Transcriptomic, epigenomic, and spatial metabolomic cell profiling redefines regional human kidney anatomy.

A large-scale multimodal atlas that includes major kidney regions is lacking. Here, we employed simultaneous high-throughput single-cell ATAC/RNA sequencing (SHARE-seq) and spatially resolved metabolomics to profile 54 human samples from distinct kidney anatomical regions. We generated transcriptomes of 446,267 cells and chromatin accessibility profiles of 401,875 cells and developed a package to analyze 408,218 spatially resolved metabolomes. We find that the same cell type, including thin limb, thick ascending limb loop of Henle and principal cells, display distinct transcriptomic, chromatin accessibility, and metabolomic signatures, depending on anatomic location. Surveying metabolism-associated gene profiles revealed non-overlapping metabolic signatures between nephron segments and dysregulated lipid metabolism in diseased proximal tubule (PT) cells. Integrating multimodal omics with clinical data identified PLEKHA1 as a disease marker, and its in vitro knockdown increased gene expression in PT differentiation, suggesting possible pathogenic roles. This study highlights previously underrepresented cellular heterogeneity underlying the human kidney anatomy.

Nomogram for predicting the risk of nonalcoholic fatty liver disease in older adults in Qingdao, China: A cross-sectional study.

BACKGROUND AND OBJECTIVES: To explore the risk factors for non-alcoholic fatty liver disease (NAFLD) and to establish a non-invasive tool for the screening of NAFLD in an older adult population. METHODS AND STUDY DESIGN: A total of 131,161 participants were included in this cross-sectional study. Participants were randomly divided into training and validation sets (7:3). The least absolute shrinkage and selection operator method was used to screen risk factors. Multivariate logistic regression was employed to develop a nomogram, which was made available online. Receiver operating characteristic curve analysis, calibration plots, and decision curve analysis were used to validate the discrimination, calibration, and clinical practicability of the nomogram. Sex and age subgroup analyses were conducted to further validate the reliability of the model. RESULTS: Nine variables were identified for inclusion in the nomogram (age, sex, waist circumference, body mass index, exercise frequency, systolic blood pressure, fasting plasma glucose, alanine aminotransferase, and low-density lipoprotein cholesterol). The area under the receiver operating characteristic curve values were 0.793 and 0.790 for the training set and the validation set, respectively. The calibration plots and decision curve analyses showed good calibration and clinical utility. Subgroup analyses demonstrated consistent discriminatory ability in different sex and age subgroups. CONCLUSIONS: This study established and validated a new nomogram model for evaluating the risk of NAFLD among older adults. The nomogram had good discriminatory performance and is a non-invasive and convenient tool for the screening of NAFLD in older adults.

TNIK depletion induces inflammation and apoptosis in injured renal proximal tubule epithelial cells.

In the aftermath of acute kidney injury (AKI), surviving proximal tubule epithelia repopulate injured tubules to promote repair. However, a portion of cells fail to repair [termed failed-repair proximal tubule cells (FR-PTCs)] and exert ongoing proinflammatory and profibrotic effects. To better understand the molecular drivers of the FR-PTC state, we reanalyzed a mouse ischemia-reperfusion injury single-nucleus RNA-sequencing (snRNA-seq) atlas to identify Traf2 and Nck interacting kinase (Tnik) to be exclusively expressed in FR-PTCs but not in healthy or acutely injured proximal tubules after AKI (2 and 6 wk) in mice. We confirmed expression of Tnik protein in injured mouse and human tissues by immunofluorescence. Then, to determine the functional role of Tnik in FR-PTCs, we depleted TNIK with siRNA in two human renal proximal tubule epithelial cell lines (primary and immortalized hRPTECs) and analyzed each by bulk RNA-sequencing. Pathway analysis revealed significant upregulation of inflammatory signaling pathways, whereas pathways associated with differentiated proximal tubules such as organic acid transport were significantly downregulated. TNIK gene knockdown drove reduced cell viability and increased apoptosis, including differentially expressed poly(ADP-ribose) polymerase (PARP) family members, cleaved PARP-1 fragments, and increased annexin V binding to phosphatidylserine. Together, these results indicate that Tnik upregulation in FR-PTCs acts in a compensatory fashion to suppress inflammation and promote proximal tubule epithelial cell survival after injury. Modulating TNIK activity may represent a prorepair therapeutic strategy after AKI.NEW & NOTEWORTHY The molecular drivers of successful and failed repair in the proximal tubule after acute kidney injury (AKI) are incompletely understood. We identified Traf2 and Nck interacting kinase (Tnik) to be exclusively expressed in failed-repair proximal tubule cells after AKI. We tested the effect of siTNIK depletion in two proximal tubule cell lines followed by bulk RNA-sequencing analysis. Our results indicate that TNIK acts to suppress inflammatory signaling and apoptosis in injured renal proximal tubule epithelial cells to promote cell survival.

