Unsupervised Clustering Algorithm Research Articles

GrFrauder: A Novel Unsupervised Clustering Algorithm for Identification Group Spam Reviewers

GrFrauder: A Novel Unsupervised Clustering Algorithm for Identification Group Spam Reviewers

Unveiling mitophagy-mediated molecular heterogeneity and development of a risk signature model for colorectal cancer by integrated scRNA-seq and bulk RNA-seq analysis.

Accumulating researchers have recognized mitophagy as a key player in tumors, but few studies have investigated its role in the tumor microenvironment (TME). Advances in the technology of single-cell RNA sequencing (scRNA-seq) have allowed unveiling the concealed features of the TME at cellular resolution. This study aimed to elucidate the role of mitophagy within the TME of colorectal cancer (CRC) and to establish a mitophagy-mediated risk model. We assessed mitophagy-related pathway activities at both single-cell and tissue levels. Subsequently, an unsupervised clustering algorithm was employed to identify mitophagy-mediated subtypes. Furthermore, we developed a mitophagy-mediated risk signature (MMRS) using least absolute shrinkage and selection operator (LASSO) Cox analysis and constructed a MMRS model incorporating the risk score and clinical variables. Subsequently, we used quantitative reverse transcription polymerase chain reaction analysis to verify the expression of the screened genes. We retrieved and annotated a total of 14,719 cells from eight samples in the scRNA-seq GSE132465 data set. The activities of mitophagy-related pathways were uniformly upregulated in cancer cells. Integrating with bulk RNA-seq data, we identified two mitophagy-mediated clusters (C1 and C2) with distinct characteristics and prognoses. C2 was identified as a mitophagy-high cluster. Then, we developed a five-gene MMRS via LASSO Cox analysis in The Cancer Genome Atlas (TCGA) cohort. We utilized the GSE39582 cohort to validate the efficacy of our model. The expression of CX3CL1 and INHBB was upregulated in CRC tissues. The present study identified two mitophagy-mediated CRC subtypes with distinct features. Our MMRS may provide potential therapeutic strategies for CRC. The findings of our work offer novel insights into the involvement of mitophagy in CRC.

A new thinking: deciphering the aberrance and clinical implication of copper-death signatures in clear cell renal cell carcinoma

RationaleRecent research has indicated that cuprotosis, or copper induced cell death, is a novel type of cell death that could be utilized as a new weapon for cancer management. However, the characteristics and implications of such signatures in cancers, especially in clear cell renal cell cancer (ccRCC), remain elusive.MethodsExpression, methylation, mutation, clinical information, copy number variation, functional implication, and drug sensitivity data at the pan-cancer level were collected from The Cancer Genome Atlas. An unsupervised clustering algorithm was applied to decipher ccRCC heterogeneity. Immune microenvironment construction, immune therapy response, metabolic pattern, and cancer progression signature between subgroups were also investigated.ResultsCuprotosis related genes were specifically downregulated in various cancer tissues compared with normal tissues and were correlated with hypermethylation and copy number variation. Cuprotosis scores were also dysregulated in tumor tissues, and we found that such a signature could positively regulate oxidative phosphorylation and Myc and negatively regulate epithelial mesenchymal translation and myogenesis pathways. CPCS1 (cuprotosis scores high) and CPCS2 (cuprotosis scores low) in ccRCC displayed distinctive clinical profiles and biological characteristics; the CPCS2 subtype had a higher clinical stage and a worse prognosis and might positively regulate cornification and epidermal cell differentiation to fuel cancer progression. CPCS2 also displayed a higher tumor mutation burden and low tumor stemness index, while it led to a low ICI therapy response and dysfunctional tumor immunity state. The genome-copy numbers of CPCS2, including arm- gain and arm- loss, were higher than those of CPCS1. The prognostic model constructed based on subgroup biomarkers exerted satisfactory performance in both the training and validation cohorts. In addition, overexpression of the copper death activator DLAT suppressed the malignant ability, including cell migration and proliferation, of renal cell lines in vitro and in vivo. Finally, activation of cuprotosis in tumors could enhance antitumor immunity through dsDNA-cGAS-STING signaling in ccRCC.ConclusionThe activation of cuprotosis might function as a promising approach among multiple cancers. The cuprotosis related signatures could reshape tumor immunity in the ccRCC microenvironment via cGAS-STING signal, thus activating tumor antigen-presenting process. Upregulation of DLAT expression in ccRCC cell lines could reactivate the copper death pattern and be treated as a suitable target for ccRCC.

