Clustering Algorithm Research Articles

Genomic analysis defines distinct pancreatic and neuronal subtypes of lung carcinoid.

Lung carcinoids (L-CDs) are rare, poorly characterised neuroendocrine tumours (NETs). L-CDs are more common in women and are not the consequence of cigarette smoking. They are classified histologically as typical carcinoids (TCs) or atypical carcinoids (ACs). ACs confer a worse survival. Histological classification is imperfect, and there is increasing interest in molecular markers. We therefore investigated global transcriptomic and epigenomic profiles of 15 L-CDs resected with curative intent at Royal Brompton Hospital. We identified underlying mutations and structural abnormalities through whole-exome sequencing (WES) and single nucleotide polymorphism (SNP) genotyping. Transcriptomic clustering algorithms identified two distinct L-CD subtypes. These showed similarities either to pancreatic or neuroendocrine tumours at other sites and so were named respectively L-CD-PanC and L-CD-NeU. L-CD-PanC tumours featured upregulation of pancreatic and metabolic pathway genes matched by promoter hypomethylation of genes for beta cells and insulin secretion (p < 1 × 10-6). These tumours were centrally located and showed mutational signatures of activation-induced deaminase/apolipoprotein B editing complex activity, together with genome-wide DNA methylation loss enriched in repetitive elements (p = 2.2 × 10-16). By contrast, the L-CD-NeU group exhibited upregulation of neuronal markers (adjusted p < 0.01) and was characterised by focal spindle cell morphology (p = 0.04), peripheral location (p = 0.01), high mutational load (p = 2.17 × 10-4), recurrent copy number alterations, and enrichment for ACs. Mutations affected chromatin remodelling and SWI/SNF complex pathways. L-CD-NeU tumours carried a mutational signature attributable to aflatoxin and aristolochic acid (p = 0.05), suggesting a possible environmental exposure in their pathogenesis. Immunologically, myeloid and T-cell markers were enriched in L-CD-PanC and B-cell markers in L-CD-NeU tumours. The substantial epigenetic and non-coding differences between L-CD-PanC and L-CD-NeU open new possibilities for biomarker selection and targeted treatment of L-CD. © 2024 The Author(s). The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Assessing clustering methods using Shannon's entropy

Unsupervised clustering techniques are crucial for effectively partitioning datasets into meaningful subgroups. In this paper, we introduce a novel clustering fuzziness metric based on Shannon's entropy, which quantifies the level of uncertainty in clustering assignments. This metric is employed to develop a statistical test for evaluating clustering algorithms and to propose parametric corrections to enhance their performance. We provide empirical evidence showing that our approach significantly improves clustering quality compared to well-known algorithms such as Fuzzy K-Means (FKM) and Fanny. For instance, applying our correction method led to a notable decrease in Jensen-Shannon Divergence (JSD) values, indicating enhanced clustering accuracy. Our methodology is demonstrated through comprehensive experiments on both simulated and real-world datasets. Additionally, our approach proves effective in determining the optimal number of clusters, showing superiority over traditional methods. These results highlight the practical benefits of our metric and correction techniques for improving clustering performance.

Identification of key ferroptosis genes and subtypes in kidney renal clear cell carcinoma

Tumour immunity is highly important for the occurrence and development of tumours, and many cancers are resistant to ferroptosis. This study aims to explore the relationship between ferroptosis-related genes (FRGs) and the immunological characteristics of kidney renal clear cell carcinoma (KIRC). We obtained RNA-seq profiles and clinical data of KIRC patients from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases and identified CD44 and GLRX5 as the key FRGs involved in KIRC immune infiltration through Spearman’s correlation analysis. Based on the expression of CD44 and GLRX5, the consensus clustering algorithm was used to classify the TCGA-KIRC samples into two clusters. A nomogram was constructed to evaluate the prognosis of KIRC patients. ESTIMATE, CIBERSORT, and single-sample gene set enrichment analysis (ssGSEA) were performed to evaluate immune infiltration between the two clusters. A weighted gene co-expression network analysis (WGCNA) was used to identify the most relevant genes to the clusters and immunity. Then, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed. The external dataset GSE53757 was used to validate the immunological features between the two clusters. Cluster 2 patients had more active immune infiltration and might be more sensitive to immunotherapy; Cluster 2 patients also had a worse prognosis and might be at a more advanced stage of KIRC. We identified key ferroptosis-related genes and subgroups involved in the immune infiltration of KIRC, which is highly important for exploring the molecular mechanisms and treatments of KIRC.

