Unsupervised Clustering Method Research Articles

Integrated single-cell and bulk RNA-sequencing data reveal molecular subtypes based on lactylation-related genes and prognosis and therapeutic response in glioma

ObjectivesGlioma, the most common and aggressive form of brain cancer, possesses a complex biology, which makes elucidating its underlying mechanisms and developing effective treatment strategies challenging. Lactylation is a recently discovered post-translational modification and has emerged as a novel research target to understand its role in various biological processes and diseases. Herein, we explored the role of lactylation in gliomas. MethodsSingle-cell RNA-sequencing (scRNA-seq) data were downloaded from the Tumour Immune Single-Cell Hub database. The R package ‘Seurat’ was used for processing the scRNA-seq data. Lactylation-related genes were identified from published literature and the Molecular Signatures Database. An unsupervised clustering method was used to identify glioma subtypes based on identified lactylation-related genes. Differences among the various clusters were examined, including clinical features, differentially expressed genes (DEGs), enriched pathways and immune cell infiltrates. A lactylation score was generated to predict the overall survival (OS) of patients with glioma using DEGs between the two clusters. ResultsThe lactylation-related genes were obtained from the scRNA-seq data, identifying two molecular subtypes, and a prognostic signature was established to stratify patients with glioma into high- and low-score groups. Analysis of the tumour immune microenvironment revealed that patients in the high-score group exhibited increased immune cell infiltration, chemokine expression and immune checkpoint expression but exhibited worse OS and better response to immunotherapy. ConclusionsAltogether, we established a novel signature based on lactylation-related clusters for robust survival prediction and immunotherapeutic response in gliomas.

Biological characteristics of molecular subtypes of ulcerative colitis characterized by ferroptosis and neutrophil infiltration

Clinical ulcerative colitis (UC) is a heterogeneous condition. Moreover, medical interventions are nonspecific, and thus, treatment responses are inconsistent. The aim of this study was to explore the molecular subtypes and biological characteristics of UC based on ferroptosis and neutrophil gene sets. Multiple intestinal mucosa gene expression profiles of UC patients in the Gene Expression Omnibus (GEO) database were downloaded. Unsupervised clustering methods were used to identify potential molecular subtypes based on ferroptosis and neutrophil gene sets. Multiple immune infiltration algorithms were used to evaluate the biological characteristics of the molecular subtypes. Machine learning identifies hub genes for molecular subtypes and analyses their diagnostic efficacy for UC and predictive performance for drug therapy. The relevant conclusions were verified by clinical samples and animal experiments. Four molecular subtypes were identified according to the ferroptosis and neutrophil gene sets: neutrophil, ferroptosis, mixed and quiescent. The subtypes have different biological characteristics and immune infiltration levels. Multiple machine learning methods jointly identified four hub genes (FTH1, AQP9, STEAP3 and STEAP4). Receiver operating characteristic (ROC) curve analysis revealed that the four hub genes could be used as diagnostic markers for UC. The clinical response profile data of infliximab treatment patients showed that AQP9 and STEPA4 were reliable predictors of infliximab treatment response. In human samples the AQP9 and STEAP4 protein were shown to be increased in UC intestinal samples. In animal experiments, the ferroptosis and neutrophil phenotype were confirmed. Dual analysis of ferroptosis and neutrophil gene expression revealed four subgroups of UC patients. The molecular subtype-associated hub genes can be used as diagnostic markers for UC and predict infliximab treatment response.

Auroral breakup detection in all-sky images by unsupervised learning

Abstract. Due to a large number of automatic auroral camera systems on the ground, image data analysis requires more efficiency than what human expert visual inspection can provide. Furthermore, there is no solid consensus on how many different types or shapes exist in auroral displays. We report the first attempt to classify auroral morphological forms by an unsupervised learning method on an image set that contains both nightside and dayside aurora. We used 6 months of full-colour auroral all-sky images captured at a high-Arctic observatory on Svalbard, Norway, in 2019–2020. The selection of images containing aurora was performed manually. These images were then input into a convolutional neural network called SimCLR for feature extraction. The clustered and fused features resulted in 37 auroral morphological clusters. In the clustering of auroral image data with two different time resolutions, we found that the occurrence of 8 clusters strongly increased when the image cadence was high (24 s), while the occurrence of 14 clusters experienced little or no change with changes in input image cadence. We therefore investigated the temporal evolution of a group of eight “active aurora” clusters. Time periods for which this active aurora persisted for longer than two consecutive images with a maximum cadence of 6 min coincided with ground-magnetic deflections, and their occurrence was found to maximize around magnetic midnight. The active aurora onsets typically included vortical auroral structures and equivalent current patterns typical for substorms. Our findings therefore suggest that our unsupervised image clustering method can be used to detect auroral breakups in ground-based image datasets with a temporal accuracy determined by the image cadence.