Biophysical informatics reveals distinctive phenotypic signatures and functional diversity of single-cell lineages.

In this work, we present an analytical method for quantifying both single-cell morphologies and cell network topologies of tumor cell populations and use it to predict 3D cell behavior. We utilized a supervised deep learning approach to perform instance segmentation on label-free live cell images across a wide range of cell densities. We measured cell shape properties and characterized network topologies for 136 single-cell clones derived from the YUMM1.7 and YUMMER1.7 mouse melanoma cell lines. Using an unsupervised clustering algorithm, we identified six distinct morphological subclasses. We further observed differences in tumor growth and invasion dynamics across subclasses in an in vitro 3D spheroid model. Compared to existing methods for quantifying 2D or 3D phenotype, our analytical method requires less time, needs no specialized equipment and is capable of much higher throughput, making it ideal for applications such as high-throughput drug screening and clinical diagnosis. https://github.com/trevor-chan/Melanoma_NetworkMorphology. Supplementary data are available at Bioinformatics online.

Seasonal Patterns of Greenland Ice Velocity From Sentinel‐1 SAR Data Linked to Runoff

AbstractAccurate projections of the mass loss from the Greenland Ice Sheet (GrIS) require a complete understanding of the ice‐dynamic response to climate forcings on seasonal and interannual timescales and would greatly benefit from more observational evidence. Here, we analyze a 5‐year, high‐resolution data set of ice velocities of the GrIS using K‐means, an unsupervised clustering algorithm, to identify ice‐sheet wide characteristic seasonal flow patterns. We include all areas flowing >0.3 m/d and obtain an ice‐sheet wide overview of the seasonality and the interannual variability. It shows both a spatial and interannual variability in seasonal flow patterns, both along individual glaciers and between glaciers. We compare with runoff from a regional climate model and infer that the ice‐sheet wide patterns are linked to the availability of water penetrating to the base of the ice.

Deep Cascade AdaBoost with Unsupervised Clustering in Autonomous Vehicles

In recent years, deep learning has achieved excellent performance in a growing number of application fields. With the help of high computation and large-scale datasets, deep learning models with huge parameters constantly enhance the performance of traditional algorithms. Additionally, the AdaBoost algorithm, as one of the traditional machine learning algorithms, has a minimal model and performs well on small datasets. However, it is still challenging to select the optimal classification feature template from a large pool of features in any scene quickly and efficiently. Especially in the field of autonomous vehicles, images taken by onboard cameras contain all kinds of targets on the road, which means the images are full of multiple features. In this paper, we propose a novel Deep Cascade AdaBoost model, which effectively combines the unsupervised clustering algorithm based on deep learning and the traditional AdaBoost algorithm. First, we use the unsupervised clustering algorithm to classify the sample data automatically. We can obtain classification subsets with small intra-class and large inter-class errors by specifying positive and negative samples. Next, we design a training framework for Cascade-AdaBoost based on clustering and mathematically demonstrate that our framework has better detection performance than the traditional Cascade-AdaBoost framework. Finally, experiments on the KITTI dataset demonstrate that our model performs better than the traditional Cascade-AdaBoost algorithm in terms of accuracy and time. The detection time was shortened by 30%, and the false detection rate was reduced by 20%. Meanwhile, the training time of our model is significantly shorter than the traditional Cascade-AdaBoost algorithm.

Self-supervised image co-saliency detection

Self-supervised learning has been shown to be highly effective to avoid the expensive cost of collecting and annotating large-scale datasets. In this paper, we explore its use in image co-saliency detection task which detects and segments the co-concurrent and conspicuous patterns in one image group, and a self-supervised framework is presented that contains three novel module designs. Specifically, we first design an unsupervised graph clustering algorithm to generate an initial detection result from extracted semantical features of convolutional networks. Then, the post-processed output is taken as the pseudo labels of the second module that learns a neural network for more accurate sample affinity estimation, followed by a manifold ranking based label smoothing. Finally, with the smoothed labels as supervision, we design a saliency network that aggregates a hierarchy of multi-scale features to predict final optimized co-saliency maps. In addition, the proposed framework can be readily extended to image saliency object detection task when giving a single image as input. Extensive evaluations on three co-saliency and five saliency object detection benchmark datasets demonstrate that the proposed framework achieves favorable performance against a variety of supervised and unsupervised state-of-the-arts.

