iKraph AI and the AdaptAge Epigenetic Clock

NotebookLM: "...iKraph [is] a massive biomedical knowledge graph designed to improve automated discovery by organizing unstructured scientific literature into structured data. Researchers utilized a top-performing natural language processing pipeline to extract over 30 million relations from PubMed abstracts, integrating them with data from 40 public databases and genomic sequences. A key innovation is the Probabilistic Semantic Reasoning (PSR) algorithm, which identifies indirect causal links between biological entities to suggest new uses for existing medications. The study demonstrates the system's effectiveness through real-time drug repurposing for COVID-19 and cystic fibrosis, achieving high accuracy comparable to human experts. Ultimately, this platform provides a comprehensive digital map of medical knowledge to help scientists navigate the information explosion and accelerate therapeutic breakthroughs." "[AdaptAge is] a novel framework for measuring biological age by distinguishing between deleterious molecular damage and beneficial adaptive responses within the epigenome. While traditional epigenetic clocks primarily identify correlations between DNA methylation and chronological age, the authors utilize Mendelian randomization to establish causal links between specific CpG sites and aging-related health outcomes. By categorizing these sites, they developed three distinct models: DamAge, which tracks harmful changes; AdaptAge, which measures protective adaptations; and CausAge, which combines both to estimate overall mortality risk. Their findings demonstrate that DamAge is a superior predictor of mortality and better captures the rejuvenation effects of cellular reprogramming, whereas AdaptAge identifies physiological efforts to maintain health. Ultimately, this work suggests that high-precision aging interventions must target causal damage rather than simply altering all age-related biomarkers, some of which may actually be helping the body survive." https://doi.org/10.1038/s42256-025-01014-w https://doi.org/10.1038/s43587-023-00557-0