A new commercial boundary dataset for metropolitan areas in the USA and Canada, built from open data

The purpose of this study is to define the geographic boundaries of commercial areas by creating a consistent definition, combining various commercial area types, including downtowns, retail centres, financial districts, and other employment subcentres. Our research involved the collection of office, retail and job density data from 69 metropolitan regions across USA and Canada. Using this data, we conducted an unsupervised image segmentation model and clustering methods to identify distinctive commercial geographic boundaries. As a result, we identified 23,751 commercial areas, providing a detailed perspective on the commercial landscape of metropolitan areas in the USA and Canada. In addition, the generated boundaries were successfully validated through comparison with previously established commerce-related boundaries. The output of this study has implications for urban and regional planning and economic development, delivering valuable insights into the overall commercial geography in the region. The commercial boundary and used codes are freely available on the School of Cities Github, and users can reuse, reproduce and modify the boundaries.

Early life adversity is associated with differential gene expression in immune cells: A cluster-based analysis across an acute psychosocial stressor

Elucidating mechanisms by which early-life adversity (ELA) contributes to increased disease risk is important for mitigating adverse health outcomes. Prior work has found differences in immune cell gene expression related to inflammation and mitochondrial activity. Using a within-person between-group experimental design, we investigated differences in gene expression clusters across acute psychosocial stress and no-stress conditions. Participants were young adults (N = 29, aged 18 – 25 years, 62 % female, 47 % with a history of ELA). Gene expression was assessed in peripheral blood mononuclear cells collected at 8 blood draws spanning two 5-hour sessions (stress vs. no-stress) separated by a week, 4 across each session (number of observations = 221). We applied two unsupervised gene clustering methods – latent profile analysis (LPA) and weighted gene co-expression analysis (WGCNA) – to cluster genes with similar expression patterns across participants. LPA identified 11 clusters, 7 of which were significantly associated with ELA-status. WGCNA identified 5 clusters, 3 of which were significantly associated with ELA-status. LPA- and WGCNA-identified clusters were correlated, and all clusters were highly preserved across sessions and time. There was no significant effect of acute stress on cluster gene expression, but there was a significant effect of time, and significant differences by ELA-status. ELA-associated clusters related to RNA splicing/processing, inflammation, leukocyte differentiation and division, and mitochondrial activity were differentially expressed across time: ELA-exposed individuals showed decreased expression of these clusters at 90-minutes while controls showed increased expression. Our findings replicate previous work in this area and highlight additional mechanisms by which ELA may contribute to disease risk.

Association of Integrated Proteomic and Metabolomic Modules with Risk of Kidney Disease Progression.

Proteins and metabolites play crucial roles in various biological functions and are frequently interconnected through enzymatic or transport processes. We present an integrated analysis of 4,091 proteins and 630 metabolites in the Chronic Renal Insufficiency Cohort Study (N=1,708; average follow-up for kidney failure [KF], 9.5 years, with 537 events). Proteins and metabolites were integrated using an unsupervised clustering method and we assessed associations between clusters and CKD progression and kidney failure using Cox proportional hazards models. Analyses were adjusted for demographics and risk factors including the estimated glomerular filtration rate (eGFR) and urine protein-creatinine ratio. Associations were identified in a discovery sample (random two-thirds, N=1139) and then evaluated in a replication sample (one-third, N=569). We identified 139 modules of correlated proteins and metabolites, which were represented by their principal components (PC). Modules and PC loadings were projected onto the replication sample which demonstrated a consistent network structure. Two modules, representing a total of 236 proteins and 82 metabolites, were robustly associated with both CKD progression and kidney failure in both discovery and validation samples. Using gene set enrichment, several transmembrane related terms were identified as over-represented in these modules. Transmembrane-ephrin receptor activity displayed the largest odds (OR = 13.2, P-value = 5.5×10 -5 ). A module containing CRIM1 and NPNT expressed in podocytes demonstrated particularly strong associations with kidney failure (P-value = 2.6×10 -5 ). This study demonstrates that integration of the proteome and metabolome can identify functions of pathophysiologic importance in kidney disease.