A 5-Pathway Signature Predicts Prognosis Based on Immune-Derived lncRNAs in Patients with Breast Cancer.

Currently, predictive models were not developed based on the signaling pathway signatures of immune-related lncRNAs in breast cancer (BRCA) patients. We selected unsupervised hierarchical clustering algorithm to classify patients with BRCA based on the significant immune-derived lncRNAs from the TCGA dataset. And different methods including ESTIMATE, ImmuneCellAI, and CIBERSORT were performed to evaluate the immune infiltration of tumor microenvironment. Using Lasso regression algorithm, we filtered the significant signaling pathways enriched by GSEA, GSVA, or PPI analysis to develop a prognostic model. And a nomogram integrated with clinical factors and significant pathways was constructed to predict the precise probability of overall survival (OS) of BRCA patients in the TCGA dataset (n = 1,098) and another two testing sets (n = 415). BRCA patients were stratified into the PC (n = 571) and GC (n = 527) subgroup with significantly different prognosis with 550 immune-related lncRNAs in the TCGA dataset. Integrated analysis revealed different immune response, oncogenic signaling, and metabolic reprograming pathways between these two subgroups. And a 5-pathway signature could predict the prognosis of BRCA patients between these two subgroups independently in the TCGA dataset, which was confirmed in another two cohorts from the GEO dataset. In the TCGA dataset, 5-year OS rate was 78% (95% CI: 73-84) vs. 82% (95% CI: 77-87) for the PC and GC group (HR = 1.63 (95% CI: 1.17-2.28), p = 0.004). The predictive power was similar in another two testing sets (HR > 1.20, p < 0.01). Finally, a nomogram is developed for clinical application, which integrated this signature and age to accurately predict the survival probability in BRCA patients. This 5-pathway signature correlated with immune-derived lncRNAs was able to precisely predict the prognosis for patients with BRCA and provided a rich source characterizing immune-related lncRNAs and further informed strategies to target BRCA vulnerabilities.

Development of criteria for cognitive dysfunction in post-COVID syndrome: the IC-CoDi-COVID approach

BackgroundWe aimed to develop objective criteria for cognitive dysfunction associated with the post-COVID syndrome. MethodsFour hundred and four patients with post-COVID syndrome from two centers were evaluated with comprehensive neuropsychological batteries. The International Classification for Cognitive Disorders in Epilepsy (IC-CoDE) framework was adapted and implemented. A healthy control group of 145 participants and a complementary data-driven approach based on unsupervised machine-learning clustering algorithms were also used to evaluate the optimal classification and cutoff points. ResultsAccording to the developed criteria, 41.2% and 17.3% of the sample were classified as having at least one cognitive domain impaired using -1 and -1.5 standard deviations as cutoff points. Attention/processing speed was the most frequently impaired domain. There were no differences in base rates of cognitive impairment between the two centers. Clustering analysis revealed two clusters, although with an important overlap (silhouette index 0.18-0.19). Cognitive impairment was associated with younger age and lower education levels, but not hospitalization. ConclusionsWe propose a harmonization of the criteria to define and classify cognitive impairment in the post-COVID syndrome. These criteria may be extrapolated to other neuropsychological batteries and settings, contributing to the diagnosis of cognitive deficits after COVID-19 and facilitating multicenter studies to guide biomarker investigation and therapies.

Single-Cell Sequencing of Malignant Ascites Reveals Transcriptomic Remodeling of the Tumor Microenvironment during the Progression of Epithelial Ovarian Cancer

Epithelial ovarian cancer (EOC) is the main cause of mortality among gynecological malignancies worldwide. Although patients with EOC undergo aggregate treatment, the prognosis is often poor. Peritoneal malignant ascites is a distinguishable clinical feature in EOC patients and plays a pivotal role in tumor progression and recurrence. The mechanisms of the tumor microenvironment (TME) in ascites in the regulation of tumor progression need to be explored. We comprehensively analyzed the transcriptomes of 4680 single cells from five EOC patients (three diagnostic samples and two recurrent samples) derived from Gene Expression Omnibus (GEO) databases. Batch effects between different samples were removed using an unsupervised deep embedding single-cell cluster algorithm. Subcluster analysis identified the different phenotypes of cells. The transition of a malignant cell state was confirmed using pseudotime analysis. The landscape of TME in malignant ascites was profiled during EOC progression. The transformation of epithelial cancer cells into mesenchymal cells was observed to lead to the emergence of related anti-chemotherapy and immune escape phenotypes. We found the activation of multiple biological pathways with the transition of tumor-associated macrophages and fibroblasts, and we identified the infiltration of CD4+CD25+ T regulatory cells in recurrent samples. The cell adhesion molecules mediated by integrin might be associated with the formation of the tumorsphere. Our study provides novel insights into the remodeling of the TME heterogeneity in malignant ascites during EOC progression, which provides evidence for identifying novel therapeutic targets and promotes the development of ovarian cancer treatment.

