Dataset Information

Methods for evaluating clustering algorithms for gene expression data using a reference set of functional classes.

ABSTRACT:

Background

A cluster analysis is the most commonly performed procedure (often regarded as a first step) on a set of gene expression profiles. In most cases, a post hoc analysis is done to see if the genes in the same clusters can be functionally correlated. While past successes of such analyses have often been reported in a number of microarray studies (most of which used the standard hierarchical clustering, UPGMA, with one minus the Pearson's correlation coefficient as a measure of dissimilarity), often times such groupings could be misleading. More importantly, a systematic evaluation of the entire set of clusters produced by such unsupervised procedures is necessary since they also contain genes that are seemingly unrelated or may have more than one common function. Here we quantify the performance of a given unsupervised clustering algorithm applied to a given microarray study in terms of its ability to produce biologically meaningful clusters using a reference set of functional classes. Such a reference set may come from prior biological knowledge specific to a microarray study or may be formed using the growing databases of gene ontologies (GO) for the annotated genes of the relevant species.

Results

In this paper, we introduce two performance measures for evaluating the results of a clustering algorithm in its ability to produce biologically meaningful clusters. The first measure is a biological homogeneity index (BHI). As the name suggests, it is a measure of how biologically homogeneous the clusters are. This can be used to quantify the performance of a given clustering algorithm such as UPGMA in grouping genes for a particular data set and also for comparing the performance of a number of competing clustering algorithms applied to the same data set. The second performance measure is called a biological stability index (BSI). For a given clustering algorithm and an expression data set, it measures the consistency of the clustering algorithm's ability to produce biologically meaningful clusters when applied repeatedly to similar data sets. A good clustering algorithm should have high BHI and moderate to high BSI. We evaluated the performance of ten well known clustering algorithms on two gene expression data sets and identified the optimal algorithm in each case. The first data set deals with SAGE profiles of differentially expressed tags between normal and ductal carcinoma in situ samples of breast cancer patients. The second data set contains the expression profiles over time of positively expressed genes (ORF's) during sporulation of budding yeast. Two separate choices of the functional classes were used for this data set and the results were compared for consistency.

Conclusion

Functional information of annotated genes available from various GO databases mined using ontology tools can be used to systematically judge the results of an unsupervised clustering algorithm as applied to a gene expression data set in clustering genes. This information could be used to select the right algorithm from a class of clustering algorithms for the given data set.

SUBMITTER: Datta S

PROVIDER: S-EPMC1590054 | biostudies-literature | 2006 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Methods for evaluating clustering algorithms for gene expression data using a reference set of functional classes.

Datta Susmita S Datta Somnath S

BMC bioinformatics 20060831

<h4>Background</h4>A cluster analysis is the most commonly performed procedure (often regarded as a first step) on a set of gene expression profiles. In most cases, a post hoc analysis is done to see if the genes in the same clusters can be functionally correlated. While past successes of such analyses have often been reported in a number of microarray studies (most of which used the standard hierarchical clustering, UPGMA, with one minus the Pearson's correlation coefficient as a measure of dis ...[more]

PMID: 16945146

Similar Datasets

Project description:This article contains data related to the research article 'Application of bi-clustering of gene expression data and gene set enrichment analysis methods to identify potentially disease causing nanomaterials' (Williams and Halappanavar, 2015) [1]. The presence of diverse types of nanomaterials (NMs) in commerce has grown significantly in the past decade and as a result, human exposure to these materials in the environment is inevitable. The traditional toxicity testing approaches that are reliant on animals are both time- and cost- intensive; employing which, it is not possible to complete the challenging task of safety assessment of NMs currently on the market in a timely manner. Thus, there is an urgent need for comprehensive understanding of the biological behavior of NMs, and efficient toxicity screening tools that will enable the development of predictive toxicology paradigms suited to rapidly assessing the human health impacts of exposure to NMs. In an effort to predict the long term health impacts of acute exposure to NMs, in Williams and Halappanavar (2015) [1], we applied bi-clustering and gene set enrichment analysis methods to derive essential features of altered lung transcriptome following exposure to NMs that are associated with lung-specific diseases. Several datasets from public microarray repositories describing pulmonary diseases in mouse models following exposure to a variety of substances were examined and functionally related bi-clusters showing similar gene expression profiles were identified. The identified bi-clusters were then used to conduct a gene set enrichment analysis on lung gene expression profiles derived from mice exposed to nano-titanium dioxide, carbon black or carbon nanotubes (nano-TiO2, CB and CNTs) to determine the disease significance of these data-driven gene sets. The results of the analysis correctly identified all NMs to be inflammogenic, and only CB and CNTs as potentially fibrogenic. Here, we elaborate on the details of the statistical methods and algorithms used to derive the disease relevant gene signatures. These details will enable other investigators to use the gene signature in future Gene Set Enrichment Analysis studies involving NMs or as features for clustering and classifying NMs of diverse properties.