Fault traceability of power grid dispatching system based on DPHS-MDS and LambdaMART algorithm

Under the background of increasing system service data, it is difficult to trace the faults of power dispatching system. The current fault tracing method has some problems, such as low precision, low efficiency and difficult troubleshooting. To solve this problem, a fault tracing method based on data partition hybrid sampling method and multiple incremental regression tree algorithm is proposed. In this paper, it first uses the hybrid sampling method of data partition and dynamic selection technology to detect the business anomaly, and then applies the clustering algorithm and the information difference graph model to realize the fault tracing of system components. The experimental results showed that the F-metric value and geometric mean value of the study design method were 0.964 and 0.685, respectively. In addition, normalized discounted cumulative gains were observed in the top 10, and the mean average precision of the top 10 was 0.752 and 0.186, respectively. The proposed method can effectively improve the fault tracing efficiency of power grid operation and maintenance personnel, and provide strong data support for the safety maintenance of power grid dispatching system.

Multiple-perspective consumer segmentation using improved weighted Fuzzy k-prototypes clustering and swarm intelligence algorithm for fresh apricot market

Leveraging clustering technology for consumer segmentation is crucial for discerning the nuanced differences among fresh apricot consumer groups and subsequently executing precise marketing strategies. To achieve a more comprehensive and lucid consumer segmentation and identify typical characteristics of apricot consumers in different clusters, this research constructs a novel multiple-perspective segmentation indicator system for fresh apricot consumers. Given the diverse degrees of importance and types of consumer segmentation variables, and the inherent sensitivity of the original Fuzzy k-prototypes (FKP) algorithm to clustering centers, we proposed the weighted Fuzzy k-prototypes (WFKP) algorithms for mixed data (MD) optimized by the particle swarm optimization (PSO) algorithm (MDPSO-WFKP) and mixed data sparrow search algorithm (SSA) (MDSSA-WFKP), both incorporating information entropy weighting for mixed attributes. We test the proposed algorithms on four University of California Irvine machine learning repository (UCI) datasets and the consumer segmentation dataset, and the performance of all selected evaluation indexes shows significant improvement. These findings unequivocally validate the efficacy of the proposed methodologies. Since the MDSSA-WFKP algorithm has the best comprehensive effect on the evaluation indexes, we use it to conduct in-depth apricot consumer segmentation research and find that the apricot consumers can be subdivided into three groups with differentiation: ‘Buddhist-like youths’, ‘Upscale attribute enthusiasts’, and ‘Quality-oriented consumers’. Finally, this paper gives the corresponding marketing suggestions based on the characteristics of the segmented groups.

Enhanced Energy Efficient Clustering and Routing Algorithm in Wireless Sensor Network

Enhanced Energy Efficient Clustering and Routing Algorithm in Wireless Sensor Network

Optimizing large-scale CO2 pipeline networks using a geospatial splitting approach

CO2 transport infrastructure is the backbone of carbon capture and storage (CCS) technology for the mitigation of carbon emissions and project deployment viability. In conventional large-scale CO2 pipeline network designs, the storage sites are generally assumed as the centroids of the major geologic basins, however, this approach might provide suboptimal solutions since the large extension of some storage formations significantly increases the length of the CO2 transportation networks. To address this situation and obtain optimal pipeline routes, we present a novel geospatial splitting framework that partitions large basins into multiple sub-sinks. In our approach, we used a large number of reservoir models varying petrophysical properties and CO2 injection rates to compute pressure plumes through numerical simulations, leading to the calculation of the number of subregions for each basin as a function of the extension of pressure interference areas and boundaries. Finally, we applied K-means clustering and Voronoi polygon algorithms to partition large basins into subregions and obtain their sink coordinates. To demonstrate the capability of the developed workflow, we investigated two CO2 pipeline network modeling case studies using our splitting approach: one regional case study focusing on the Intermountain West (I-West) region and one nationwide case study covering the lower 48 states in the U.S. In both case studies, we compared the optimal pipeline routes using the original and new storage locations and examined the major differences. The use of the developed geospatial approach resulted in both cases in a shortening of the total pipeline network length by 13% and 10%, compared to the pipeline modeling with the original basins, leading to cost reductions of 25% and 17%, respectively, demonstrating that the location of point sinks has a critical impact on the length and expenses of pipelines to efficiently transport CO2 to distant storage sites. Therefore, the workflow presented here contributes to the proper and realistic modeling of case studies that support decision-making in CCS deployment.