Predicting proximal tubule failed repair drivers through regularized regression analysis of single cell multiomic sequencing.

Renal proximal tubule epithelial cells have considerable intrinsic repair capacity following injury. However, a fraction of injured proximal tubule cells fails to undergo normal repair and assumes a proinflammatory and profibrotic phenotype that may promote fibrosis and chronic kidney disease. The healthy to failed repair change is marked by cell state-specific transcriptomic and epigenomic changes. Single nucleus joint RNA- and ATAC-seq sequencing offers an opportunity to study the gene regulatory networks underpinning these changes in order to identify key regulatory drivers. We develop a regularized regression approach to construct genome-wide parametric gene regulatory networks using multiomic datasets. We generate a single nucleus multiomic dataset from seven adult human kidney samples and apply our method to study drivers of a failed injury response associated with kidney disease. We demonstrate that our approach is a highly effective tool for predicting key cis- and trans-regulatory elements underpinning the healthy to failed repair transition and use it to identify NFAT5 as a driver of the maladaptive proximal tubule state.

Electronic structure and magnetothermal properties of two-dimensional ScCl.

Two-dimensional ferromagnetic materials with intrinsic half-metallic properties have strong application advantages in nanoscale spintronics. Herein, density functional theory calculations show that monolayer ScCl is a ferromagnetic metallic material when undoped (n = 0), and the transition from metal to half-metal occurs with the continuous doping of holes. On the contrary, as the concentration of doped electrons increases, the system will exhibit metallic characteristics, which is particularly evident from a variation in spin polarizability. Furthermore, we have discussed how doped carriers affect the shape of the Fermi surface and the Fermi velocity of electrons. Most importantly, Monte Carlo simulations show that the ScCl monolayer is particularly regulated by carrier concentration (n) and magnetic field (h). Additionally, trends in energy and magnetic exchange coupling in different magnetic configurations (AFM phase and FM phase) with different doping concentrations are presented. When n < -0.16, the material is not only a half-metallic material that easily flips the magnetic axis, but also proves to be a candidate ferromagnetic material that works stably at room temperature in terms of dynamic stability. In addition, the origin of magnetocrystalline anisotropy is analyzed, and the contribution of different orbitals to spin-orbit coupling is presented. Moreover, we note that when magnetic field is small (h < 1 T), the change in size has a significant effect on ferromagnetic phase transition. However, when the system size is large (size >15 nm), TC is less sensitive to magnetic field. In addition, hole doping and size effect will greatly affect the hC of the system, but when the hole doping exceeds the critical value (n = -0.16), its influence on the hysteresis loop is no longer obvious. These interesting magnetic phenomena and easily adjustable physical properties show us that monolayer ScCl will be a promising functional material.

Epigenetic reprogramming driving successful and failed repair in acute kidney injury.

Acute kidney injury (AKI) causes epithelial damage followed by subsequent repair. While successful repair restores kidney function, this process is often incomplete and can lead to chronic kidney disease (CKD) in a process called failed repair. To better understand the epigenetic reprogramming driving this AKI-to-CKD transition we generated a single nucleus multiomic atlas for the full mouse AKI time course, consisting of ~280,000 single nucleus transcriptomes and epigenomes. We reveal cell-specific dynamic alterations in gene regulatory landscapes reflecting especially activation of proinflammatory pathways. We further generated single nucleus multiomic data from four human AKI samples including validation by genome-wide identification of NF-kB binding sites. A regularized regression analysis identifies key regulators involved in both successful and failed repair cell fate, identifying the transcription factor CREB5 as a regulator of both successful and failed tubular repair that also drives proximal tubule cell proliferation after injury. Our interspecies multiomic approach provides a foundation to comprehensively understand cell states in AKI.