Project description:BackgroundDNA microarrays, which determine the expression levels of tens of thousands of genes from a sample, are an important research tool. However, the volume of data they produce can be an obstacle to interpretation of the results. Clustering the genes on the basis of similarity of their expression profiles can simplify the data, and potentially provides an important source of biological inference, but these methods have not been tested systematically on datasets from complex human tissues. In this paper, four clustering methods, CRC, k-means, ISA and memISA, are used upon three brain expression datasets. The results are compared on speed, gene coverage and GO enrichment. The effects of combining the clusters produced by each method are also assessed.Resultsk-means outperforms the other methods, with 100% gene coverage and GO enrichments only slightly exceeded by memISA and ISA. Those two methods produce greater GO enrichments on the datasets used, but at the cost of much lower gene coverage, fewer clusters produced, and speed. The clusters they find are largely different to those produced by k-means. Combining clusters produced by k-means and memISA or ISA leads to increased GO enrichment and number of clusters produced (compared to k-means alone), without negatively impacting gene coverage. memISA can also find potentially disease-related clusters. In two independent dorsolateral prefrontal cortex datasets, it finds three overlapping clusters that are either enriched for genes associated with schizophrenia, genes differentially expressed in schizophrenia, or both. Two of these clusters are enriched for genes of the MAP kinase pathway, suggesting a possible role for this pathway in the aetiology of schizophrenia.ConclusionConsidered alone, k-means clustering is the most effective of the four methods on typical microarray brain expression datasets. However, memISA and ISA can add extra high-quality clusters to the set produced by k-means, so combining these three methods is the method of choice.

Project description:BackgroundmicroRNA (miRNA) is a short RNA (~ 22 nt) that regulates gene expression at the posttranscriptional level. Aberration of miRNA expressions could affect their targeting mRNAs involved in cancer-related signaling pathways. We conduct clustering analysis of miRNA and mRNA using expression data from the Cancer Genome Atlas (TCGA). We combine the Hungarian algorithm and blossom algorithm in graph theory. Data analysis is done using programming language R and Python.MethodsWe first quantify edge-weights of the miRNA-mRNA pairs by combining their expression correlation coefficient in tumor (T_CC) and correlation coefficient in normal (N_CC). We thereby introduce a bipartite graph partition procedure to identify cluster candidates. Specifically, we propose six weight formulas to quantify the change of miRNA-mRNA expression T_CC relative to N_CC, and apply the traditional hierarchical clustering to subjectively evaluate the different weight formulas of miRNA-mRNA pairs. Among these six different weight formulas, we choose the optimal one, which we define as the integrated mean value weights, to represent the connections between miRNA and mRNAs. Then the Hungarian algorithm and the blossom algorithm are employed on the miRNA-mRNA bipartite graph to passively determine the clusters. The combination of Hungarian and the blossom algorithms is dubbed maximum weighted merger method (MWMM).ResultsMWMM identifies clusters of different sizes that meet the mathematical criterion that internal connections inside a cluster are relatively denser than external connections outside the cluster and biological criterion that the intra-cluster Gene Ontology (GO) term similarities are larger than the inter-cluster GO term similarities. MWMM is developed using breast invasive carcinoma (BRCA) as training data set, but can also applies to other cancer type data sets. MWMM shows advantage in GO term similarity in most cancer types, when compared to other algorithms.ConclusionsmiRNAs and mRNAs that are likely to be affected by common underlying causal factors in cancer can be clustered by MWMM approach and potentially be used as candidate biomarkers for different cancer types and provide clues for targets of precision medicine in cancer treatment.

Dataset Information

Methods for evaluating clustering algorithms for gene expression data using a reference set of functional classes.

Background

Results

Conclusion

Publications

Methods for evaluating clustering algorithms for gene expression data using a reference set of functional classes.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets