Translated title. Original: Nobelpreisträger Demis Hassabis präsentiert bahnbrechendes Meisterwerk AlphaGenome 2: 98 % der menschlichen genetischen Dunklen Materie erfolgreich entschlüsselt
The text claims Google DeepMind has introduced AlphaGenome 2, a model that predicts how DNA mutations affect gene activity at single-cell resolution. It reports improved benchmark results and describes a free library of mutation-effect scores, but the supplied text ends mid-sentence.
Google DeepMind has introduced AlphaGenome 2, an AI model designed to predict how DNA changes affect different cells. The work may help researchers study disease causes and identify possible treatments.
Google DeepMind’s AlphaGenomes 2 predicts genetic variant effects using DNA sequence and cell-specific conditions. On a single-cell immune dataset, it improved causal-variant identification and gene-expression direction predictions over its previous version; the results are not measures of clinical accuracy.
Google DeepMind’s AlphaGenome 2 predicts how DNA variants affect specific cell types, helping researchers link genetic findings to disease causes. It is available for non-commercial research through the AlphaGenome API.
Translated title. Original: OpenAI patauge encore en annonçant des « découvertes » scientifiques sans en être sûr - Next
OpenAI and Anthropic face scrutiny for releasing unverified scientific claims and a large volume of AI-generated preprints, with concerns over peer review, proper attribution, and publication practices across mathematics, computer science, and physics.
Translated title. Original: İnsan vücudunda ilk kez DNA düzenlemesi yapıldı: Ömür boyu ilaç kullanmaya son verebilir | TGRT Haber
Scientists achieved direct in-body DNA editing for Wilson disease using prime editing, aiming for a one-time, permanent treatment delivered to liver cells via lipid nanoparticles. While promising for reducing lifelong therapy costs and risks, experts caution that ensuring precise edits and targeted delivery remains challenging.
Translated title. Original: Forskare använder modell för att skapa genetiska ritningar av virus - Tyrone's info
Researchers at Stanford used a generative AI to propose genetic blueprints for microscopic viruses, a step that could enable design of new life forms in the future. While not creating fully formed life now, the work prompts ethical questions and consideration of future biotechnological capabilities.
Anthropic says an effort involving about 950 AI agents found CRISPR-like DNA in viruses after searching sequence databases for more than 21 hours. A later report said a University of Copenhagen scientist had studied the system for years and uploaded related information to Anthropic’s models.
Translated title. Original: Olihně působí jako tvorové z jiné planety. Umějí přepisovat instrukce pro vlastní bílkoviny. DNA přitom nechávají beze změny
Researchers examined RNA editing in squid, finding thousands of edits across hundreds of proteins that alter protein forms and functions, shedding light on how RNA-level regulation contributes to cellular complexity without changing DNA.
Carbon-X is an open DNA-reading model from Hugging Face, trained on DNA and biology data, that can discuss mutation effects in English and runs on a single 80 GB GPU under Apache 2.0. It performs competitively on DNA benchmarks, is not medical-grade, and may mispredict without full context.