Necroptosis-related lncRNA-based novel signature to predict the prognosis and immune landscape in soft tissue sarcomas

BackgroundNecroptosis-related long noncoding RNAs (lncRNAs) play crucial roles in cancer initiation and progression. Nevertheless, the role and mechanism of necroptosis-related lncRNAs in soft tissue sarcomas (STS) is so far unknown and needs to be explored further.MethodsClinical and genomic data were obtained from the UCSC Xena database. All STS patients’ subclusters were performed by unsupervised consensus clustering method based on the prognosis-specific lncRNAs, and then assessed their survival advantage and immune infiltrates. In addition, we explored the pathways and biological processes in subclusters through gene set enrichment analysis. At last, we established the necroptosis-related lncRNA-based risk signature (NRLncSig) using the least absolute shrinkage and selection operator (LASSO) method, and explored the prediction performance and immune microenvironment of this signature in STS.ResultsA total of 911 normal soft tissue samples and 259 STS patients were included in current study. 39 prognosis-specific necroptosis-related lncRNAs were selected. Cluster 2 had a worse survival than the cluster 1 and characterized by different immune landscape in STS. A worse outcome in the high-risk group was observed by survival analysis and indicated an immunosuppressive microenvironment. The ROC curve analyses illustrated that the NRLncSig performing competitively in prediction of prognosis for STS patients. In addition, the nomogram presents excellent performance in predicting prognosis, which may be more beneficial towards STS patients’ treatment.ConclusionsOur result indicated that the NRLncSig could be a good independent predictor of prognosis, and significantly connected with immune microenvironment, thereby providing new insights into the roles of necroptosis-related lncRNAs in STS.

Inter and intra‐operator reliability of Lekholm and Zarb classification and proposal of a novel radiomic data‐driven clustering for qualitative assessment of edentulous alveolar ridges

AbstractObjectivesThe present study was conducted to evaluate the reproducibility of Lekholm and Zarb classification system (L&Z) for bone quality assessment of edentulous alveolar ridges and to investigate the potential of a data‐driven approach for bone quality classification.Materials and MethodsTwenty‐six expert clinicians were asked to classify 110 CBCT cross‐sections according to L&Z classification (T0). The same evaluation was repeated after one month with the images put in a different order (T1). Intra‐ and inter‐examiner agreement analyses were performed using Cohen's kappa coefficient (CK) and Fleiss' kappa coefficient (FK), respectively. Additionally, radiomic features extraction was performed from 3D edentulous ridge blocks derived from the same 110 CBCTs, and unsupervised clustering using 3 different clustering methods was used to identify patterns in the obtained data.ResultsIntra‐examiner agreement between T0 and T1 was weak (CK 0.515). Inter‐examiner agreement at both time points was minimal (FK at T0: 0.273; FK at T1: 0.243). The three different unsupervised clustering methods based on radiomic features aggregated the 110 CBCTs in three groups in the same way.ConclusionsThe results showed low agreement among clinicians when using L&Z classification, indicating that the system may not be as reliable as previously thought. The present study suggests the possible application of a reproducible data‐driven approach based on radiomics for the classification of edentulous alveolar ridges, with potential implications for improving clinical outcomes. Further research is needed to determine the clinical significance of these findings and to develop more standardized and accurate methods for assessing bone quality of edentulous alveolar ridges.