A novel thinking: DDR axis refines the classification of ccRCC with distinctive prognosis, multi omics landscape and management strategy.

DNA damage response and repair (DDR) related signatures play an important role in maintaining genome stability and other biological processes. It also affects the occurrence, development, and treatment of cancer. However, in renal cell carcinoma (RCC), especially clear cell renal carcinoma (ccRCC), the potential association between DDR-related signatures and tumor heterogeneity and tumor microenvironment (TME) remains unclear. Utilizing unsupervised clustering algorithm, we divided RCC into two subgroups, DCS1 and DCS2, according to the differences in DDR gene expression, and compared the characteristics of the two subgroups through multiple dimensions. Compared with DCS1, DCS2 patients have higher clinical stage/grade and worse prognosis, which may be related to active metabolic status and immunosuppression status. At the same time, the high mutation rate in DCS2 may also be an important reason for the prognosis. We also analyzed the sensitivity of the two subgroups to different therapeutic agents and established a subtypes' biomarkers-based prognostic system with good validation results to provide ideas for clinical diagnosis and treatment. Finally, we identified a pivotal role for DDX1 in the DDR gene set, which may serve as a future therapeutic target. This study showed that DDR has an important impact on the development and treatment of RCC. DCS2 subtypes have a poor prognosis, and more personalized treatment and follow-up programs may be needed. The assessment of DDR gene mutations in patients may be helpful for clinical decision-making. DDX1 may be one of the effective targets for RCC treatment in the future.

Computational Flow Cytometry Provides Accurate and Sensitive Measurable Residual Disease Assessment in Chronic Lymphocytic Leukemia

Measurable residual disease (MRD) is a highly sensitive reflection of disease burden that correlates with time-to-event outcomes in chronic lymphocytic leukemia (CLL). Undetectable MRD at the end of fixed-duration treatment is an independent prognostic marker and associated with favourable progression-free and overall survival. Muti-colour flow cytometry (FC) is the most common method of MRD assessment and often performed using a harmonized protocol established by the European Research Initiative on CLL. Conventional FC analysis involves sequential manual gating strategies performed on 2-dimensional scatter plots with varying axis parameters. While straightforward with a small number of parameters, this process becomes too complex and time-consuming with state-of-the-art mass flow cytometers where up to 40 markers are simultaneously measured and large multidimensional datasets are generated. Over the past decade, a variety of computational FC algorithms have been developed to automate and assist in tasks such as population identification, enumeration, and classification. These mathematical approaches are based on dimension reduction and clustering methods that require up to several hours and have not been compatible with routine diagnostic workflows until recently. Flow-Self Organizing Maps (FlowSOM) (Van Gassen et al. Cytometry A 2015) is an unsupervised clustering algorithm (R package from Bioconductor) with fast analysis times and multidimensional visualization. Implementation of FlowSOM in a diagnostic laboratory offers several advantages including lower technical variability, reduced bias, and improved efficiency. We developed a computational pipeline to assess CLL MRD using FlowSOM. CLL flow cytometry standard (FCS) files were identified from the Peter MacCallum Cancer Centre and included cases with typical and atypical immunophenotypes from peripheral blood (PB) or bone marrow (BM). Atypical CLL was defined as cases with different expression of one or more antigens from the typical immunophenotype. The diagnostic workflow was developed using a training set (n=25) containing representative samples from MRD negative (n=5), typical CLL (n=5), atypical CLL (n=10), and healthy bone marrow (n=5) cases. Computational accuracy was evaluated against expert human analysis in a validation cohort (n=293) consisting of PB typical (n=139), PB atypical (n=91), BM typical (n=30), and BM atypical CLL (n=33) samples. FCS files were preprocessed (PeacoQC, FlowCore R packages) to remove low-quality events, apply compensation, and transform the data prior to FlowSOM analysis. Two-level hierarchical clustering was performed to achieve maximum separation between normal and MRD events (level 1: lymphocytes, level 2: B cells). The training set was used to generate a diagnostic map of 225 clusters containing the full breadth of immature and mature B cells along with disease immunophenotypes (Figure 1). Clusters were verified by manual inspection in Kaluza software. Computational MRD was then determined by applying the FlowSOM map to validation samples and counting events within the disease clusters. Computational MRD demonstrated excellent correlation with expert human analysis. Linear regression R2 was 0.96, 0.79, 0.96, and 0.91 for PB typical, PB atypical, BM typical, and BM atypical CLL cohorts, respectively (Figure 2). Positive or negative MRD was highly accurate using a threshold of 0.01% for MRD. Computational analysis demonstrated high sensitivity (99%, 99%, 100%, 90%), specificity (92%, 92%, 100%, 85%), and concordance (95%, 98%, 100%, 88%) across all four validation groups (PB typical, PB atypical, BM typical and BM atypical CLL, respectively). The kappa coefficient was 0.92 across the entire validation cohort. CLL MRD assessment in BM samples can be difficult due to overlapping immunophenotypes particularly between disease and transitional B cells. We subsequently included an additional BM cohort consisting of only MRD negative samples to further assess detectable/undetectable accuracy. Samples were highly concordant (94%) except for two cases with low disease signal (0.010% and 0.012%). In summary, FlowSOM computational CLL MRD is highly accurate and offers the potential for an automated workflow. This method can be adapted to other haematological malignancies including acute leukemia and lymphoproliferative disorders. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Dynamic autonomic nervous system states arise during emotions and manifest in basal physiology.

