Search

Cancer is curable if it can be detected early. One way to detect cancer is by analyzing the change in expression of genes in the suspected tissue. Serial analysis of gene expression (SAGE) is a sequencing technique used for measuring the expression levels of genes. Cancer detection problem can be posed as binary classification problem like whether a tissue is cancerous or normal. SAGE libraries contain expression levels of thousands of genes which are the features. It is impossible to consider all these features for classification and also the general feature selection algorithms are not efficient with this data. In this paper, closed frequent itemset mining is proposed as a feature selection technique for identifying a small set of features which can distinguish the two classes efficiently. The performance of the proposed technique is evaluated on SAGE data related to breast tissue and a group of 26 genes are selected as best features. Two well known classifiers, extreme learning machine (ELM) and support vector machine (SVM), are used to evaluate the effectiveness of the selected features in classification and found that the proposed method works well with these classifiers. [ABSTRACT FROM AUTHOR]

Abstract: Mass spectrometry (MS)-based proteomics has been established as a standard way for biomarker discovery and early detection of disease from the proteome level. However, the data generated by MS technology are noisy, have a lot of redundancy, and most importantly have low sample-to-dimension ratio. Therefore, successful analysis of MS data heavily relies on the algorithmic and data mining techniques used to extract information. In this paper, we briefly discuss some of the commonly used algorithms applied to mass-to-charge (m/z) feature selection, and propose a k-means clustering based hybrid algorithm to address the high-dimension and high-correlation issues associated with this task. Our k-means clustering based hybrid algorithm incorporates the advantages of both filter- and wrapper-based feature selection algorithms. A special iterative procedure is introduced to overcome the instability of k-means clustering and genetic algorithm. By comparing the proposed hybrid system with several popular m/z feature selection algorithms using 10 different classifiers, we show that the proposed algorithm is able to select m/z features with lower correlation while achieving higher sample classification accuracies. The m/z feature selection results also indicate that the proposed hybrid algorithm is very stable despite its stochastic nature. [ABSTRACT FROM AUTHOR]