An integrated geomechanical and petrophysical multiparameter approach for gas reservoir evaluation

AbstractIntegrating petrophysical and geomechanical parameters is an efficient approach to evaluating shale gas reservoir potential. The high cost of corings and their limited number, coupled with time-intensive investigation, led researchers to use this alternative combination approach. In the Jiaoshiba area, from single-pilot well core data and log measurements, petrophysical and geomechanical parameters such as shale volume, total organic carbon, gas content, as well as pore pressure, stress components, and mineral brittleness were first estimated using established methods. In the second phase, based on logging curves, the reservoir electro-facies (EF) classification was performed using the unsupervised multi-resolution graph-based clustering method on a series of twenty wells, identifying five EF with different intrinsic characteristics. Unsupervised analyses were developed using the multilayer artificial neural network while incorporating the K-nearest neighbors and graphical classification algorithms. The results from the first and second phases indicate reservoir richness in organic matter, with the best reservoir exhibited by EF2 and EF3. In addition, effective stress components (SV, SH, and Sh) evaluation shows a normal stress regime with hydraulic fracture systems perpendicular to the minimum horizontal stress at each measured depth of the reservoir (Sv > SH > Sh). This research workflow can efficiently evaluate shale reservoirs with a realistic approach for identifying favorable fracturing positions while reducing errors due to human interference.

Hybrid methodology development for lubrication regimes identification based on measurements, simulation, and data clustering

The lubrication regimes of a laboratory scale journal bearing were analysed with wide band acoustic emission (AE) measurements. The analysis was supported by data-based clustering of AE data. Digital twin of the journal bearing was generated with Simpack multi-body simulation software to study opportunities for developing a hybrid methodology for more complex systems monitoring. The AE and data-based clustering approach can be effectively used to reveal fundamental lubrication modes, i.e., hydrodynamic (HL), mixed (ML) and boundary (BL) lubrication as a function of Hersey number, which can be evaluated in situ by utilizing digital twin. Besides AE the other parameters monitored were friction torque, bearing temperature, loading, sliding velocity and oil pressure. The materials used in the experiments were case-hardened 18CrNiMo7–6 steel and nitrided 42CrMo7 steel. The tests were lubricated with synthetic extreme-pressure gear oil (SGN 320) and the bearing temperature was kept constant during the tests. The bearing loading and sliding velocity during tests were varied in the wide range resulting in different lubrication situations. The acoustic emission signals power and frequency content was analysed, and essential features were extracted for data clustering. For lubrication regime change identification the parameters such as signal RMS and coefficient of variation (CV) proved to be important, while signal kurtosis showed to be the most sensitive in discovering anomalies. The high sensitivity requires data filtering to remove erroneous peaks and to reveal real trends more clearly. It is also interesting to notice the changes in AE frequency during changing to different lubrication regime. In literature different clustering and classification methods has been proposed and applied for journal bearing status identification. Here the selected unsupervised clustering method was the mean-shift clustering due to fact, that the lubrication regimes in the Stribeck curve form an inseparable continuum. The algorithm does not require specifying the number of clusters in advance, i.e., the clusters are determined by the algorithm with respect to the data. The results were compared to simulations with digital twin. i.e., by comparing simulated digital twin film thickness and measured AE kurtosis relative to measured Hersey number. It was concluded that digital twins can be utilized as virtual sensor for in situ detecting of lubrication regimes, if it is possible to calibrate the simulation with sensitive measurements, e.g., AE. It is obvious that simulations alone cannot predict suddenly appearing anomalies, such as impurities or surface fatigue failures.

Abstract 2310: Integrating real-world histopathological and clinicogenomic data from 1799 lung cancer patients by applying unsupervised deep learning

