Single-cell RNA-seq (scRNA-seq) enables a quantitative cell-type characterisation based on global transcriptome profiles. We present Single-Cell Consensus Clustering (SC3), a user-friendly tool for unsupervised clustering which achieves high accuracy and robustness by combining multiple clustering solutions through a consensus approach. We demonstrate that SC3 is capable of identifying subclones based on the transcriptomes from neoplastic cells collected from patients.

As a stress response, senescence is a dynamic process involving multiple effector mechanisms whose combination determines the phenotypic quality. Here we identify autophagy as a new effector mechanism of senescence. Autophagy is activated during senescence and its activation is correlated with negative feedback in the PI3K-mammalian target of rapamycin (mTOR) pathway. A subset of autophagy-related genes are up-regulated during senescence: Overexpression of one of those genes, ULK3, induces autophagy and senescence. Furthermore, inhibition of autophagy delays the senescence phenotype, including senescence-associated secretion. Our data suggest that autophagy, and its consequent protein turnover, mediate the acquisition of the senescence phenotype.Supplemental material is available at http://www.genesdev.org.Received December 24, 2008; revised version accepted February 11, 2009. Cellular senescence is a state of stable cell cycle arrest with active metabolism. Similar to apoptosis, senescence can be a failsafe program against a variety of cellular insults. In contrast to apoptosis, however, in which the cytotoxic signals converge to a common mechanism, senescence is typically a delayed stress response involving multiple effector mechanisms. These effector mechanisms include epigenetic regulation (Narita et al. 2006;Adams 2007), the DNA damage response (Bartkova et al. 2006;Di Micco et al. 2006;Mallette et al. 2007), and the senescence-associated secretion phenotype (Kortlever et al. 2006;Acosta et al. 2008;Coppé et al. 2008;Kuilman et al. 2008;Wajapeyee et al. 2008). The relative contribution of these effectors varies depending on the trigger and cell type.Oncogene-induced senescence (OIS) illustrates well the tumor-suppressive role of senescence (Collado and Serrano 2006). The initial phenotype of oncogene induction is a highly proliferative state, which mimics transformation. However, this mitotic burst is gradually replaced by senescence. Although it has been proposed that global and progressive epigenetic alterations play a crucial role in OIS (Narita et al. 2006), the precise mechanism by which cells can achieve such a dramatic change is still unclear.Autophagy is a genetically regulated program responsible for the turnover of cellular proteins and damaged or superfluous organelles. This evolutionarily conserved process is characterized by the formation of doublemembrane cytosolic vesicles, autophagosomes, which sequester cytoplasmic content and deliver it to lysosomes (Ohsumi 2001;Klionsky et al. 2007;Mizushima et al. 2008). Autophagy is often associated with acute metabolic changes and rapid protein replacement. For example, autophagy is required for preimplantation development, where maternal proteins are recycled by autophagy (Tsukamoto et al. 2008). Autophagy is also required for survival in the early neonatal starvation period (Kuma et al. 2004;Komatsu et al. 2005). In addition to these physiological conditions, cytotoxic stimuli can also activate autophagy, but its precise role as a stress response...

Treatment-refractory rheumatoid arthritis (RA) is a major clinical challenge. Drug-free remission is uncommon but provides proof-of-concept that articular immune-homeostasis can be reinstated. Here we identify active cellular and molecular mechanisms of sustained remission in RA. Single-cell transcriptomics (32,000 cells) identified phenotypic changes in synovial tissue macrophages (STM) spanning health, early/active RA, treatment-refractory/active RA, and RA in sustained remission. Each clinical state is characterised by different frequencies of 9 discrete phenotypes of 4 distinct STM subpopulations with diverse homeostatic, regulatory and inflammatory functions. This cellular atlas combined with deep-phenotypic, spatial and functional analyses of synovial biopsy FACSsorted STMs revealed two STM subpopulations (MerTK/TREM2 high and MerTK/FOLR2/LYVE1 pos ) with unique remission transcriptomic signatures enriched in negative regulators of inflammation. In response to damage signals these cells are potent producers of inflammation-resolving lipid mediators and are the only STMs that induce the repair response of synovial fibroblasts. A low proportion of MerTK pos STMs in remission RA is a prognostic biomarker predictive of flare after treatment cessation. Therapeutic enhancement of MerTK pos STM-subsets could encourage resolution of inflammation and reinstate synovial homeostasis in inflammatory arthritis.

Using single-cell RNA-seq (scRNA-seq), the full transcriptome of individual cells can be acquired, enabling a quantitative cell-type characterisation based on expression profiles. However, due to the large variability in gene expression, identifying cell types based on the transcriptome remains challenging. We present Single-Cell Consensus Clustering (SC3), a tool for unsupervised clustering of scRNA-seq data. SC3 achieves high accuracy and robustness by consistently integrating different clustering solutions through a consensus approach. Tests on twelve published datasets show that SC3 outperforms five existing methods while remaining scalable, as shown by the analysis of a large dataset containing 44,808 cells. Moreover, an interactive graphical implementation makes SC3 accessible to a wide audience of users, and SC3 aids biological interpretation by identifying marker genes, differentially expressed genes and outlier cells. We illustrate the capabilities of SC3 by characterising newly obtained transcriptomes from subclones of neoplastic cells collected from patients.