Multiobjective Simulated Annealing-Based Clustering of Tissue Samples for Cancer Diagnosis

Abstract

In the field of pattern recognition, the study of the gene expression profiles of different tissue samples over different experimental conditions has become feasible with the arrival of microarray-based technology. In cancer research, classification of tissue samples is necessary for cancer diagnosis, which can be done with the help of microarray technology. In this paper, we have presented a multiobjective optimization (MOO)-based clustering technique utilizing archived multiobjective simulated annealing(AMOSA) as the underlying optimization strategy for classification of tissue samples from cancer datasets. The presented clustering technique is evaluated for three open source benchmark cancer datasets [Brain tumor dataset, Adult Malignancy, and Small Round Blood Cell Tumors (SRBCT)]. In order to evaluate the quality or goodness of produced clusters, two cluster quality measures viz, adjusted rand index and classification accuracy ( % CoA) are calculated. Comparative results of the presented clustering algorithm with ten state-of-the-art existing clustering techniques are shown for three benchmark datasets. Also, we have conducted a statistical significance test called t-test to prove the superiority of our presented MOO-based clustering technique over other clustering techniques. Moreover, significant gene markers have been identified and demonstrated visually from the clustering solutions obtained. In the field of cancer subtype prediction, this study can have important impact.

Multi-objective semi-supervised clustering of tissue samples for cancer diagnosis

“…Our proposed semisupervised clustering mechanism is applied on three open source cancer data sets, namely, brain tumor, adult malignancy and SRBCT. Our proposed technique is also compared with some state-of-the-art clustering algorithms including two recently developed MOO-based techniques (Mukhopadhyay et al 2010;Acharya et al 2015) with respect to two clustering performance metrics, namely, ARI (Mukhopadhyay et al 2010) and Percentage Classification Accuracy (CoA%) (Mukhopadhyay et al 2010). Experimental results support our assumptions that the proposed semi-supervised clustering technique will perform much better than the existing clustering algorithms.…”

“…From Table 1, we can conclude that our proposed semi-supervised clustering technique based on the search capabilities of AMOSA performs much better than all of existing clustering algorithms provided in table for clustering tissue samples of adult malignancy, brain tumor and SRBCT data sets in terms of ARI and %Co A. Results also prove that the proposed technique performs much better than two recently developed multi-objective based clustering techniques, AMOSA based unsupervised clustering technique (Acharya et al 2015) and MOGASVM clustering technique. In (Acharya et al 2015), a multi-objective based clustering technique is developed for cancer tissue sample classification using the search capability of AMOSA.…”

“…The performance of the proposed semi-supervised clustering technique using the search capability of AMOSA is compared with other state-of-the-art clustering algorithms like UnAMOSA (unsupervised clustering using the search capability of AMOSA) (Acharya et al 2015), MOGASVM (Mukhopadhyay et al 2010), Expectation Maximization Clustering (EM) (Jain et al 1999), K-means clustering (Jain and Dubes 1988), hierarchical average linkage clustering (Jain and Dubes 1988), SiMM-TS clustering (Bandyopadhyay et al 2007), Self Organizing Map (SOM) clustering ) and consensus clustering (Strehl and Ghosh 2003). Consensus clustering contains three approaches to ensemble cluster, which are Cluster based Similarity Partitioning Algorithm (CSPA), Meta-Clustering Algorithm (MCLA) and HyperGraph Partitioning Algorithm (HGPA).…”

“…With the development of micro-array technology, it is feasible to study the gene expression profiles of different tissue samples over different experimental conditions Alizadeh et al 2000;Acharya et al 2015). The automatic classification of tissue samples into different classes like benign (non-cancerous) cells or malignant (cancerous) cells has important impact on cancer research.…”

“…It involves the training and testing time of SVM which would increase with the increase in sample size. In order to get rid of this time complexity, in (Acharya et al 2015) authors have proposed another multi-objective based solution framework for classification of cancer tissue samples. Here a simulated annealing based multi-objective algorithm is developed for cancer cell classification.…”