A Delaware judge ruled that the federal government’s license to university inventions does not automatically protect a private supplier. Moderna’s separate government-contract defense—and the underlying infringement claims—remain unresolved.
A compact deep-learning model distinguished young from aged mouse blood stem cells by reading subtle patterns in DAPI-stained nuclei. Its most immediate value is as a research-screening tool—not a clinical age test or proof of cellular rejuvenation.
Researchers carbonized a laboratory-made papyrus scroll, detected lead-spiked writing with handheld X-ray fluorescence and reconstructed readable text from CT data. The proof of concept could help identify which surviving Herculaneum scrolls are best suited for virtual unrolling.
A programmable trapped-ion device reproduced a simplified model of particle-pair formation and tracked the process in space and time. The experiment is a physics benchmark—not yet a quantum advantage demonstration.
A population-scale atlas of the human prefrontal cortex links cell-specific gene activity across Alzheimer’s, Parkinson’s, schizophrenia and five other conditions—but it is a research map, not a diagnostic test or treatment.
Anthropic says a coordinated group of Claude agents found a reverse-transcriptase system with CRISPR-like repeat arrays, then human scientists confirmed its molecular components in the lab. Its biological function remains unknown.
Fourteen research teams will test whether gene editing, RNA treatments and engineered microbes can help threatened species withstand disease and climate stress — under a contained-research rule that bars releases into the wild during the programme.