Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “unsupervised clustering”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 397 records · Page 22Linked to original sources
PubMed1995

Application of fuzzy c-means segmentation technique for tissue differentiation in MR images of a hemorrhagic glioblastoma multiforme.

The application of a raw data-based, operator-independent MR segmentation technique to differentiate boundaries of tumor from edema or hemorrhage is demonstrated. A case of a glioblastoma multiforme with gross and histopathologic correlation is presented. The MR image data set was segmented into tissue classes based on three different MR weighted image parameters (T1-, proton density-, and T2-weighted) using unsupervised fuzzy c-means (FCM) clustering algorithm technique for pattern recognition. A radiological examination of the MR images and correlation with fuzzy clustering segmentations was performed. Results were confirmed by gross and histopathology which, to the best of our knowledge, reports the first application of this demanding approach. Based on the results of neuropathologic correlation, the application of FCM MR image segmentation to several MR images of a glioblastoma multiforme represents a viable technique for displaying diagnostically relevant tissue contrast information used in 3D volume reconstruction. With this technique, it is possible to generate segmentation images that display clinically important neuroanatomic and neuropathologic tissue contrast information from raw MR image data.

Adult↗
PubMed2004

Global search for chromosomal abnormalities in infiltrating ductal carcinoma of the breast using array-comparative genomic hybridization.

Array-comparative genomic hybridization (a-CGH) is a molecular cytogenetic technique for detection of multiple chromosomal abnormalities in genomic DNA samples. Using an a-CGH with 287 probes, we examined 14 cases of breast infiltrating ductal carcinoma (IDCA) that had previously been classified by fluorescent in situ hybridization (FISH) as either human epidermal growth factor receptor-2 positive (HER2+) or HER2- and analyzed the data by hierarchical, K-means, and principal component analyses. The aim of the study was to identify the genetic abnormalities that are present in breast IDCAs and determine if the global status of 287 cytogenetic locations could be used as a more objective method for breast IDCA classification. Concordance between FISH and a-CGH at the HER2 locus was 78.6% (11/14). In general, a-CGH detected more abnormalities in HER2+ cases. In HER 2+ cases, chromosomes 1, 2, 3, 7, 9, 17, and 20 had more regions that showed statistically significant (P < or = 0.01) changes in DNA copy number. Among all the aberrant cytogenetic locations detected, 20q13, 7p12.3 approximately p12.1, and 17q23.2 approximately q25.3, which contain among others, genes for TNFRSF6B, EGFR, and TK1 showed statistically significant gains (P < or = 0.01) in 83, 66.7, and 50% of the HER2+ IDCA cases, respectively. Chromosome location 8q24.12 approximately q24.13 was the only region that showed consistent amplification in approximately 50% of the HER2- cases. Unsupervised hierarchical and K-means cluster analyses and principal component analysis using the DNA copy number status of 287 cytogenetic locations or the 177 cytogenetic locations that showed statistically significant differences revealed a cluster consisting of mainly HER2- IDCA cases. Even though this study demonstrates the usefulness of a-CGH in the rapid identification of aberrant DNA regions in tumor samples, we conclude that an array-CGH with more than 287 probes will be needed for a more precise mapping of DNA aberrations at the global level.

Breast Neoplasms↗
PubMedDate not supplied

Automated analysis of proton NMR spectra from combinatorial rapid parallel synthesis using self-organizing maps.

It is now quite routine to acquire proton NMR spectra of compounds in 96-well plates prepared in a rapid parallel synthesis fashion using a flow-NMR automation setup. However, the analysis of 96 NMR spectra obtained in this manner is often laborious and painstakingly slow. We have developed a new, automated method for rapidly analyzing 96 NMR spectra of compounds synthesized in an 8 x 12 matrix using self-organizing maps (SOM). This unsupervised neural network is capable of clustering together NMR spectra containing a common pattern of -R groups and identifying outliers from within such clusters. Analysis of these outlier spectra can quickly help indicate the presence of undesired products, impurities, starting materials, and other unexpected errors in a 96-well plate synthesis by focusing the chemists' attention on the aberrant NMR spectra. Thus, SOM can be a valuable tool in performing efficient quality control on combinatorial libraries.

