Search Results for « UMAP »

Abstract Background The application of Uniform Manifold Approximation and Projection (UMAP) for dimensionality reduction and visualization has revolutionized the analysis of single-cell RNA expression and population genetics. However, its potential in single-cell DNA sequencing data analysis, particularly for visualizing gene mutation information, has not been fully explored. Results We introduce Mugen-UMAP, a novel Python-based program that extends UMAP's utility to single-cell DNA sequencing data. This innovative tool provides a comprehensive pipeline for processing gene annotation files of single-cell somatic single-nucleotide variants and metadata to the visualization of UMAP projections for identifying clusters, along with various statistical analyses. Employing Mugen-UMAP, we analyzed whole-exome sequencing data from 365 single-cell samples across 12 non-small cell lung cancer (NSCLC) patients, revealing distinct clusters associated with histological subtypes and uncovering patterns of gene mutation that suggest a deeper insight into tumor heterogeneity and evolution. Conclusions The application of Mugen-UMAP demonstrates its capacity to provide valuable insights into the visualization and interpretation of single-cell DNA sequencing data. Mugen-UMAP can be found at https://github.com/tengchn/Mugen-UMAP

Abstract Background The application of Uniform Manifold Approximation and Projection (UMAP) for dimensionality reduction and visualization has revolutionized the analysis of single-cell RNA expression and population genetics. However, its potential in single-cell DNA sequencing data analysis, particularly for visualizing gene mutation information, has not been fully explored. Results We introduce Mugen-UMAP, a novel Python-based program that extends UMAP’s utility to single-cell DNA sequencing data. This innovative tool provides a comprehensive pipeline for processing gene annotation files of single-cell somatic single-nucleotide variants and metadata to the visualization of UMAP projections for identifying clusters, along with various statistical analyses. Employing Mugen-UMAP, we analyzed whole-exome sequencing data from 365 single-cell samples across 12 non-small cell lung cancer (NSCLC) patients, revealing distinct clusters associated with histological subtypes of NSCLC. Moreover, to demonstrate the general utility of Mugen-UMAP, we applied the program to 9 additional single-cell WES datasets from various cancer types, uncovering interesting patterns of cell clusters that warrant further investigation. In summary, Mugen-UMAP provides a quick and effective visualization method to uncover cell cluster patterns based on the gene mutation information from single-cell DNA sequencing data. Conclusions The application of Mugen-UMAP demonstrates its capacity to provide valuable insights into the visualization and interpretation of single-cell DNA sequencing data. Mugen-UMAP can be found at https://github.com/tengchn/Mugen-UMAP

Abstract We have already explained the geometric phenomena employed to replace Lemma 1 in the UMAP algorithm in Chap. 3 and discussed the aspects from category theory that are necessary to design this pipeline in an abstract manner in Chap. 4 .

Objective/context: This article focuses on information about the Military Units to Aid Production found in Cuban periodical publications like El Mundo, Granma and Verde Olivo between 1965 and 1968 about the Unidades Militares de Ayuda a la Producción (UMAP). Originality: Very few studies have dealt with this subject as a way to reconstruct the economic, socio-political and cultural aspects of the period in which this revolutionary institution arose. Methodology: This investigation has been guided by a focus which favors an analysis of documents, due to the absence of other kinds of historical sources. Conclusions: These Units were created with the aim of reforming some persons, like priests, homosexuals, dissidents and the unemployed, through agricultural work, which was regarded as a corrective measure on an individual and political-ideological level.

Abstract In this chapter, we provide a comprehensive description of the IsUMap scheme, an alternative to the UMAP algorithm which was already briefly introduced in Sect. 5.2.5 . The development of IsUMap has been presented gradually throughout the text, particularly in the Sect. 3.6 in Chap. 3 and in Sect. 5.2.5 in Chap. 5 . Here, we now offer a detailed exposition of the algorithm, directing readers to the corresponding sections for a deeper understanding of the theoretical foundation of each step.

Uniform Manifold Approximation and Projection (UMAP) relies on a neighborhood graph constructed from local distances; errors in this graph can therefore contribute to manifold fragmentation and the compression of geometrically distinct regions in low-dimensional embeddings. We present JORC-UMAP, a graph-stage modification that reweights UMAP's fuzzy neighborhood graph using Ollivier-Ricci curvature (ORC) and Jaccard neighborhood overlap. ORC provides an edgewise geometric signal, while Jaccard overlap serves as an empirical gate: negative-curvature edges with sufficient overlap are reinforced, low-overlap candidates are strongly attenuated, and nonnegative-curvature edges are retained or attenuated through a bounded response. The reweighted graph is then passed to the standard UMAP optimization stage. We evaluate JORC-UMAP on synthetic manifolds, noisy bridge constructions, and five labeled benchmarks against t-SNE, UMAP, TriMap, and PaCMAP. Controlled Trefoil Knot and Swiss Roll experiments illustrate reduced structural collapse and manifold fragmentation relative to standard UMAP, while a three-ring ablation illustrates how Jaccard gating can suppress low-overlap noisy connections that curvature-only reweighting may amplify. On the five labeled benchmarks, descriptive grid-optimized comparisons are mixed: JORC-UMAP achieves the highest class-median distance-ordering accuracy on USPS and the second-highest on MNIST, Fashion-MNIST, and 20 Newsgroups. These results provide preliminary evidence that curvature- and overlap-based graph reweighting is a useful intervention within UMAP, while broader repeated evaluations and more efficient curvature estimation remain necessary.

Penelitian ini bertujuan memetakan struktur wacana publik mengenai Danantara sebagai lembaga pengelola investasi negara melalui pemodelan topik berbasis BERTopic yang dioptimasi dengan UMAP dan HDBSCAN. Ketidakpastian informasi dan meningkatnya volume percakapan digital terkait Danantara menuntut pendekatan analitis yang mampu menangkap keragaman isu secara lebih komprehensif dibandingkan analisis sentimen konvensional. Data diperoleh dari himpunan cuitan yang telah melalui tahapan pra-pemrosesan, penghapusan duplikasi, serta deteksi anomali temporal untuk meminimalkan distorsi akibat ledakan percakapan tidak organik. BERTopic dikonfigurasi dengan embedding IndoSBERT-large yang secara spesifik dirancang untuk representasi semantik teks Bahasa Indonesia, serta penyesuaian parameter UMAP–HDBSCAN guna menghasilkan struktur klaster yang stabil. Hasil pemodelan menghasilkan lima belas topik yang merepresentasikan spektrum wacana mulai dari dukungan terhadap penguatan ekonomi nasional, pembahasan energi dan BUMN, kritik terhadap isu legalitas dan akuntabilitas, hingga percakapan ringan yang bersifat humor. Temuan ini menunjukkan bahwa percakapan publik mengenai Danantara bersifat heterogen dan dipengaruhi oleh dinamika sosial-politik di media sosial. Secara keseluruhan, pendekatan BERTopic dengan optimalisasi UMAP–HDBSCAN menunjukkan efektivitas dalam mengungkap struktur diskursus yang berlapis berdasarkan tingkat koherensi dan keragaman topik yang dihasilkan serta menyediakan dasar berbasis data bagi analisis persepsi publik dan formulasi strategi komunikasi kebijakan.