A New Ensemble Method for Detecting Anomalies in Gene Expression Matrices

Laura Selicato,Attilio Guarini,Maria Carmela Vegliante,Nicoletta Del Buono,Gian Maria Zaccaria,Flavia Esposito,Giuseppina Opinto,Grazia Gargano,Sabino Ciavarella

doi:10.3390/math9080882

Abstract

One of the main problems in the analysis of real data is often related to the presence of anomalies. Namely, anomalous cases can both spoil the resulting analysis and contain valuable information at the same time. In both cases, the ability to detect these occurrences is very important. In the biomedical field, a correct identification of outliers could allow the development of new biological hypotheses that are not considered when looking at experimental biological data. In this work, we address the problem of detecting outliers in gene expression data, focusing on microarray analysis. We propose an ensemble approach for detecting anomalies in gene expression matrices based on the use of Hierarchical Clustering and Robust Principal Component Analysis, which allows us to derive a novel pseudo-mathematical classification of anomalies.

Highlights

Real datasets often contain observations that behave differently from the majority of the data
We address the problem of detecting outliers in Gene Expression Profiling (GEP) data, focusing on microarray data containing gene expression values for a given number of samples labeled with a biological class
We present a new ensemble approach to anomaly detection that combines

Summary

Introduction

Real datasets often contain observations that behave differently from the majority of the data. If an occurrence differs from the dominant part of the data, or if it is sufficiently unlikely under the assumed data probability model, it is considered an anomaly or outlier. Anomalies may adversely affect the conclusions drawn from data analysis; on the other hand, they may contain important information. Robust statistics are designed to detect outliers by first fitting the majority of the data and flagging data points that deviate [2]. The correct identification of outliers is of great importance: depending on the type of analysis to be performed, biologists can decide whether or not these data should be removed

Methods

Discussion

Conclusion