Abstract Hematoxylin and eosin (H&E) stained histologic sections contain invaluable information that remains largely untapped because of its complexity. To this end, AI applications employing deep learning (DL) can facilitate the translation of image data to enable human interpretation and yield novel oncological insights that would have otherwise remained imperceptible. DL-based methodologies are multimodal, capable of integrating imaging with clinicogenomic data to furnish a more holistic perspective and affording more accurate predictions in oncology. Here, we developed an unsupervised DL workflow to analyze 1,799 H&E images of lung cancer (NSCLC n = 951; SCLC n = 50; others n = 798) incorporating comprehensive patient-level clinical data (electronic health records [ConcertAI]) integrated with genomics (WES and RNA-seq [Caris Labs]). There are three steps in our approach: (1) image preprocessing and filtering, yielding > 30 million image patches; (2) utilizing pretrained SimCLR models from 57 public oncology histopathology datasets with ResNet-18 as a backbone structure to extract 512-dimensional-feature vectors for each patch; (3) using three unsupervised clustering methods (kmeans, DBSCAN, Leiden clustering) to cluster patches and selected Leiden clustering. We identified 635 primary imaging clusters using an elbow method and generated an image feature matrix by calculating correlations between each patch and cluster centroids; these were aggregated and mapped back to source slides. In this proof-of-concept, distinct image feature patterns characterized SCLC and NSCLC samples. For SCLC, one of the salient features was the presence of hemorrhage, which may be associated with higher rates of fine-needle aspiration biopsy procedure for SCLC compared with NSCLC which was confirmed in the EHR data (p = 0.032). Derived morphological clusters were correlated with tumor-immune genomic features (Tumor Mutational Burden [TMB], Immunologic Constant of Rejection [ICR], and Miracle scores1) serving as predictors of response to immune-checkpoint inhibitor therapy. By applying linear models, we detected 11, 96 and 249 significantly associated imaging clusters, respectively, highly enriched with immune cells e.g., plasma cells, macrophages, lymphocytes, and supporting an infiltrated and inflamed tumor-immune microenvironment. In summary, a multimodal, unsupervised deep learning workflow combining H&E imaging with clinicogenomic data was developed to identify histologic feature clusters associated with well-established tumor-immune genomic signatures of NSCLC immune infiltration and molecular phenotypes. These studies demonstrate enormous potential to yield histopathological and translational insights in NSCLC and SCLC that can empower clinicians to make better therapeutic response predictions. Citation Format: Si Wu, Yujie Zhao, Hugo Luo, Kevin Kolahi, Thanh Bui, Xu Shi, Aditee Shrotre, Alexander Liede, Xi Zhao, Josue Samayoa, Weilong Zhao. Integrating real-world histopathological and clinicogenomic data from 1799 lung cancer patients by applying unsupervised deep learning [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2310.

Abstract 7417: SpatialCS: A tool to effectively integrate histological annotations and spatial transcriptomics analysis

Abstract Spatial transcriptomics is revolutionizing our understanding of tumor ecosystems, providing insights into the spatial dimensions of cell locations, their organization, and interactions within tissues. A key challenge in this field is integrating histological annotations with spatial domains identified through unsupervised clustering methods like SpaGCN. This integration is crucial for accurate spatial domain identification and further analyses. However, clustering results often vary based on parameters such as resolution, leading to inconsistencies with histological annotations. This variation necessitates optimized parameter selection to achieve the best clustering outcomes. We here introduce SpatialCS, a novel tool specifically designed to both visualize and quantify the alignment between histological annotations and the clustering results obtained from spatial transcriptomics data. SpatialCS employs a new index based on the asymmetric Wallace indices. This index penalizes spot pairs that are clustered together but do not share the same histology annotation type, with penalties scaled according to their transcriptional similarity. Additionally, we propose a purity index, derived by calculating the average transcriptional entropy of all clusters. This index serves as a measure of the homogeneity within each cluster. Furthermore, SpatialCS includes a unique visualization function. This function projects histological annotations onto the clustering results as outlines, making it easier to assess the consistency between histology annotations and clustering outcomes. We have applied SpatialCS to Visium spatial transcriptomics data generated on four colorectal cancer liver metastases, each accompanied by comprehensive spot-level histological annotations in both tumor and normal compartments. The histologic annotations include but not limited to are tumor cells, tumor cell cluster subtypes, tumor stroma, tumor in the vessels, portal tract, individual bile ducts, portal vein and liver cell plates and non-neoplastic stroma. The application of SpatialCS has proven its effectiveness in identifying the most consistent clustering results, which were further confirmed by experienced pathologists. Notably, the output of SpatialCS offers helpful guidance for improving histological annotations. Together, SpatialCS not only provides a quantitative approach to evaluate the consistency between histological annotations and unsupervised clustering analysis but also showcases its potential to significantly enhance histological annotations. Citation Format: Yanshuo Chu, Kyung Serk Cho, Isha Khanduri, Tieling Zhou, Jiahui Jiang, Yunhe Liu, Xinmiao Yan, Guangsheng Pei, Yibo Dai, Yang Liu, Ruiping Wang, Akshaya S Jadhav, Sharia Hernandez, Humam Kadara, Luisa Maren Solis Soto, Scott Kopetz, Dipen Maheshbhai Maru, Linghua Wang. SpatialCS: A tool to effectively integrate histological annotations and spatial transcriptomics analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7417.

Unsupervised Clustering-Assisted Method for Consensual Quantitative Analysis of Methanol-Gasoline Blends by Raman Spectroscopy.

Methanol-gasoline blends have emerged as a promising and environmentally friendly bio-fuel option, garnering widespread attention and promotion globally. The methanol content within these blends significantly influences their quality and combustion performance. This study explores the qualitative and qualitative analysis of methanol-gasoline blends using Raman spectroscopy coupled with machine learning methods. Experimentally, methanol-gasoline blends with varying methanol concentrations were artificially configured, commencing with initial market samples. For qualitative analysis, the partial least squares discriminant analysis (PLS-DA) model was employed to classify the categories of blends, demonstrating high prediction performance with an accuracy of nearly 100% classification. For the quantitative analysis, a consensus model was proposed to accurately predict the methanol content. It integrates member models developed on clustered variables, using the unsupervised clustering method of the self-organizing mapping neural network (SOM) to accomplish the regression prediction. The performance of this consensus model was systemically compared to that of the PLS model and uninformative variable elimination (UVE)-PLS model. Results revealed that the unsupervised consensus model outperformed other models in predicting the methanol content across various types of methanol gasoline blends. The correlation coefficients for prediction sets consistently exceeded 0.98. Consequently, Raman spectroscopy emerges as a suitable choice for both qualitative and quantitative analysis of methanol-gasoline blend quality. This study anticipates an increasing role for Raman spectroscopy in analysis of fuel composition.

Association of TNFRSF19 with a TNF family-based prognostic model and subtypes in gliomas using machine learning

PurposeTNF family members (TFMs) play a crucial role in different types of cancers, with TNF Receptor Superfamily Member 19 (TNFRSF19) standing out as a particularly important member in this category. Further research is necessary to investigate the potential impact of TFMs on prognosis prediction and to elucidate the function and potential therapeutic targets linked to TNFRSF19 expression in gliomas. MethodsThree databases provided the data on gene expression and clinical information. Fourteen prognostic members were found through univariate Cox analysis and subsequently utilized to construct TFMs-based model in LASSO and multivariate Cox analyses. TFMs-based subtypes based on the expression profile were identified using an unsupervised clustering method. Machine learning algorithm identified key genes linked to prognostic model and subtype. A sequence of immune infiltrations was evaluated using the ssGSEA and ESTIMATE algorithms. Immunohistochemistry was used to examine the patterns of expression and the clinical significance of TNFRSF19. ResultsOur development of a prognostic model and subtypes based on the TNF family was successful, resulting in accurate predictions of prognosis. The findings indicate that TNFRSF19 exhibited strong performance. Upregulation of TNFRSF19 was correlated with malignant phenotypes and poor prognosis, which was confirmed through immunohistochemistry. TNFRSF19 played a role in reshaping the immunosuppressive microenvironment in gliomas, and multiple drug-targeted TNFRSF19 molecules were identified. ConclusionsThe TMF-based prognostic model and subtype can facilitate treatment decisions for glioma. TNFRSF19 is an outstanding representative of a predictor of prognosis and immunotherapy effect in gliomas.

Deep clustering in subglacial radar reflectance reveals subglacial lakes

Abstract. Ice-penetrating radar (IPR) imaging is a valuable tool for observing the internal structure and bottom of ice sheets. Subglacial water bodies, also known as subglacial lakes, generally appear as distinct, bright, flat, and continuous reflections in IPR images. In this study, we use available IPR images from the Gamburtsev Subglacial Mountains to extract one-dimensional reflector waveform features of the ice–bedrock interface. We apply a deep-learning method to reduce the dimension of the reflector features. An unsupervised clustering method is then used to separate different types of reflector features, including a reflector type corresponding to subglacial lakes. The derived clustering labels are then used to detect features of subglacial lakes in IPR images. Using this method, we compare the new detections with a known-lakes inventory. The results indicate that this new method identified additional subglacial lakes that were not previously detected, and some previously known lakes are found to correspond to other reflector clusters. This method can offer automatic detections of subglacial lakes and provide new insight for subglacial studies.