High resolution spatial profiling of kidney injury and repair using RNA hybridization-based in situ sequencing.

Emerging spatially resolved transcriptomics technologies allow for the measurement of gene expression in situ at cellular resolution. We apply direct RNA hybridization-based in situ sequencing (dRNA HybISS, Cartana part of 10xGenomics) to compare male and female healthy mouse kidneys and the male kidney injury and repair timecourse. A pre-selected panel of 200 genes is used to identify cell state dynamics patterns during injury and repair. We develop a new computational pipeline, CellScopes, for the rapid analysis, multi-omic integration and visualization of spatially resolved transcriptomic datasets. The resulting dataset allows us to resolve 13 kidney cell types within distinct kidney niches, dynamic alterations in cell state over the course of injury and repair and cell-cell interactions between leukocytes and kidney parenchyma. At late timepoints after injury, C3+ leukocytes are enriched near pro-inflammatory, failed-repair proximal tubule cells. Integration of snRNA-seq dataset from the same injury and repair samples also allows us to impute the spatial localization of genes not directly measured by dRNA HybISS.

Sequencing of Physically Interacting Cells in Human Kidney Allograft Rejection to Infer Contact-dependent Immune Cell Transcription.

BACKGROUND: Rejection requires cell-cell contact involving immune cells. Inferring the transcriptional programs of cell-cell interactions from single-cell RNA-sequencing (scRNA-seq) data is challenging as spatial information is lost. METHODS: We combined a CD45 pos enrichment strategy with Cellular Indexing of Transcriptomes and Epitopes by sequencing based quantification of leukocyte surface proteins to analyze cell-cell interactions in 11 human kidney transplant biopsies encompassing a spectrum of rejection diagnoses. scRNA-seq was performed using the 10X Genomics platform. We applied the sequencing physically interacting cells computational method to deconvolute the transcriptional profiles of heterotypic physically interacting cells. RESULTS: The 11 human allograft biopsies generated 31 203 high-quality single-cell libraries. Clustering was further refined by combining Cellular Indexing of Transcriptomes and Epitopes by sequencing data from 6 different leukocyte-specific surface proteins. Three of 6 doublet clusters were identified as physically interacting cell complexes; macrophages or dendritic cells bound to B cells or plasma cells; natural killer (NK) or T cells bound to macrophages or dendritic cells and NK or T cells bound to endothelial cells. Myeloid-lymphocyte physically interacting cell complexes expressed activated and proinflammatory genes. Lymphocytes physically interacting with endothelial cells were enriched for NK and CD4 T cells. NK cell-endothelial cell contact caused increased expression of endothelial proinflammatory genes CXCL9 and CXCL10 and NK cell proinflammatory genes CCL3 , CCL4 , and GNLY . CONCLUSIONS: The transcriptional profiles of physically interacting cells from human kidney transplant biopsies can be inferred from scRNA-seq data using the sequencing physically interacting cells method. This approach complements previous methods that estimate cell-cell physical contact from scRNA-seq data.

Genetically engineering endothelial niche in human kidney organoids enables multilineage maturation, vascularization and de novo cell types.

Vascularization plays a critical role in organ maturation and cell type development. Drug discovery, organ mimicry, and ultimately transplantation in a clinical setting thereby hinges on achieving robust vascularization of in vitro engineered organs. Here, focusing on human kidney organoids, we overcome this hurdle by combining an inducible ETS translocation variant 2 (ETV2) human induced pluripotent stem cell (iPSC) line, which directs endothelial fate, with a non-transgenic iPSC line in suspension organoid culture. The resulting human kidney organoids show extensive vascularization by endothelial cells with an identity most closely related to endogenous kidney endothelia. Vascularized organoids also show increased maturation of nephron structures including more mature podocytes with improved marker expression, foot process interdigitation, an associated fenestrated endothelium, and the presence of renin+ cells. The creation of an engineered vascular niche capable of improving kidney organoid maturation and cell type complexity is a significant step forward in the path to clinical translation. Furthermore, this approach is orthogonal to native tissue differentiation paths, hence readily adaptable to other organoid systems and thus has the potential for a broad impact on basic and translational organoid studies.

A single-cell multiomic analysis of kidney organoid differentiation.