Parallel density clustering algorithm based on MapReduce and optimized cuckoo algorithm

Parallel density clustering algorithm based on MapReduce and optimized cuckoo algorithm

Community detection using Jaya optimization algorithm based on deep learning methods and KNN graph-based clustering

Purpose In this paper, community identification has been considered as the most critical task of social network analysis. The purpose of this paper is to organize the nodes of a given network graph into distinct clusters or known communities. These clusters will therefore form the different communities available within the social network graph. Design/methodology/approach To date, numerous methods have been developed to detect communities in social networks through graph clustering techniques. The k-means algorithm stands out as one of the most well-known graph clustering algorithms, celebrated for its straightforward implementation and rapid processing. However, it has a serious drawback because it is insensitive to initial conditions and always settles on local optima rather than finding the global optimum. More recently, clustering algorithms that use a reciprocal KNN (k-nearest neighbors) graph have been used for data clustering. It skillfully overcomes many major shortcomings of k-means algorithms, especially about the selection of the initial centers of clusters. However, it does face its own challenge: sensitivity to the choice of the neighborhood size parameter k, which is crucial for selecting the nearest neighbors during the clustering process. In this design, the Jaya optimization method is used to select the K parameter in the KNN method. Findings The experiment on real-world network data results show that the proposed approach significantly improves the accuracy of methods in community detection in social networks. On the other hand, it seems to offer some potential for discovering a more refined hierarchy in social networks and thus becomes a useful tool in the analysis of social networks. Originality/value This paper introduces an enhancement to the KNN graph-based clustering method by proposing a local average vector method for selecting the optimal neighborhood size parameter k. Furthermore, it presents an improved Jaya algorithm with KNN graph-based clustering for more effective community detection in social network graphs.

Multi-view clustering based on pairwise co-regularization and robust dual graph non-negative matrix factorization

Multi-view clustering has been a hotspot in the field of machine learning and pattern recognition, and methods based on non-negative matrix factorization have gained attention for their simplicity and interpretability. Despite these methods achieving great clustering performance, there may also be some limitations, such as the full structural information of data and the similarity between different views not being considered. This paper proposes a novel multi-view clustering algorithm based on pairwise co-regularization and robust dual graph non-negative matrix factorization. Firstly, the l2,p-norm is applied in non-negative matrix factorization framework to enhance model robustness. Next, pairwise co-regularization is utilized to extract the inter-view information of all views. Then, graph dual regularization is applied to preserve the structure information of the data and feature spaces. Lastly, an auto-weighted strategy is introduced to assign appropriate weights to each view. In addition, an iterative updating optimization scheme for the proposed algorithm is developed, and the convergence proof of the scheme is provided. The experimental results on twelve real-world datasets show that the proposed algorithm is both effective and efficient, compared with three classical and six state-of-the-art algorithms.

A survey of genetic algorithms for clustering: Taxonomy and empirical analysis

Clustering, an unsupervised learning technique, aims to group patterns into clusters where similar patterns are grouped together, while dissimilar ones are placed in different clusters. This task can present itself as a complex optimization problem due to the extensive search space generated by all potential data partitions. Genetic Algorithms (GAs) have emerged as efficient tools for addressing this task. Consequently, significant advancements and numerous proposals have been developed in this field.This work offers a comprehensive and critical review of state-of-the-art mono-objective Genetic Algorithms (GAs) for partitional clustering. From a more theoretical standpoint, it examines 22 well-known proposals in detail, covering their encoding strategies, objective functions, genetic operators, local search methods, and parent selection strategies. Based on this information, a specific taxonomy is proposed. In addition, from a more practical standpoint, a detailed experimental study is conducted to discern the advantages and disadvantages of approaches. Specifically, 22 different cluster validation indices are considered to compare the performance of clustering techniques. This evaluation is performed across 94 datasets encompassing diverse configurations, including the number of classes, separation between classes, and pattern dimensionality. Results reveal interesting findings, such as the key role of local search in optimizing results and reducing search space. Additionally, representations based on centroids and labels demonstrate greater efficiency and crossover and mutation operators do not prove to be as relevant. Ultimately, while the results are satisfactory, real-world clustering problems introduce additional complexity, especially for algorithms aiming to determine the number of clusters, resulting in diminished performance and the need for new approaches to be explored. Code, datasets and instructions to run algorithms in the LEAL library are available in an associated repository, in order to facilitate future experiments in this environment.