FVCNet: Detection obstacle method based on feature visual clustering network in power line inspection

AbstractPower line inspection is an important means to eliminate hidden dangers of power lines. It is a difficult research problem how to solve the low accuracy of power line inspection based on deep neural network (DNN) due to the problems of multi‐view‐shape, small‐size object. In this paper, an automatic detection model based on Feature visual clustering network (FVCNet) for power line inspection is established. First, an unsupervised clustering method for power line inspection is proposed, and applied to construct a detection model which can recognize multi‐view‐shape objects and enhanced object features. Then, the bilinear interpolation method is used to Feature enhancement method, and the enhanced high‐level semantics and low‐level semantics are fused to solve the problems of small object size and single sample. In this paper, FVCNet is applied to the MS‐COCO 2017 data set and self‐made power line inspection data set, and the test accuracy is increased to 61.2% and 82.0%, respectively. Compared with other models, especially for the categories that are greatly affected by multi‐view‐shape, the test accuracy has been improved significantly.

Unsupervised AI reveals insect species-specific genome signatures.

Insects are a highly diverse phylogeny and possess a wide variety of traits, including the presence or absence of wings and metamorphosis. These diverse traits are of great interest for studying genome evolution, and numerous comparative genomic studies have examined a wide phylogenetic range of insects. Here, we analyzed 22 insects belonging to a wide phylogenetic range (Endopterygota, Paraneoptera, Polyneoptera, Palaeoptera, and other insects) by using a batch-learning self-organizing map (BLSOM) for oligonucleotide compositions in their genomic fragments (100-kb or 1-Mb sequences), which is an unsupervised machine learning algorithm that can extract species-specific characteristics of the oligonucleotide compositions (genome signatures). The genome signature is of particular interest in terms of the mechanisms and biological significance that have caused the species-specific difference, and can be used as a powerful search needle to explore the various roles of genome sequences other than protein coding, and can be used to unveil mysteries hidden in the genome sequence. Since BLSOM is an unsupervised clustering method, the clustering of sequences was performed based on the oligonucleotide composition alone, without providing information about the species from which each fragment sequence was derived. Therefore, not only the interspecies separation, but also the intraspecies separation can be achieved. Here, we have revealed the specific genomic regions with oligonucleotide compositions distinct from the usual sequences of each insect genome, e.g., Mb-level structures found for a grasshopper Schistocerca americana. One aim of this study was to compare the genome characteristics of insects with those of vertebrates, especially humans, which are phylogenetically distant from insects. Recently, humans seem to be the "model organism" for which a large amount of information has been accumulated using a variety of cutting-edge and high-throughput technologies. Therefore, it is reasonable to use the abundant information from humans to study insect lineages. The specific regions of Mb length with distinct oligonucleotide compositions have also been previously observed in the human genome. These regions were enriched by transcription factor binding motifs (TFBSs) and hypothesized to be involved in the three-dimensional arrangement of chromosomal DNA in interphase nuclei. The present study characterized the species-specific oligonucleotide compositions (i.e., genome signatures) in insect genomes and identified specific genomic regions with distinct oligonucleotide compositions.

An unsupervised automatic organization method for Professor Shirakawa’s hand-notated documents of oracle bone inscriptions

AbstractAs one of the most influential Chinese cultural researchers in the second half of the twentieth-century, Professor Shirakawa is active in the research field of ancient Chinese characters. He has left behind many valuable research documents, especially his hand-notated oracle bone inscriptions (OBIs) documents. OBIs are one of the world’s oldest characters and were used in the Shang Dynasty about 3600 years ago for divination and recording events. The organization of OBIs is not only helpful in better understanding Prof. Shirakawa’s research and further study of OBIs in general and their importance in ancient Chinese history. This paper proposes an unsupervised automatic organization method to organize Prof. Shirakawa’s OBIs and construct a handwritten OBIs data set for neural network learning. First, a suite of noise reduction is proposed to remove strangely shaped noise to reduce the data loss of OBIs. Secondly, a novel segmentation method based on the supervised classification of OBIs regions is proposed to reduce adverse effects between characters for more accurate OBIs segmentation. Thirdly, a unique unsupervised clustering method is proposed to classify the segmented characters. Finally, all the same characters in the hand-notated OBIs documents are organized together. The evaluation results show that noise reduction has been proposed to remove noises with an accuracy of 97.85%, which contains number information and closed-loop-like edges in the dataset. In addition, the accuracy of supervised classification of OBIs regions based on our model achieves 85.50%, which is higher than eight state-of-the-art deep learning models, and a particular preprocessing method we proposed improves the classification accuracy by nearly 11.50%. The accuracy of OBIs clustering based on supervised classification achieves 74.91%. These results demonstrate the effectiveness of our proposed unsupervised automatic organization of Prof. Shirakawa’s hand-notated OBIs documents. The code and datasets are available at http://www.ihpc.se.ritsumei.ac.jp/obidataset.html.

