Algorithms for Bounded-Error Correlation of High Dimensional Data in Microarray Experiments

The problem of clustering continuous valued data has been well studied in literature. Its application to microarray analysis relies on such algorithms as -means, dimensionality reduction techniques (including SVD, principal components, KL transforms, etc.), and graph-based approaches for building dendrograms of sample data. In contrast, similar problems for discrete-attributed data are relatively unexplored. An instance of analysis of discrete-attributed data arises in detecting co-regulated samples in microarrays. Here, our interest is in detecting samples that are up- and down-regulated together. Efficient solutions to this problem enable more refined correlation between presence of sequence motifs and underlying regulation patterns in microarray data. One of the key challenges associated with clustering discrete-attributed data is that these problems typically are NP-hard and few effective heuristics are known for solving them. In this paper, we present an algorithm and a softwar...

• Type: Conference
• Year: 2003
• Where: CSB
• Authors: Mehmet Koyutürk, Ananth Grama, Wojciech Szpankowski
• Keywords: Continuous Valued Data, CSB 2003, Discrete-attributed Data, Microarray Data