Kidney organoids differentiated from pluripotent stem cells are powerful models of kidney development and disease but are characterized by cell immaturity and off-target cell fates. Comparing the cell-specific gene regulatory landscape during organoid differentiation with human adult kidney can serve to benchmark progress in differentiation at the epigenome and transcriptome level for individual organoid cell types. Using single-cell multiome and histone modification analysis, we report more broadly open chromatin in organoid cell types compared to the human adult kidney. We infer enhancer dynamics by cis-coaccessibility analysis and validate an enhancer driving transcription of HNF1B by CRISPR interference both in cultured proximal tubule cells and also during organoid differentiation. Our approach provides an experimental framework to judge the cell-specific maturation state of human kidney organoids and shows that kidney organoids can be used to validate individual gene regulatory networks that regulate differentiation.

Drug­resistant Acinetobacter baumannii: From molecular mechanisms to potential therapeutics (Review).

Bacterial drug resistance is increasingly becoming an important problem that needs to be solved urgently in modern clinical practices. Infection caused by Acinetobacter baumannii is a serious threat to the life and health of patients. The drug resistance rate of Acinetobacter baumannii strains is increasing, thus research on the drug resistance of Acinetobacter baumannii has also seen an increase. When patients are infected with drug-resistant Acinetobacter baumannii, the availability of suitable antibiotics commonly used in clinical practices is becoming increasingly limited and the prognosis of patients is worsening. Studying the molecular mechanism of the drug resistance of Acinetobacter baumannii is fundamental to solving the problem of drug-resistant Acinetobacter baumannii and potentially other 'super bacteria'. Drug resistance mechanisms primarily include enzymes, membrane proteins, efflux pumps and beneficial mutations. Research on the underlying mechanisms provides a theoretical basis for the use and development of antibiotics and the development of novel treatment methods.

Ferromagnetic vanadium disulfide VS2 monolayers with high Curie temperature and high spin polarization.

The structural, electronic, and magnetic properties of vanadium disulfide VS2 monolayers were investigated using first-principles calculations and Monte Carlo (MC) simulations. The results of molecular dynamics simulations and phonon dispersion showed that the VS2 monolayer has good dynamic and thermodynamic stabilities. Based on the results of the band structure, we also explore the effect of carrier concentrations on the spin gap, spin polarization and the direction of the easy magnetic axis. Our results demonstrated that doping an appropriate amount of holes can cause the reversal of the easy magnetic axis and maintain nearly 100% spin polarization, which greatly improves the application possibility of the VS2 monolayer as a spintronic device. The contribution of different orbits to the spin-orbit coupling (SOC) effect is given in magnetocrystalline anisotropy energy, which provides a theoretical basis for explaining the origin of magnetic crystal anisotropy. Based on the MC simulations, we also showed the influences of different parameters (carrier concentrations, magnetic field and crystal field) on the magnetothermal properties of the VS2 monolayer. It is found that the increase of hole doping concentrations can promote the increase of the Curie temperature, while the increase of electron doping concentrations will greatly weaken the Curie temperature. Furthermore, according to the influences of different parameters on the Curie temperature and spin polarization, we conclude that a suitably enhanced magnetic field and appropriate hole concentrations will not only make the system maintain high spin polarization, but also make the system exhibit ferromagnetic properties above room temperature.

Lineage Tracing and Single-Nucleus Multiomics Reveal Novel Features of Adaptive and Maladaptive Repair after Acute Kidney Injury.