The outflow of the autonomic nervous system (ANS) is continuous and dynamic, but its functional organization is not well understood. Whether ANS patterns accompany emotions, or arise in basal physiology, remain unsettled questions in the field. Here, we searched for brief ANS patterns amidst continuous, multichannel physiological recordings in 45 healthy older adults. Participants completed an emotional reactivity task in which they viewed video clips that elicited a target emotion (awe, sadness, amusement, disgust, or nurturant love); each video clip was preceded by a pre-trial baseline period and followed by a post-trial recovery period. Participants also sat quietly for a separate 2-min resting period to assess basal physiology. Using principal components analysis and unsupervised clustering algorithms to reduce the second-by-second physiological data during the emotional reactivity task, we uncovered five ANS states. Each ANS state was characterized by a unique constellation of patterned physiological changes that differentiated among the trials of the emotional reactivity task. These ANS states emerged and dissipated over time, with each instance lasting several seconds on average. ANS states with similar structures were also detectable in the resting period but were intermittent and of smaller magnitude. Our results offer new insights into the functional organization of the ANS. By assembling short-lived, patterned changes, the ANS is equipped to generate a wide range of physiological states that accompany emotions and that contribute to the architecture of basal physiology.

Self-Organized Fuzzy Neural Network Nonlinear System Modeling Method Based on Clustering Algorithm

In this paper, an improved self-organizing fuzzy neural network (SOFNN-CA) based on a clustering algorithm is proposed for nonlinear systems modeling in industrial processes. In order to reduce training time and increase training speed, we combine offline learning and online identification. The unsupervised clustering algorithm is used to generate the initial centers of the network in the offline learning phase, and, in the self-organizing phase of the system, the Mahalanobis distance (MD) index and error criterion are adopted to add neurons to learn new features. A new density potential index (DPI) combined with neuron local field potential (LFP) is designed to adjust the neuron width, which further improves the network generalization. The similarity index calculated by the Gaussian error function is used to merge neurons to reduce redundancy. Meanwhile, the convergence of SOFNN-CA in the case of structural self-organization is demonstrated. Simulations and experiments results show that the proposed SOFNN-CA has a more desirable modeling accuracy and convergence speed compared with SOFNN-ALA and SOFNN-AGA.

