This paper provides a comprehensive study of Federated Learning (FL) with an emphasis on enabling software and hardware platforms, protocols, real-life applications and use-cases. FL can be applicable to multiple domains but applying it to different industries has its own set of obstacles. FL is known as collaborative learning, where algorithm(s) get trained across multiple devices or servers with decentralized data samples without having to exchange the actual data. This approach is radically different from other more established techniques such as getting the data samples uploaded to servers or having data in some form of distributed infrastructure. FL on the other hand generates more robust models without sharing data, leading to privacy-preserved solutions with higher security and access privileges to data. This paper starts by providing an overview of FL. Then, it gives an overview of technical details that pertain to FL enabling technologies, protocols, and applications. Compared to other survey papers in the field, our objective is to provide a more thorough summary of the most relevant protocols, platforms, and real-life use-cases of FL to enable data scientists to build better privacy-preserving solutions for industries in critical need of FL. We also provide an overview of key challenges presented in the recent literature and provide a summary of related research work. Moreover, we explore both the challenges and advantages of FL and present detailed service use-cases to illustrate how different architectures and protocols that use FL can fit together to deliver desired results.

Heterogeneous mental disorders such as Autism Spectrum Disorder (ASD) are notoriously difficult to diagnose, especially in children. The current psychiatric diagnostic process is based purely on the behavioral observation of symptomology (DSM-5/ICD-10) and may be prone to misdiagnosis. In order to move the field toward more quantitative diagnosis, we need advanced and scalable machine learning infrastructure that will allow us to identify reliable biomarkers of mental health disorders. In this paper, we propose a framework called ASD-DiagNet for classifying subjects with ASD from healthy subjects by using only fMRI data. We designed and implemented a joint learning procedure using an autoencoder and a single layer perceptron (SLP) which results in improved quality of extracted features and optimized parameters for the model. Further, we designed and implemented a data augmentation strategy, based on linear interpolation on available feature vectors, that allows us to produce synthetic datasets needed for training of machine learning models. The proposed approach is evaluated on a public dataset provided by Autism Brain Imaging Data Exchange including 1, 035 subjects coming from 17 different brain imaging centers. Our machine learning model outperforms other state of the art methods from 10 imaging centers with increase in classification accuracy up to 28% with maximum accuracy of 82%. The machine learning technique presented in this paper, in addition to yielding better quality, gives enormous advantages in terms of execution time (40 min vs. 7 h on other methods). The implemented code is available as GPL license on GitHub portal of our lab (https://github.com/pcdslab/ASD-DiagNet).

G protein-coupled receptors (GPCRs) regulate diverse physiological processes, and many human diseases are due to defects in GPCR signaling. To identify the dynamic response of a signaling network downstream from a prototypical G s -coupled GPCR, the vasopressin V2 receptor, we have carried out multireplicate, quantitative phosphoproteomics with iTRAQ labeling at four time points following vasopressin exposure at a physiological concentration in cells isolated from rat kidney. A total of 12,167 phosphopeptides were identified from 2,783 proteins, with 273 changing significantly in abundance with vasopressin. Two-dimensional clustering of phosphopeptide time courses and Gene Ontology terms revealed that ligand binding to the V2 receptor affects more than simply the canonical cyclic adenosine monophosphateprotein kinase A and arrestin pathways under physiological conditions. The regulated proteins included key components of actin cytoskeleton remodeling, cell-cell adhesion, mitogen-activated protein kinase signaling, Wnt/␤-catenin signaling, and apoptosis pathways. These data suggest that vasopressin can regulate an array of cellular functions well beyond its classical role in regulating water and solute transport. These results greatly expand the current view of GPCR signaling in a physiological context and shed new light on

No abstract

Vasopressin regulates water excretion, in part, by controlling the abundances of the water channel aquaporin-2 (AQP2) protein and regulatory proteins in the renal collecting duct. To determine whether vasopressin-induced alterations in protein abundance result from modulation of protein production, protein degradation, or both, we used protein mass spectrometry with dynamic stable isotope labeling in cell culture to achieve a proteome-wide determination of protein half-lives and relative translation rates in mpkCCD cells. Measurements were made at steady state in the absence or presence of the vasopressin analog, desmopressin (dDAVP). Desmopressin altered the translation rate rather than the stability of most responding proteins, but it significantly increased both the translation rate and the half-life of AQP2. In addition, proteins associated with vasopressin action, including Mal2, Akap12, gelsolin, myosin light chain kinase, annexin-2, and Hsp70, manifested altered translation rates. Interestingly, desmopressin increased the translation of seven glutathione S-transferase proteins and enhanced protein S-glutathionylation, uncovering a previously unexplored vasopressin-induced post-translational modification. Additional bioinformatic analysis of the mpkCCD proteome indicated a correlation between protein function and protein half-life. In particular, processes that are rapidly regulated, such as transcription, endocytosis, cell cycle regulation, and ubiquitylation are associated with proteins with especially short half-lives. These data extend our understanding of the mechanisms underlying vasopressin signaling and provide a broad resource for additional investigation of collecting duct function (http://helixweb.nih.gov/ESBL/Database/ ProteinHalfLives/index.html). 24: 179324: -180524: , 201324: . doi: 10.1681 Vasopressin controls water excretion by regulating the molecular water channel aquaporin-2 (AQP2) in two fundamental ways: (1) regulated trafficking, 1 seen in a time frame of 40 seconds to about 40 minutes, 2 and (2) regulation of AQP2 protein abundance, typically seen after many hours of exposure to vasopressin. 3,4 Studies in animal models of various water balance disorders have revealed that most clinically important water balance abnormalities are associated with dysregulation of AQP2 abundance rather than trafficking. 5 High levels of circulating vasopressin for periods of hours to days in animals are associated with increases in AQP2 mRNA in the kidney. 6,7 Such observations have led to the widely accepted view that vasopressin regulates AQP2 abundance by transcriptional mechanisms. [8][9][10][11] However, recent studies in cultured inner medullary collecting duct (IMCD) cells by Nedvetsky et al. 12 have shown J Am Soc Nephrol

