We present a Drosophila model demonstrating that repeated cycles of tissue regeneration following localized apoptosis can induce tumorigenesis in the absence of oncogene activation or tumor suppressor loss. Chronic regenerative stress leads to epithelial disorganization, loss of cell polarity, immune cell infiltration, DNA damage, centrosomal defects, and aneuploidy–classical hallmarks of cancer. A subset of ”walking dead” cells resists apoptosis and contributes to tissue overgrowth and heterogeneity. Single-cell transcriptomics identifies a distinct cell population, absent from normal tissue, emerging under both regenerative and tumorigenic conditions and characterized by enrichment in JNK pathway components. The tumorigenic signature specifically features distinctive JNK regulatory elements while excluding chaperone proteins. Notably, inhibition of JNK signaling or modulation of DNA repair suppresses neoplastic transformation. This model reveals how chronic injury and defective regeneration drive cancer initiation, establishing a mechanistic link between wound healing and tumorigenesis, and providing a powerful system to explore early tumor development and therapeutic intervention.
The Hidden Power of Normalization: Exponential Capacity Control in Deep Neural Networks
arXiv:2511.00958v1 Announce Type: cross Abstract: Normalization methods are fundamental components of modern deep neural networks (DNNs). Empirically, they are known to stabilize optimization dynamics and