Comparing simulated Milky Way satellite galaxies with observations using unsupervised clustering

ABSTRACT We develop a new analysis method that allows us to compare multidimensional observables to a theoretical model. The method is based on unsupervised clustering algorithms which assign the observational and simulated data to clusters in high dimensionality. From the clustering result, a goodness of fit (the p-value) is determined with the Fisher–Freeman–Halton test. We first show that this approach is robust for 2D Gaussian distributions. We then apply the method to the observed MW satellites and simulated satellites from the fiducial model of our semi-analytic code a-sloth . We use the following five observables of the galaxies in the analysis: stellar mass, virial mass, heliocentric distance, mean stellar metallicity [Fe/H], and stellar metallicity dispersion σ[Fe/H]. A low p-value returned from the analysis tells us that our a-sloth fiducial model does not reproduce the mean stellar metallicity of the observed MW satellites well. We implement an ad hoc improvement to the physical model and show that the number of dark matter merger trees which have a p-value &amp;gt; 0.01 increases from 3 to 6. This method can be extended to data with higher dimensionality easily. We plan to further improve the physical model in a-sloth using this method to study elemental abundances of stars in the observed MW satellites.

Multispectral intraoperative imaging for the detection of the hemodynamic response to interictal epileptiform discharges.

Interictal epileptiform discharges (IEDs) are brief neuronal discharges occurring between seizures in patients with epilepsy. The characterization of the hemodynamic response function (HRF) specific to IEDs could increase the accuracy of other functional imaging techniques to localize epileptiform activity, including functional near-infrared spectroscopy and functional magnetic resonance imaging. This study evaluated the possibility of using an intraoperative multispectral imaging system combined with electrocorticography (ECoG) to measure the average HRF associated with IEDs in eight patients. Inter-patient variability of the HRF is illustrated in terms of oxygenated hemoglobin peak latency, oxygenated hemoglobin increase/decrease following IEDs, and signal-to-noise ratio. A sub-region was identified using an unsupervised clustering algorithm in three patients that corresponded to the most active area identified by ECoG.

ODP028 Association Between Serum Steroid Profile and Metabolic Risks in Adults with Classic Congenital Adrenal Hyperplasia Due to 21-Hydroxylase Deficiency

Abstract Objective In adult patients with congenital adrenal hyperplasia (CAH), long-term health outcomes, including metabolic and cardiovascular health, are a priority for the management of CAH. The enzyme 21-hydroxylase deficiency, which accounts for a majority of CAH, leads to a large perturbation in serum steroid profile under glucocorticoid replacement therapy. Serum levels of several adrenal steroids, including 17-hydroxyprogesterone and androstenedione, are associated with disease severity and sexual dysfunction of CAH patients. However, the association between the serum steroid profiles with metabolic risk in CAH patients is not elucidated yet. We aimed to investigate the serum steroid profile of adult CAH patients using liquid chromatography-mass spectrometry (LC-MS) and their association with metabolic risk. Methods Adult patients with classic CAH due to 21-hydroxylase deficiency were included in this study (36 females and 27 males). A multi-steroid panel composed of 25 steroids and their metabolites was applied to morning serum samples of study subjects using LC-MS. An unsupervised clustering algorithm was applied to the serum steroid panel to discover any unique pattern which divides CAH subjects into clusters. The association between serum steroid profiles and clinical characteristics, including age, sex, body mass index (BMI), subtypes (simple virilizing and salt wasting), glucocorticoid dose, and the presence of hypertension, diabetes mellitus (DM), and metabolic syndrome (MetS) was analyzed. The discriminatory power of each steroid or a combination of steroids was estimated using the area under the curve of receiver operating characteristics (AUROC) for metabolic diseases. Results Fifteen steroids including pregnenolone sulfate, 17α-hydroxyprogesterone, cortisol, cortisone, 21-deoxycortisol, 20α-dihydrocortisol, 20α-dihydrocortisone, α-cortolone, β-cortolone, tetrahydrocortisone, dehydroepiandrosterone sulfate, testosterone, androstenedione, 11β-hydroxytestosterone, and 11β-hydroxyandrostenedione were successfully quantified in all subjects. The median age of subjects was 28 years (interquartile range: 25-35). The prevalence of DM, hypertension, and MetS was 3.2% (2/63), 15.8% (10/63), and 28.6% (18/63), respectively. The subjects were divided into two clusters based on the pattern of serum steroid profile using unsupervised hierarchical clustering. The prevalence of MetS was significantly different between the two clusters (cluster 1, 37.8% [17/45] vs. cluster 2, 5.6% [1/18], P = 0. 011). The prevalence of HTN was numerically higher in cluster 1 than cluster 2 (20. 0% [9/45] vs. 5.6% [1/18], P = 0.257). Other clinical characteristics, including age, sex, BMI, subtypes, and glucocorticoid dose, were not different between the two clusters. Among fifteen steroids, the level of tetrahydrocortisone showed the highest discriminatory power for MetS (AUROC 0.795, 95% confidence interval: 0.675-0.914). The multivariate logistic regression model of all 15 steroids showed AUROC of 0.832 (95% CI: 0.732-0.933) for MetS. Conclusion The serum steroid profiles of CAH patients were significantly associated with the presence of MetS. This suggests that serum steroid signatures can guide the optimal management of adult CAH patients to minimize the risk of MetS. Presentation: No date and time listed