Combinatorial Chemistry Techniques↗
PubMed2006

Multivariable difference gel electrophoresis and mass spectrometry: a case study on transforming growth factor-beta and ERBB2 signaling.

Multivariable DIGE/MS was used to investigate proteins altered in expression and/or post-translational modification in response to activation of transforming growth factor (TGF)-beta receptors in MCF10A mammary epithelial cells overexpressing the HER2/Neu (ErbB2) oncogene. Proteome changes were monitored in response to exogenous TGF-beta over time (0, 8, 24, and 40 h), and proteins were resolved using medium range (pH 4-7) and narrow range (pH 5.3-6.5) isoelectric focusing combined with up to 2 mg of protein to allow inspection of lower abundance proteins. Triplicate samples were prepared independently and analyzed together across multiple DIGE gels using a pooled sample internal standard to quantify expression changes with statistical confidence. Unsupervised principle component analysis and hierarchical clustering of the individual DIGE proteome expression maps provided independent confirmation of distinct expression patterns from the individual experiments and demonstrated high reproducibility between replicate samples. Fifty-nine proteins (including some isoforms) that exhibited significant kinetic expression changes were identified using mass spectrometry and database interrogation and were mapped to existing biological networks involved in TGF-beta signaling. Several proteins with a potential role in breast cancer, such as maspin and cathepsin D, were identified as novel molecules associated with TGF-beta signaling.

Amino Acid Sequence↗
PubMed2002

Unsupervised technique for robust target separation and analysis of DNA microarray spots through adaptive pixel clustering.

MOTIVATION: Microarray images challenge existing analytical methods in many ways given that gene spots are often comprised of characteristic imperfections. Irregular contours, donut shapes, artifacts, and low or heterogeneous expression impair corresponding values for red and green intensities as well as their ratio R/G. New approaches are needed to ensure accurate data extraction from these images. RESULTS: Herein we introduce a novel method for intensity assessment of gene spots. The technique is based on clustering pixels of a target area into foreground and background. For this purpose we implemented two clustering algorithms derived from k-means and Partitioning Around Medoids (PAM), respectively. Results from the analysis of real gene spots indicate that our approach performs superior to other existing analytical methods. This is particularly true for spots generally considered as problematic due to imperfections or almost absent expression. Both PX(PAM) and PX(KMEANS) prove to be highly robust against various types of artifacts through adaptive partitioning, which more correctly assesses expression intensity values. AVAILABILITY: The implementation of this method is a combination of two complementary tools Extractiff (Java) and Pixclust (free statistical language R), which are available upon request from the authors.

Algorithms↗
PubMed2001

Classification of chronic radiation sickness cases using neural networks and classification trees.

Chronic radiation sickness is a deterministic radiation health effect observed among the Mayak Production Association workers in Russia. In this study, unsupervised neural networks were used to cluster hematological measurements in a subset (n = 88) of the Mayak Production Association population while excluding from the analysis the radiation dose and the historical clinical diagnosis. Clusters of observations that had lower average leukocyte and thrombocyte counts were labeled "affected" and those having higher average blood cell counts were labeled "unaffected." The class (cluster) membership for each individual was used subsequently as a dependent variable in a classification tree model in order to identify significant features of the underlying classification model. After re-classification of cases using this method, the results showed a better data separation between the blood cell counts for affected vs. unaffected groups compared to those based on historical classification, and a greater difference between group means for differential blood counts was observed than for the historical diagnosis. The reclassification of diagnostic groups changed the group mean radiation doses. The geometric means (and 95% CL) of cumulative radiation dose equivalent from external exposures, based on the historical diagnosis, are 0.31 (0.0035, 3.4) vs. 1.7 (0.0007, 18) Sv. After clustering and classification tree analyses, the group geometric means were 0.78 (0.0014, 8.6) vs. 1.5 (0.0007, 17) and 0.82 (0.0013, 9.0) vs. 1.4 (0.0008, 16) Sv, using (respectively) whole blood cell counts or differential counts as the independent variables. The approach presented here is useful as a diagnostic aid for both retrospective analyses and in the event of future radiation accidents.