A clustering tool for generating biological geometries for computational modeling in radiobiology.

To develop a computational tool that converts biological images into geometries compatible with computational software dedicated to the Monte Carlo simulation of radiation transport (TOPAS), and subsequent biological tissue responses (CompuCell3D). The depiction of individual biological entities from segmentation images is essential in computational radiobiological modeling for two reasons: image pixels or voxels representing a biological structure, like a cell, should behave as a single entity when simulating biological processes, and the action of radiation in tissues is described by the association of biological endpoints to physical quantities, as radiation dose, scored the entire group of voxels assembling a cell.&#xD;Approach: The tool is capable of cropping and resizing the images and performing clustering of image voxels to create independent entities (clusters) by assigning a unique identifier to these voxels conforming to the same cluster. The clustering algorithm is based on the adjacency of voxels with image values above an intensity threshold to others already assigned to a cluster. The performance of the tool to generate geometries that reproduced original images was evaluated by the Dice Similarity Coefficient (DSC), and by the number of individual entities in both geometries. A set of tests consisting of segmentation images of cultured neuroblastoma cells, two cell nucleus populations, and the vasculature of a mouse brain were used.&#xD;Main results: The DSC was 1.0 in all images, indicating that original and generated geometries were identical, and the number of individual entities in both geometries agreed, proving the ability of the tool to cluster voxels effectively following user-defined specifications. The potential of this tool in computational radiobiological modeling, was shown by evaluating the spatial distribution of DNA Double-Strand-Breaks after microbeam irradiation in a segmentation image of a cell culture.&#xD;Significance: This tool enables the use of realistic biological geometries in computational radiobiological studies.

FunMaps: a method for parcellating functional brain networks using resting-state functional MRI data.

Parcellations of resting-state functional magnetic resonance imaging (rs-fMRI) data are widely used to create topographical maps of functional networks in the human brain. While such network maps are highly useful for studying brain organization and function, they usually require large sample sizes to make them, thus creating practical limitations for researchers that would like to carry out parcellations on data collected in their labs. Furthermore, it can be difficult to quantitatively evaluate the results of a parcellation since networks are usually identified using a clustering algorithm, like principal components analysis, on the results of a single group-averaged connectivity map. To address these challenges, we developed the FunMaps method: a parcellation routine that intrinsically incorporates stability and replicability of the parcellation by keeping only network distinctions that agree across halves of the data over multiple random iterations. Here, we demonstrate the efficacy and flexibility of FunMaps, while describing step-by-step instructions for running the program. The FunMaps method is publicly available on GitHub (https://github.com/persichetti-lab/FunMaps). It includes source code for running the parcellation and auxiliary code for preparing data, evaluating the parcellation, and displaying the results.

Research on the Correlation between the Dynamic Distribution Patterns of Urban Population Density and Land Use Morphology Based on Human–Land Big Data: A Case Study of the Shanghai Central Urban Area

The dynamic distribution of urban population density and the interaction with land use elements involve mutual constraints and guidance. However, in the existing research on the relationship between urban population density and land use, the discussion on the distribution patterns of urban population density typically spans long time periods and uses large spatial units, lacking analysis of the dynamic changes in population density within high granularity land parcels over a day. In studies related to the urban built environment, the complex relationships between different-dimensional land use elements and the dynamic distribution of population density also need further exploration. To address these bottlenecks, this study takes Shanghai’s central urban area as an example. Based on 24 h mobile signaling data on weekdays, weekends, and typical holidays, as well as urban land use data, clustering algorithms are used to summarize patterns of dynamic population density distribution. Pearson correlation analysis is then employed to study the correlation between dynamic population density distribution patterns and different land use elements. The results indicate that various urban land use factors such as locational centrality, functional diversity, transportation accessibility, compactness, and landscape quality have different impacts on the dynamic distribution of population density in spatial units, and the dynamic distribution patterns of population density in different land use types also vary. This research contributes to guiding the optimization of spatial quality and formulating planning and management measures that more effectively match construction intensity with population activity density.