COMBING: Clustering in Oncology for Mathematical and Biological Identification of Novel Gene Signatures.

Precision medicine is a paradigm shift in healthcare relying heavily on genomics data. However, the complexity of biological interactions, the large number of genes as well as the lack of comparisons on the analysis of data, remain a tremendous bottleneck regarding clinical adoption. In this paper, we introduce a novel, automatic and unsupervised framework to discover low-dimensional gene biomarkers. Our method is based on the LP-Stability algorithm, a high dimensional center-based unsupervised clustering algorithm. It offers modularity as concerns metric functions and scalability, while being able to automatically determine the best number of clusters. Our evaluation includes both mathematical and biological criteria to define a quantitative metric. The recovered signature is applied to a variety of biological tasks, including screening of biological pathways and functions, and characterization relevance on tumor types and subtypes. Quantitative comparisons among different distance metrics, commonly used clustering methods and a referential gene signature used in the literature, confirm state of the art performance of our approach. In particular, our signature, based on 27 genes, reports at least 30 times better mathematical significance (average Dunn's Index) and 25% better biological significance (average Enrichment in Protein-Protein Interaction) than those produced by other referential clustering methods. Finally, our signature reports promising results on distinguishing immune inflammatory and immune desert tumors, while reporting a high balanced accuracy of 92% on tumor types classification and averaged balanced accuracy of 68% on tumor subtypes classification, which represents, respectively 7% and 9% higher performance compared to the referential signature.

Unsupervised clustering algorithms improve the reproducibility of dynamic contrast-enhanced magnetic resonance imaging pulmonary perfusion quantification in muco-obstructive lung diseases.

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) allows the assessment of pulmonary perfusion, which may play a key role in the development of muco-obstructive lung disease. One problem with quantifying pulmonary perfusion is the high variability of metrics. Quantifying the extent of abnormalities using unsupervised clustering algorithms in residue function maps leads to intrinsic normalization and could reduce variability. We investigated the reproducibility of perfusion defects in percent (QDP) in clinically stable patients with cystic fibrosis (CF) and chronic obstructive pulmonary disease (COPD). 15 CF (29.3 ± 9.3y, FEV1%predicted = 66.6 ± 15.8%) and 20 COPD (66.5 ± 8.9y, FEV1%predicted = 42.0 ± 13.3%) patients underwent DCE-MRI twice 1 month apart. QDP, pulmonary blood flow (PBF), and pulmonary blood volume (PBV) were computed from residue function maps using an in-house quantification pipeline. A previously validated MRI perfusion score was visually assessed by an expert reader. Overall, mean QDP, PBF, and PBV did not change within 1 month, except for QDP in COPD (p < 0.05). We observed smaller limits of agreement (± 1.96 SD) related to the median for QDP (CF: ± 38%, COPD: ± 37%) compared to PBF (CF: ± 89%, COPD: ± 55%) and PBV (CF: ± 55%, COPD: ± 51%). QDP correlated moderately with the MRI perfusion score in CF (r = 0.46, p < 0.05) and COPD (r = 0.66, p < 0.001). PBF and PBV correlated poorly with the MRI perfusion score in CF (r =-0.29, p = 0.132 and r =-0.35, p = 0.067, respectively) and moderately in COPD (r =-0.57 and r =-0.57, p < 0.001, respectively). In patients with muco-obstructive lung diseases, QDP was more robust and showed a higher correlation with the MRI perfusion score compared to the traditionally used perfusion metrics PBF and PBV.

Structure Identification for Force-Induced Reaction Using Single-Molecule Conductance Measurement

Structure Identification for Force-Induced Reaction Using Single-Molecule Conductance Measurement