Chronic Disease↗
PubMed2006

Structured one-class classification.

The one-class classification problem aims to distinguish a target class from outliers. The spherical one-class classifier (SOCC) solves this problem by finding a hypersphere with minimum volume that contains the target data while keeping outlier samples outside. SOCC achieves satisfactory performance only when the target samples have the same distribution tendency in all orientations. Therefore, the performance of the SOCC is limited in the way that many superfluous outliers might be mistakenly enclosed. The authors propose to exploit target data structures obtained via unsupervised methods such as agglomerative hierarchical clustering and use them in calculating a set of hyperellipsoidal separating boundaries. This method is named the structured one-class classifier (TOCC). The optimization problem in TOCC can be formulated as a series of second-order cone programming problems that can be solved with acceptable efficiency by primal-dual interior-point methods. The experimental results on artificially generated data sets and benchmark data sets demonstrate the advantages of TOCC.

Journal Article↗
PubMed2005

Quantitative transcriptional profiling of ATDC5 mouse progenitor cells during chondrogenesis.

During the differentiation of a mouse chondroprogenitor cell line, ATDC5, an analysis of the transcription cartilage-related genes was carried out using real-time RT-PCR in a semiquantitative fashion. A total number of 104 genes both previously linked to chondrogenesis and hitherto not associated with the development of cartilage were analyzed. Parametric statistics, and unsupervised hierarchical and K-medians clustering approaches were used to analyze the gene expression during the sequential processes of proliferation, condensation, differentiation, maturation, and hypertrophic conversion of ATDC5 cells. The obtained data provided a robust determination of expression patterns that make possible an accurate assessment of the molecular events along the chondrogenic differentiation pathway. In addition, time-course expression profiles were described for eight highly regulated genes that have not been associated with chondrogenesis as yet. These included Cryab, Rcor2, Hig1, Bnip3, Mst4, Calml4, Gng2, and Islr.

Animals↗
PubMedDate not supplied

The use of urine proteomic and metabonomic patterns for the diagnosis of interstitial cystitis and bacterial cystitis.

The advent of systems biology approaches that have stemmed from the sequencing of the human genome has led to the search for new methods to diagnose diseases. While much effort has been focused on the identification of disease-specific biomarkers, recent efforts are underway toward the use of proteomic and metabonomic patterns to indicate disease. We have developed and contrasted the use of both proteomic and metabonomic patterns in urine for the detection of interstitial cystitis (IC). The methodology relies on advanced bioinformatics to scrutinize information contained within mass spectrometry (MS) and high-resolution proton nuclear magnetic resonance (1H-NMR) spectral patterns to distinguish IC-affected from non-affected individuals as well as those suffering from bacterial cystitis (BC). We have applied a novel pattern recognition tool that employs an unsupervised system (self-organizing-type cluster mapping) as a fitness test for a supervised system (a genetic algorithm). With this approach, a training set comprised of mass spectra and 1H-NMR spectra from urine derived from either unaffected individuals or patients with IC is employed so that the most fit combination of relative, normalized intensity features defined at precise m/z or chemical shift values plotted in n-space can reliably distinguish the cohorts used in training. Using this bioinformatic approach, we were able to discriminate spectral patterns associated with IC-affected, BC-affected, and unaffected patients with a success rate of approximately 84%.

Bacterial Infections↗
PubMed2006

Gene-expression profiling of systemic anaplastic large-cell lymphoma reveals differences based on ALK status and two distinct morphologic ALK+ subtypes.

