Multi-level gene/MiRNA feature selection using deep belief nets and active learning

Abstract

Selecting the most discriminative genes/miRNAs has been raised as an important task in bioinformatics to enhance disease classifiers and to mitigate the dimensionality curse problem. Original feature selection methods choose genes/miRNAs based on their individual features regardless of how they perform together. Considering group features instead of individual ones provides a better view for selecting the most informative genes/miRNAs. Recently, deep learning has proven its ability in representing the data in multiple levels of abstraction, allowing for better discrimination between different classes. However, the idea of using deep learning for feature selection is not widely used in the bioinformatics field yet. In this paper, a novel multi-level feature selection approach named MLFS is proposed for selecting genes/miRNAs based on expression profiles. The approach is based on both deep and active learning. Moreover, an extension to use the technique for miRNAs is presented by considering the biological relation between miRNAs and genes. Experimental results show that the approach was able to outperform classical feature selection methods in hepatocellular carcinoma (HCC) by 9%, lung cancer by 6% and breast cancer by around 10% in F1-measure. Results also show the enhancement in F1-measure of our approach over recently related work in [1] and [2].

In silico Pathway Activation Network Decomposition Analysis (iPANDA) as a method for biomarker development

“…Thus, using the iPANDA signalling pathway activation scores as input for deep learning methods could bring results closer to experimental settings and make them more interpretable to bench biologists. One of the most difficult steps of multilayer perceptron training is the dimension reduction and feature selection procedures, which aim to generate the appropriate input for further learning44. Signalling pathway activation scoring using iPANDA will likely help reduce the dimensionality of expression data without losing biological relevance and may be used as an input to deep learning methods especially for drug discovery applications.…”

Machine Learning-based state-of-the-art methods for the classification of RNA-Seq data

“…Features from gene expression were extracted with regions of noncoding transcripts (miRNA) using DBNs and active learning. Here deep learning feature extractors were used to reduce the dimensionality of six cancer data sets and outperformed basic feature selection methods [71].…”

A scalable saliency-based feature selection method with instance-level information

“…There are only a few research works that address the problem of FS in Big Data environments. In [12,8] [13], the authors claimed to use a deep feature selection technique to reduce the input space in short-term wind forecasting models. However, their approach uses Recursive Feature Elimination (RFE) which requires to exponentially train several different models, making it unable to be used with a high number of features.…”