A general question in molecular physiology is how to identify candidate protein kinases corresponding to a known or hypothetical phosphorylation site in a protein of interest. It is generally recognized that the amino acid sequence surrounding the phosphorylation site provides information that is relevant to identification of the cognate protein kinase. Here, we present a mass spectrometry-based method for profiling the target specificity of a given protein kinase as well as a computational tool for the calculation and visualization of the target preferences. The mass spectrometry-based method identifies sites phosphorylated in response to in vitro incubation of protein mixtures with active recombinant protein kinases followed by standard phosphoproteomic methodologies. The computational tool, called "PhosphoLogo," uses an information-theoretic algorithm to calculate position-specific amino acid preferences and anti-preferences from the mass-spectrometry data (http://helixweb.nih.gov/PhosphoLogo/). The method was tested using protein kinase A (catalytic subunit α), revealing the well-known preference for basic amino acids in positions -2 and -3 relative to the phosphorylated amino acid. It also provides evidence for a preference for amino acids with a branched aliphatic side chain in position +1, a finding compatible with known crystal structures of protein kinase A. The method was also employed to profile target preferences and anti-preferences for 15 additional protein kinases with potential roles in regulation of epithelial transport: CK2, p38, AKT1, SGK1, PKCδ, CaMK2δ, DAPK1, MAPKAPK2, PKD3, PIM1, OSR1, STK39/SPAK, GSK3β, Wnk1, and Wnk4.

Background: Few studies have explored dialysis patients’ perspectives on dialysis decision-making and end-of-life-care (EoLC) preferences. We surveyed a racially diverse cohort of maintenance dialysis patients in the Cleveland, OH, USA, metropolitan area. Materials and methods: In this cross-sectional study, we administered a 41-item questionnaire to 450 adult chronic dialysis patients. Items assessed patients’ knowledge of their kidney disease as well as their attitudes toward chronic kidney disease (CKD) treatment issues and EoLC issues. Results: The cohort included 67% Blacks, 27% Caucasians, 2.8% Hispanics, and 2.4% others. The response rate was 94% (423/450). Most patients considered it essential to obtain detailed information about their medical condition (80.6%) and prognosis (78.3%). Nearly 19% of respondents regretted their decision to start dialysis. 41% of patients would prefer treatment(s) aimed at relieving pain rather than prolonging life (30.5%), but a majority would want to be resuscitated (55.3%). Only 8.4% reported having a designated healthcare proxy, and 35.7% reported completing a living will. A significant percentage of patients wished to discuss their quality of life (71%), psychosocial and spiritual concerns (50.4%), and end-of-life issues (38%) with their nephrologist. Conclusion: Most dialysis patients wish to have more frequent discussions about their disease, prognosis, and EoLC planning. Findings from this study can inform the design of future interventions.

Hypokalemia (low serum potassium level) is a common electrolyte imbalance that can cause a defect in urinary concentrating ability, i.e., nephrogenic diabetes insipidus (NDI), but the molecular mechanism is unknown. We employed proteomic analysis of inner medullary collecting ducts (IMCD) from rats fed with a potassium-free diet for 1 day. IMCD protein quantification was performed by mass spectrometry using a label-free methodology. A total of 131 proteins, including the water channel AQP2, exhibited significant changes in abundance, most of which were decreased. Bioinformatic analysis revealed that many of the down-regulated proteins were associated with the biological processes of generation of precursor metabolites and energy, actin cytoskeleton organization, and cell-cell adhesion. Targeted LC-MS/MS and immunoblotting studies further confirmed the down regulation of 18 selected proteins. Electron microscopy showed autophagosomes/autophagolysosomes in the IMCD cells of rats deprived of potassium for only 1 day. An increased number of autophagosomes was also confirmed by immunofluorescence, demonstrating co-localization of LC3 and Lamp1 with AQP2 and several other down-regulated proteins in IMCD cells. AQP2 was also detected in autophagosomes in IMCD cells of potassium-deprived rats by immunogold electron microscopy. Thus, enhanced autophagic degradation of proteins, most notably including AQP2, is an early event in hypokalemia-induced NDI.