With the use of microarray gene-expression profiling, we analyzed a homogeneous series of 32 patients with systemic anaplastic large-cell lymphoma (ALCL) and 5 ALCL cell lines. Unsupervised analysis classified ALCL in 2 clusters, corresponding essentially to morphologic subgroups (ie, common type vs small cell and "mixed" variants) and clinical variables. Patients with a morphologic variant of ALCL had advanced-stage disease. This group included a significant number of patients who experienced early relapse. Supervised analysis showed that ALK+ALCL and ALK- ALCL have different gene-expression profiles, further confirming that they are different entities. Among the most significantly differentially expressed genes between ALK+ and ALK- samples, we found BCL6, PTPN12, CEBPB, and SERPINA1 genes to be overexpressed in ALK+ ALCL. This result was confirmed at the protein level for BCL-6, C/EBPbeta and serpinA1 through tissue microarrays. The molecular signature of ALK- ALCL included overexpression of CCR7, CNTFR, IL22, and IL21 genes but did not provide any obvious clues to the molecular mechanism underlying this tumor subtype. Once confirmed on a larger number of patients, the results of the present study could be used for clinical and therapeutic management of patients at the time of diagnosis.

Anaplastic Lymphoma Kinase↗
PubMed2003

Karyometry of nuclei from actinic keratosis and squamous cell cancer of the skin.

OBJECTIVE: To use karyometric analysis methods to compare actinic keratoses (AKs) to squamous cell carcinomas (SCCs) to determine if SCCs showed a logical progression beyond that seen in AKs and to explore variability within and between lesion types to better understand distinctions between the 2. STUDY DESIGN: Biopsies from 31 subjects with AKs were obtained from upper inner arm skin, forearm skin and AK lesions. Biopsies from 23 different subjects in a related subproject provided SCC biopsies for comparison. RESULTS: Karyometric measures of nuclear abnormality and sun damage were derived. Mean actinic damage levels progressed logically from inner arm to sun-exposed skin, to AK, to SCC. Considerable heterogeneity existed at the case level. Unsupervised learning methods revealed 2 distinct clusters of progressed lesions with different nuclear signatures, reflecting differing levels of actinic damage. Number of AKs and SCCs and invasiveness and differentiation of SCCs were distributed across both clusters in roughly equivalent proportions. CONCLUSION: Karyometric methods, shown previously to be capable of sensitively detecting subtle nuclear changes, revealed the possibility of 2 progression pathways, each containing AKs and SCCs. This finding may have prognostic implications.

Adolescent↗
PubMed2000

Detection of multiple sclerosis with visual evoked potentials--an unsupervised computational intelligence system.

This paper describes the application of a novel unsupervised pattern recognition system to the classification of the Visual Evoked Potentials (VEP's) of normal and multiple sclerosis (MS) patients. The method combines a traditional statistical feature extractor with a fuzzy clustering method, all implemented in a parallel neural network architecture. The optimization routine, ALOPEX, is used to train the network while decreasing the likelihood of local solutions. The unsupervised system includes a feature extraction and clustering module, trained by the optimization routine ALOPEX. Through maximization of the output variance of each node, and an architecture which excludes redundancy, the feature extraction network retains the most significant Karhunen-Loeve expansion vectors. The clustering module uses a modification to the Fuzzy c-Means (FCM) clustering algorithms, where ALOPEX adjusts a set of cluster centers to minimize an objective error function. The result combines the power of the FCM algorithms with the advantage of a more global solution from ALOPEX. The new pattern recognition system is used to cluster the VEP's of 13 normal and 12 MS subjects. The classification with this technique can, without supervision, separate the patient population into two groups which largely correspond to the MS and control subject groups. A suitable threshold can be chosen so that the recognizer chooses no false negatives. The use of multiple stimulation patterns appears to improve the reliability of the decision. The reasoning of most neural networks in their decision making cannot easily be extracted upon the completion of training. However, due to the linearity of the network nodes, the cluster prototypes of this unsupervised system can be reconstructed to illustrate the reasoning of the system. In this application, this analysis hints at the usefulness of previously unused portions of the VEP in detecting MS. It also indicates a possible use of the system as a training aide.

Adult↗
PubMed2006

Three-dimensional classification of spinal deformities using fuzzy clustering.