Deep analysis of skin molecular heterogeneities and their significance on the precise treatment of patients with psoriasis.

Psoriasis is a highly heterogeneous autoinflammatory disease. At present, heterogeneity in disease has not been adequately translated into concrete treatment options. Our aim was to develop and verify a new stratification scheme that identifies the heterogeneity of psoriasis by the integration of large-scale transcriptomic profiles, thereby identifying patient subtypes and providing personalized treatment options whenever possible. We performed functional enrichment and network analysis of upregulated differentially expressed genes using microarray datasets of lesional and non-lesional skin samples from 250 psoriatic patients. Unsupervised clustering methods were used to identify the skin subtypes. Finally, an Xgboost classifier was utilized to predict the effects of methotrexate and commonly prescribed biologics on skin subtypes. Based on the 163 upregulated differentially expressed genes, psoriasis patients were categorized into three subtypes (subtypes A-C). Immune cells and proinflammatory-related pathways were markedly activated in subtype A, named immune activation. Contrastingly, subtype C, named stroma proliferation, was enriched in integrated stroma cells and tissue proliferation-related signaling pathways. Subtype B was modestly activated in all the signaling pathways. Notably, subtypes A and B presented good responses to methotrexate and interleukin-12/23 inhibitors (ustekinumab) but inadequate responses to tumor necrosis factor-α inhibitors and interleukin-17A receptor inhibitors. Contrastly, subtype C exhibited excellent responses to tumor necrosis factor-α inhibitors (etanercept) and interleukin-17A receptor inhibitors (brodalumab) but not methotrexate and interleukin-12/23 inhibitors. Psoriasis patients can be assorted into three subtypes with different molecular and cellular characteristics based on the heterogeneity of the skin's immune cells and the stroma, determining the clinical responses of conventional therapies.

Evidence for 2 clusters among patients with noneosinophilic asthma

BackgroundAlthough asthma is often seen as an eosinophilic disease associated with atopy, patients with noneosinophilic asthma represent a substantial part of the population with asthma. ObjectiveTo apply an unsupervised clustering method in a cohort of 588 patients with noneosinophilic asthma (sputum eosinophils < 3%) recruited from an asthma clinic of a secondary care center. MethodsOur cluster analysis of the whole cohort identified 2 subgroups as cluster 1 (n = 417) and cluster 2 (n = 171). ResultsCluster 1 comprised a predominantly female group with late disease onset, a low proportion of atopy (24%), and a substantial smoking history (53%). In this cluster, treatment burden was low (<50% of inhaled corticosteroid users); asthma control and quality of life were poor, with median Asthma Control Test, Asthma Control Questionnaire, and Asthma Quality of Life scores of 16, 1.7, and 4.5, respectively, whereas lung function was preserved with a median postbronchodilation forced expiratory volume in 1 second of 93% predicted. Cluster 2 was a predominantly male group, almost exclusively comprising patients with atopy (99%) with early disease onset and a moderate treatment burden (median inhaled corticosteroids dose 800 µg/d equivalent beclomethasone). In cluster 2, asthma was partially controlled, with median Asthma Control Test and Asthma Control Questionnaire scores reaching 18 and 1.3, respectively, and lung function well preserved with a median postbronchodilation of 95% predicted. Although systemic and airway neutrophilic inflammation was the dominant pattern in cluster 1, cluster 2 essentially comprised paucigranulocytic asthma with moderately elevated fraction exhaled nitric oxide. ConclusionNoneosinophilic asthma splits into 2 clusters distinguishing by disease onset, atopic status, smoking history, systemic and airway inflammation, and disease control and quality of life.