Implementasi Algoritma K-Means Guna Pengelompokkan Data Penjualan Berdasarkan Pembelian

Information technology can help solve problems faced by humans by facilitating performance. Information technology and information systems are difficult to separate in the business world. Data mining is the core of the KDD process, which involves inferring algorithms that explore data, developing models and finding previously unknown patterns. KDD is often referred to as knowledge discovery in databases. The KDD process generally consists of 5 stages, namely data selection, pre-processing/cleaning, transformation, data mining and interpretation/evaluation. K-Means is a clustering algorithm in data mining to be able to produce groups of large amounts of data with a point-based partition method with fast and efficient computing time. Clustering is the process of dividing objects from a data set into several homogeneous clusters. The main purpose of the cluster method is to group a number of data/objects into clusters (groups) so that each cluster will contain data that is as similar as possible. This study aims to provide real solutions to UD. Martua in order to know which items are selling well and which items are not selling well so that the object can know which items need to be added to the stock and which items need to be reduced. The method used in this study is the K-Means method with stages, namely data selection, pre-processing, data transformation, information extraction and evaluation of results. The data consists of 30 item data, there are 8 as members of C1 and are best-selling items and 22 as members of C2 and are not selling items. The conclusion that can be obtained from this study is that the K-Means method can group items at UD. Martua. This study shows that the implementation of the K-Means method with the support of the RapidMiner application is effective in grouping item data at UD. Martua.

Binning Metagenomic Contigs Using Contig Embedding and Decomposed Tetranucleotide Frequency.

Metagenomic binning is a crucial step in metagenomic research. It can aggregate the genome sequences belonging to the same microbial species into independent bins. Most existing methods ignore the semantic information of contigs and lack effective processing of tetranucleotide frequency, resulting in insufficient and complex feature information extracted for binning and poor binning results. To address the above problems, we propose CedtBin, a metagenomic binning method based on contig embedding and decomposed tetranucleotide frequency. First, the improved BERT model is used to learn the contigs to obtain their embedding representation. Secondly, the tetranucleotide frequencies are decomposed using a non-negative matrix factorization (NMF) algorithm. After that, the two features are spliced and input into the clustering algorithm for binning. Considering the sensitivity of the DBSCAN clustering algorithm to input parameters, in order to solve the drawbacks of manual parameter input, we also propose an Annoy-DBSCAN algorithm that can adaptively determine the parameters of the DBSCAN algorithm. This algorithm uses Approximate Nearest Neighbors Oh Yeah (Annoy) and combines it with a grid search strategy to find the optimal parameters of the DBSCAN algorithm. On simulated and real datasets, CedtBin achieves better binning results than mainstream methods and can reconstruct more genomes, indicating that the proposed method is effective.

GenoTriplo: A SNP genotype calling method for triploids.

Triploidy is very useful in both aquaculture and some cultivated plants as the induced sterility helps to enhance growth and product quality, as well as acting as a barrier against the contamination of wild populations by escapees. To use genetic information from triploids for academic or breeding purposes, an efficient and robust method to genotype triploids is needed. We developed such a method for genotype calling from SNP arrays, and we implemented it in the R package named GenoTriplo. Our method requires no prior information on cluster positions and remains unaffected by shifted luminescence signals. The method relies on starting the clustering algorithm with an initial higher number of groups than expected from the ploidy level of the samples, followed by merging groups that are too close to each other to be considered as distinct genotypes. Accurate classification of SNPs is achieved through multiple thresholds of quality controls. We compared the performance of GenoTriplo with that of fitPoly, the only published method for triploid SNP genotyping with a free software access. This was assessed by comparing the genotypes generated by both methods for a dataset of 1232 triploid rainbow trout genotyped for 38,033 SNPs. The two methods were consistent for 89% of the genotypes, but for 26% of the SNPs, they exhibited a discrepancy in the number of different genotypes identified. For these SNPs, GenoTriplo had >95% concordance with fitPoly when fitPoly genotyped better. On the contrary, when GenoTriplo genotyped better, fitPoly had less than 50% concordance with GenoTriplo. GenoTriplo was more robust with less genotyping errors. It is also efficient at identifying low-frequency genotypes in the sample set. Finally, we assessed parentage assignment based on GenoTriplo genotyping and observed significant differences in mismatch rates between the best and second-best couples, indicating high confidence in the results. GenoTriplo could also be used to genotype diploids as well as individuals with higher ploidy level by adjusting a few input parameters.

