This is a repository copy of Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer.

Kidney fibrosis is the hallmark of chronic kidney disease progression, however, currently no antifibrotic therapies exist. This is largely because the origin, functional heterogeneity and regulation of scar-forming cells during human kidney fibrosis remains poorly understood. Here, using single cell RNA-seq, we profiled the transcriptomes of proximal tubule and non-proximal tubule cells in healthy and fibrotic human kidneys to map the entire human kidney in an unbiased approach. This enabled mapping of all matrix-producing cells at high resolution, revealing distinct subpopulations of pericytes and fibroblasts as the major cellular sources of scar forming myofibroblasts during human kidney fibrosis. We used genetic fate-tracing, time-course single cell RNA-seq and ATAC-seq experiments in mice, and spatial transcriptomics in human kidney fibrosis to functionally interrogate these findings, shedding new light on the origin, heterogeneity and differentiation of human kidney myofibroblasts and their fibroblast and pericyte precursors at unprecedented resolution. Finally, we used this strategy to facilitate target discovery, identifying Nkd2 as a myofibroblast-specific target in human kidney fibrosis.

Renal fibrosis is the common end point of virtually all progressive kidney diseases. Renal fibrosis should not be viewed as a simple and uniform 'scar', but rather as a dynamic system that involves extracellular matrix components and many, if not all, renal and infiltrating cell types. The involved cells exhibit enormous plasticity or phenotypic variability-a fact that we are only beginning to appreciate. Only a detailed understanding of the underlying mechanisms of renal fibrosis can facilitate the development of effective treatments. In this Review, we discuss the most recent advances in renal, or more specifically, tubulointerstitial fibrosis. Novel mechanisms as well as potential treatment targets based on different cell types are described. Problems that continue to plague the field are also discussed, including specific therapeutic targeting of the kidney, the development of improved diagnostic methods to assess renal fibrosis and the shortcomings of available animal models.

This is a repository copy of Pan-cancer image-based detection of clinically actionable genetic alterations.

No abstract