Biclustering of Expression Microarray Data Using Affinity Propagation

Abstract

Abstract. Biclustering, namely simultaneous clustering of genes and samples, represents a challenging and important research line in the expression microarray data analysis. In this paper, we investigate the use of Affinity Propagation, a popular clustering method, to perform biclustering. Specifically, we cast Affinity Propagation into the Couple Two Way Clustering scheme, which allows to use a clustering technique to perform biclustering. We extend the CTWC approach, adapting it to Affinity Propagation, by introducing a stability criterion and by devising an approach to automatically assemble couples of stable clusters into biclusters.Empirical results, obtained in a synthetic benchmark for biclustering, show that our approach is extremely competitive with respect to the state of the art, achieving an accuracy of 91% in the worst case performance and 100% accuracy for all tested noise levels in the best case.

A Binary Factor Graph Model for Biclustering

“…As we can see in Fig.2, increasing the noise completely corrupts the performances of the aggregation algorithm. Notice that, obtained results are competitive with other state of the art approaches (see figure 2 in [6], figure 1 in [18] or figure 3 in [9]), confirming the potentialities of the proposed approach.…”

“…In our model we used as similarity the negative of the Euclidean distance (as in [9]), which allows to retrieve only constant value biclusters. As in the original Affinity Propagation model, a proper setting of the preferences (namely the self similarities) is crucial: in our experiments we found that a good choice is represented by the first integer number below the median (which represents the standard setting [1]).…”

“…Many of such previous approaches are based on the idea of adapting a given clustering technique to the biclustering problem, for example by repeatedly performing experiments and genes clustering [8,9]. This paper follows the above-described research trend, and proposes a novel biclustering algorithm, which extends and adapts to the biclustering scenario the well known Affinity Propagation (AP) clustering algorithm [1].…”

“…Even if some variants of the AP approach have been applied to the microarray scenario -see for example [11,12] -its use in the biclustering context remains somehow unexplored, with few papers recently published (such as [9], and [13]). In particular, in [9] the AP model is used as the clustering module in an iterative rows and columns clustering scheme [8]: however no modifications to the basic AP model has been introduced, which is still used as a standard clustering method.…”

“…In particular, in [9] the AP model is used as the clustering module in an iterative rows and columns clustering scheme [8]: however no modifications to the basic AP model has been introduced, which is still used as a standard clustering method. In contrast, [13] proposes an exemplar-based strategy to find biclusters.…”