STUDY DESIGN: A prospective study of a large set of three-dimensional (3D) reconstructions of spinal deformities in adolescent idiopathic scoliosis (AIS). OBJECTIVES: To determine the value of fuzzy clustering techniques to automatically detect clinically relevant 3D curve patterns within this set of 3D spine models. SUMMARY OF BACKGROUND DATA: Classification is important for the assessment of AIS and has been mainly used to guide surgical treatment. Current classification systems are based on visual curve pattern identification using two-dimensional radiologic measurements but remain controversial because of their low interobserver and intraobserver reliability. A clinically useful 3D classification remains to be found. METHODS: An unsupervised learning algorithm, fuzzy k-means clustering, was applied on 409 3D spine models. Analysis of data distribution using clinical parameters was performed by studying similar curve patterns, near each cluster center identified. RESULTS: The algorithm determined that the entire sample of models could be segmented in five easily differentiated curve patterns similar to those of the Lenke and King classifications. Furthermore, a system with 12 classes made possible the identification of subpatterns of spinal deformity with true 3D components. CONCLUSIONS: Automatic and clinically relevant 3D classification of AIS is possible using an unsupervised learning algorithm. This approach can now be used to build a relevant 3D classification of AIS using appropriate key features of 3D models selected by a panel of expert spinal deformity surgeons.

Adolescent↗
PubMed2003

An unsupervised hierarchical dynamic self-organizing approach to cancer class discovery and marker gene identification in microarray data.

MOTIVATION: Current Self-Organizing Maps (SOMs) approaches to gene expression pattern clustering require the user to predefine the number of clusters likely to be expected. Hierarchical clustering methods used in this area do not provide unique partitioning of data. We describe an unsupervised dynamic hierarchical self-organizing approach, which suggests an appropriate number of clusters, to perform class discovery and marker gene identification in microarray data. In the process of class discovery, the proposed algorithm identifies corresponding sets of predictor genes that best distinguish one class from other classes. The approach integrates merits of hierarchical clustering with robustness against noise known from self-organizing approaches. RESULTS: The proposed algorithm applied to DNA microarray data sets of two types of cancers has demonstrated its ability to produce the most suitable number of clusters. Further, the corresponding marker genes identified through the unsupervised algorithm also have a strong biological relationship to the specific cancer class. The algorithm tested on leukemia microarray data, which contains three leukemia types, was able to determine three major and one minor cluster. Prediction models built for the four clusters indicate that the prediction strength for the smaller cluster is generally low, therefore labelled as uncertain cluster. Further analysis shows that the uncertain cluster can be subdivided further, and the subdivisions are related to two of the original clusters. Another test performed using colon cancer microarray data has automatically derived two clusters, which is consistent with the number of classes in data (cancerous and normal). AVAILABILITY: JAVA software of dynamic SOM tree algorithm is available upon request for academic use. SUPPLEMENTARY INFORMATION: A comparison of rectangular and hexagonal topologies for GSOM is available from http://www.mame.mu.oz.au/mechatronics/journalinfo/Hsu2003supp.pdf

Algorithms↗
PubMed2004

Evaluation of radiological features for breast tumour classification in clinical screening with machine learning methods.

OBJECTIVE: In this work, methods utilizing supervised and unsupervised machine learning are applied to analyze radiologically derived morphological and calculated kinetic tumour features. The features are extracted from dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) time-course data. MATERIAL: The DCE-MRI data of the female breast are obtained within the UK Multicenter Breast Screening Study. The group of patients imaged in this study is selected on the basis of an increased genetic risk for developing breast cancer. METHODS: The k-means clustering and self-organizing maps (SOM) are applied to analyze the signal structure in terms of visualization. We employ k-nearest neighbor classifiers (k-nn), support vector machines (SVM) and decision trees (DT) to classify features using a computer aided diagnosis (CAD) approach. RESULTS: Regarding the unsupervised techniques, clustering according to features indicating benign and malignant characteristics is observed to a limited extend. The supervised approaches classified the data with 74% accuracy (DT) and providing an area under the receiver-operator-characteristics (ROC) curve (AUC) of 0.88 (SVM). CONCLUSION: It was found that contour and wash-out type (WOT) features determined by the radiologists lead to the best SVM classification results. Although a fast signal uptake in early time-point measurements is an important feature for malignant/benign classification of tumours, our results indicate that the wash-out characteristics might be considered as important.

Artificial Intelligence↗
PubMed2004

Molecular profiling and classification of sporadic renal cell carcinoma by quantitative methylation analysis.

PURPOSE: Preoperative histologic classification of solid renal masses remains limited with current technology. We determine the utility of molecular profiling based on quantitative methylation analysis for characterization of sporadic renal cell carcinoma. EXPERIMENTAL DESIGN: Primary renal cell carcinomas representing three different histologic subtypes were obtained from a total of 38 patients who underwent radical nephrectomy for suspected malignant disease. Genomic DNA was isolated from tumors and was subjected to sodium bisulfite modification. The normalized index of methylation (NIM) for each sample was determined by quantitative real-time methylation-specific PCR at 17 different gene promoters. Hierarchical cluster analysis was performed by using an unsupervised neural network with binary tree topology. RESULTS: The majority of gene promoters that were analyzed in this study demonstrated very low levels of methylation (NIM <1.0). The RASSF1A gene promoter, however, was methylated in 30 of 38 (79%) cases. The frequency of RASSF1A methylation in papillary, clear-cell, and oncocytoma subtypes was 100, 90, and 25%, respectively. The highest levels of RASSF1A methylation were observed in the papillary (mean NIM = 78.9) and clear-cell (mean NIM = 13.4) subtypes. The vast majority of oncocytomas were completely unmethylated, and none demonstrated >1% methylation (mean NIM = 0.11). Hierarchical cluster analysis based on quantitative methylation levels resulted in stratification of sporadic renal cell carcinomas into their discrete histologic subtypes. CONCLUSIONS: Classification of sporadic renal cell carcinomas into histologic subtypes can be accomplished via multigene quantitative methylation profiling. Validation of this approach and selection of appropriate methylation markers may ultimately lead to use of this technology in the preoperative assessment of suspicious renal masses.

Aged↗
PubMed2003

Multiresolution fuzzy clustering of functional MRI data.

Recent developments in the analysis of functional MRI data reveal a shift from hypothesis-driven statistical tests to unsupervised strategies. One of the most promising approaches is the fuzzy clustering algorithm (FCA), whose potential to detect activation patterns has already been demonstrated. But the FCA suffers from three drawbacks: first the computational complexity, second the higher sensitivity to noise and third the dependence on the random initialization. With the multiresolution approach presented here, these weak points are significantly improved, as is demonstrated in our tests with simulated and real functional MRI data.

Algorithms↗
PubMed2005

Noise-robust soft clustering of gene expression time-course data.

Clustering is an important tool in microarray data analysis. This unsupervised learning technique is commonly used to reveal structures hidden in large gene expression data sets. The vast majority of clustering algorithms applied so far produce hard partitions of the data, i.e. each gene is assigned exactly to one cluster. Hard clustering is favourable if clusters are well separated. However, this is generally not the case for microarray time-course data, where gene clusters frequently overlap. Additionally, hard clustering algorithms are often highly sensitive to noise. To overcome the limitations of hard clustering, we applied soft clustering which offers several advantages for researchers. First, it generates accessible internal cluster structures, i.e. it indicates how well corresponding clusters represent genes. This can be used for the more targeted search for regulatory elements. Second, the overall relation between clusters, and thus a global clustering structure, can be defined. Additionally, soft clustering is more noise robust and a priori pre-filtering of genes can be avoided. This prevents the exclusion of biologically relevant genes from the data analysis. Soft clustering was implemented here using the fuzzy c-means algorithm. Procedures to find optimal clustering parameters were developed. A software package for soft clustering has been developed based on the open-source statistical language R. The package called Mfuzz is freely available.

Algorithms↗