Intrusion detection systems for the internet of things: a probabilistic anomaly detection approach

The expansion of IoT domains presents new cybersecurity challenges. In this paper, we propose an enhanced Intrusion Detection System (IDS) designed to tackle these emerging security issues. Our approach harnesses the power of unsupervised learning techniques, particularly the Variational Autoencoder (VAE), to reduce data dimensionality while preserving essential features. By integrating advanced clustering algorithms–including K-means, Gaussian Mixture Model (GMM), and Variational Bayesian Gaussian Mixture Model (VBGMM)–we facilitate precise data categorization within IoT networks. Our conducted experiments using the IoT Network Intrusion Dataset demonstrate that VAE-based methods consistently outperform traditional Autoencoders (AE), achieving superior performance across critical metrics such as False Alarm Rate (FAR), Detection Rate (DR), Accuracy, Precision, and Area Under the ROC Curve (AUC).

MAMCL: Multi-attributes Masking Contrastive Learning for explainable seismic facies analysis

Seismic facies analysis is crucial in hydrocarbon exploration and development. Traditional machine learning approaches typically require manual selection of attributes and lack interpretability analysis. We propose an interpretable framework, multi-attribute masking contrastive learning (MAMCL), designed to adaptively select, explore and aggregate seismic attributes for seismic facies analysis. The MAMCL framework includes a depthwise CNN module for feature extraction and an iTransformer module for feature aggregation. Based on the assumption that different attributes computed on the same seismic sample imply common information associated with the same geologic facies, we formulate an unsupervised strategy of contrastive learning to pre-train the MAMCL framework for refining the attributes. This pre-training method encourages the network to extract and integrate highly correlated attribute features by enhancing the expression of commonalities within the same sample, and implicitly increase the distance between features of different categories by differentiating the expressions of different samples. Ultimately, these refined features only need to be input into a simple clustering algorithm, such as K-Means, to achieve seismic facies classification. MAMCL requires no labels or manual selection of attributes and can utilize the self-attention mechanism of iTransformer to compute adaptive attribute weights, facilitating interpretability analysis. We applied MAMCL framework to both unlogged turbidite channel systems in Canterbury Basin, New Zealand, and logged Chengdao area in Bohai Bay Basin, China, achieving reliable classification results and providing interpretability analysis.

A Multi-Omics Prognostic Model Capturing Tumor Stemness and the Immune Microenvironment in Clear Cell Renal Cell Carcinoma.

Background: Cancer stem-like cells (CSCs), a distinct subset recognized for their stem cell-like abilities, are intimately linked to the resistance to radiotherapy, metastatic behaviors, and self-renewal capacities in tumors. Despite their relevance, the definitive traits and importance of CSCs in the realm of oncology are still not fully comprehended, particularly in the context of clear cell renal cell carcinoma (ccRCC). A comprehensive understanding of these CSCs' properties in relation to stemness, and their impact on the efficacy of treatment and resistance to medication, is of paramount importance. Methods: In a meticulous research effort, we have identified new molecular categories designated as CRCS1 and CRCS2 through the application of an unsupervised clustering algorithm. The analysis of these subtypes included a comprehensive examination of the tumor immune environment, patterns of metabolic activity, progression of the disease, and its response to immunotherapy. In addition, we have delved into understanding these subtypes' distinctive clinical presentations, the landscape of their genomic alterations, and the likelihood of their response to various pharmacological interventions. Proceeding from these insights, prognostic models were developed that could potentially forecast the outcomes for patients with ccRCC, as well as inform strategies for the surveillance of recurrence after treatment and the handling of drug-resistant scenarios. Results: Compared with CRCS1, CRCS2 patients had a lower clinical stage/grading and a better prognosis. The CRCS2 subtype was in a hypoxic state and was characterized by suppression and exclusion of immune function, which was sensitive to gefitinib, erlotinib, and saracatinib. The constructed prognostic risk model performed well in both training and validation cohorts, helping to identify patients who may benefit from specific treatments or who are at risk of recurrence and drug resistance. A novel therapeutic target, SAA2, regulating neutrophil and fibroblast infiltration, and, thus promoting ccRCC progression, was identified. Conclusions: Our findings highlight the key role of CSCs in shaping the ccRCC tumor microenvironment, crucial for therapy research and clinical guidance. Recognizing tumor stemness helps to predict treatment efficacy, recurrence, and drug resistance, informing treatment strategies and enhancing ccRCC patient outcomes.