LNP Efficacy Models
2026-11-08
A transformer-based analysis shows why lipid nanoparticle performance depends on interacting component identities, compositions, formulation-level variables and cellular context. Its results support formulation-wide prediction and reveal important limits: luciferase expression is a composite endpoint, in vitro rankings do not guarantee in vivo performance, and the evidence does not establish a role for 5-Methoxy-UTP itself.
What LNP Models Actually Predict
2026-11-07
A mechanistic editorial on how COMET links lipid identity, formulation context and cell-specific RNA expression, and why its results should not be interpreted as direct evidence for any particular nucleotide material.
What COMET Reveals About LNP Design
2026-11-06
A literature-focused analysis of how COMET models lipid nanoparticles as multi-component systems, what its reported prediction results establish, and where its transferability remains limited.
What COMET Reveals About LNP Formulation Biology
2026-11-06
A transformer-based model shows that lipid nanoparticle performance emerges from interactions among lipid identity, composition, formulation context and cell type. The evidence supports formulation-wide prediction of messenger RNA expression, while leaving the biological effects of 5-Methoxy-UTP itself untested.