SIGNIFICANCE STATEMENT: Understanding the mechanisms underlying adaptive and maladaptive renal repair after AKI and their long-term consequences is critical to kidney health. The authors used lineage tracing of cycling cells and single-nucleus multiomics (profiling transcriptome and chromatin accessibility) after AKI. They demonstrated that AKI triggers a cell-cycle response in most epithelial and nonepithelial kidney cell types. They also showed that maladaptive proinflammatory proximal tubule cells (PTCs) persist until 6 months post-AKI, although they decreased in abundance over time, in part, through cell death. Single-nucleus multiomics of lineage-traced cells revealed regulatory features of adaptive and maladaptive repair. These included activation of cell state-specific transcription factors and cis-regulatory elements, and effects in PTCs even after adaptive repair, weeks after the injury event. BACKGROUND: AKI triggers a proliferative response as part of an intrinsic cellular repair program, which can lead to adaptive renal repair, restoring kidney structure and function, or maladaptive repair with the persistence of injured proximal tubule cells (PTCs) and an altered kidney structure. However, the cellular and molecular understanding of these repair programs is limited. METHODS: To examine chromatin and transcriptional responses in the same cell upon ischemia-reperfusion injury (IRI), we combined genetic fate mapping of cycling ( Ki67+ ) cells labeled early after IRI with single-nucleus multiomics-profiling transcriptome and chromatin accessibility in the same nucleus-and generated a dataset of 83,315 nuclei. RESULTS: AKI triggered a broad cell cycle response preceded by cell type-specific and global transcriptional changes in the nephron, the collecting and vascular systems, and stromal and immune cell types. We observed a heterogeneous population of maladaptive PTCs throughout proximal tubule segments 6 months post-AKI, with a marked loss of maladaptive cells from 4 weeks to 6 months. Gene expression and chromatin accessibility profiling in the same nuclei highlighted differences between adaptive and maladaptive PTCs in the activity of cis-regulatory elements and transcription factors, accompanied by corresponding changes in target gene expression. Adaptive repair was associated with reduced expression of genes encoding transmembrane transport proteins essential to kidney function. CONCLUSIONS: Analysis of genome organization and gene activity with single-cell resolution using lineage tracing and single-nucleus multiomics offers new insight into the regulation of renal injury repair. Weeks to months after mild-to-moderate IRI, maladaptive PTCs persist with an aberrant epigenetic landscape, and PTCs exhibit an altered transcriptional profile even following adaptive repair.

Immune cell heterogeneity in a mouse model of diabetic kidney disease.

Typical kidney single-cell RNA-sequencing contains relatively few leukocytes, complicating efforts to understand how immune cells impact kidney disease progression. In this issue, Fu et al. use a flow sorting strategy to generate a very large immune cell single-cell RNA-sequencing atlas in a mouse model of diabetic kidney disease. These findings highlight the importance of leukocyte cell subtypes in diabetic kidney disease.

Comprehensive single-cell transcriptional profiling defines shared and unique epithelial injury responses during kidney fibrosis.

The underlying cellular events driving kidney fibrogenesis and metabolic dysfunction are incompletely understood. Here, we employed single-cell combinatorial indexing RNA sequencing to analyze 24 mouse kidneys from two fibrosis models. We profiled 309,666 cells in one experiment, representing 50 cell types/states encompassing epithelial, endothelial, immune, and stromal populations. Single-cell analysis identified diverse injury states of the proximal tubule, including two distinct early-phase populations with dysregulated lipid and amino acid metabolism, respectively. Lipid metabolism was defective in the chronic phase but was transiently activated in the very early stages of ischemia-induced injury, where we discovered increased lipid deposition and increased fatty acid ß-oxidation. Perilipin 2 was identified as a surface marker of intracellular lipid droplets, and its knockdown in vitro disrupted cell energy state maintenance during lipid accumulation. Surveying epithelial cells across nephron segments identified shared and unique injury responses. Stromal cells exhibited high heterogeneity and contributed to fibrogenesis by epithelial-stromal crosstalk.

Defining cellular complexity in human autosomal dominant polycystic kidney disease by multimodal single cell analysis.

Autosomal dominant polycystic kidney disease (ADPKD) is the leading genetic cause of end stage renal disease characterized by progressive expansion of kidney cysts. To better understand the cell types and states driving ADPKD progression, we analyze eight ADPKD and five healthy human kidney samples, generating single cell multiomic atlas consisting of ~100,000 single nucleus transcriptomes and ~50,000 single nucleus epigenomes. Activation of proinflammatory, profibrotic signaling pathways are driven by proximal tubular cells with a failed repair transcriptomic signature, proinflammatory fibroblasts and collecting duct cells. We identify GPRC5A as a marker for cyst-lining collecting duct cells that exhibits increased transcription factor binding motif availability for NF-κB, TEAD, CREB and retinoic acid receptors. We identify and validate a distal enhancer regulating GPRC5A expression containing these motifs. This single cell multiomic analysis of human ADPKD reveals previously unrecognized cellular heterogeneity and provides a foundation to develop better diagnostic and therapeutic approaches.

Multimodal single cell sequencing implicates chromatin accessibility and genetic background in diabetic kidney disease progression.

The proximal tubule is a key regulator of kidney function and glucose metabolism. Diabetic kidney disease leads to proximal tubule injury and changes in chromatin accessibility that modify the activity of transcription factors involved in glucose metabolism and inflammation. Here we use single nucleus RNA and ATAC sequencing to show that diabetic kidney disease leads to reduced accessibility of glucocorticoid receptor binding sites and an injury-associated expression signature in the proximal tubule. We hypothesize that chromatin accessibility is regulated by genetic background and closely-intertwined with metabolic memory, which pre-programs the proximal tubule to respond differently to external stimuli. Glucocorticoid excess has long been known to increase risk for type 2 diabetes, which raises the possibility that glucocorticoid receptor inhibition may mitigate the adverse metabolic effects of diabetic kidney disease.

Spatially resolved transcriptomics and the kidney: many opportunities.

Defining changes in gene expression during health and disease is critical for the understanding of human physiology. In recent years, single-cell/nuclei RNA sequencing (sc/snRNAseq) has revolutionized the definition and discovery of cell types and states as well as the interpretation of organ- and cell-type-specific signaling pathways. However, these advances require tissue dissociation to the level of the single cell or single nuclei level. Spatially resolved transcriptomics (SrT) now provides a platform to overcome this barrier in understanding the physiological contexts of gene expression and cellular microenvironment changes in development and disease. Some of these transcriptomic tools allow for high-resolution mapping of hundreds of genes simultaneously in cellular and subcellular compartments. Other tools offer genome depth mapping but at lower resolution. We review advances in SrT, considerations for using SrT in your own research, and applications for kidney biology.

Mapping the single-cell transcriptomic response of murine diabetic kidney disease to therapies.

Diabetic kidney disease (DKD) occurs in ∼40% of patients with diabetes and causes kidney failure, cardiovascular disease, and premature death. We analyzed the response of a murine DKD model to five treatment regimens using single-cell RNA sequencing (scRNA-seq). Our atlas of ∼1 million cells revealed a heterogeneous response of all kidney cell types both to DKD and its treatment. Both monotherapy and combination therapies targeted differing cell types and induced distinct and non-overlapping transcriptional changes. The early effects of sodium-glucose cotransporter-2 inhibitors (SGLT2i) on the S1 segment of the proximal tubule suggest that this drug class induces fasting mimicry and hypoxia responses. Diabetes downregulated the spliceosome regulator serine/arginine-rich splicing factor 7 (Srsf7) in proximal tubule that was specifically rescued by SGLT2i. In vitro proximal tubule knockdown of Srsf7 induced a pro-inflammatory phenotype, implicating alternative splicing as a driver of DKD and suggesting SGLT2i regulation of proximal tubule alternative splicing as a potential mechanism of action for this drug class.

Spatially Resolved Transcriptomic Analysis of Acute Kidney Injury in a Female Murine Model.

BACKGROUND: Single-cell sequencing technologies have advanced our understanding of kidney biology and disease, but the loss of spatial information in these datasets hinders our interpretation of intercellular communication networks and regional gene expression patterns. New spatial transcriptomic sequencing platforms make it possible to measure the topography of gene expression at genome depth. METHODS: We optimized and validated a female bilateral ischemia-reperfusion injury model. Using the 10× Genomics Visium Spatial Gene Expression solution, we generated spatial maps of gene expression across the injury and repair time course, and applied two open-source computational tools, Giotto and SPOTlight, to increase resolution and measure cell-cell interaction dynamics. RESULTS: An ischemia time of 34 minutes in a female murine model resulted in comparable injury to 22 minutes for males. We report a total of 16,856 unique genes mapped across our injury and repair time course. Giotto, a computational toolbox for spatial data analysis, enabled increased resolution mapping of genes and cell types. Using a seeded nonnegative matrix regression (SPOTlight) to deconvolute the dynamic landscape of cell-cell interactions, we found that injured proximal tubule cells were characterized by increasing macrophage and lymphocyte interactions even 6 weeks after injury, potentially reflecting the AKI to CKD transition. CONCLUSIONS: In this transcriptomic atlas, we defined region-specific and injury-induced loss of differentiation markers and their re-expression during repair, as well as region-specific injury and repair transcriptional responses. Lastly, we created an interactive data visualization application for the scientific community to explore these results (http://humphreyslab.com/SingleCell/).