Molecular Classification and Interpretation of Amyotrophic Lateral Sclerosis Using Deep Convolution Neural Networks and Shapley Values

Abstract

Amyotrophic lateral sclerosis (ALS) is a prototypical neurodegenerative disease characterized by progressive degeneration of motor neurons to severely effect the functionality to control voluntary muscle movement. Most of the non-additive genetic aberrations responsible for ALS make its molecular classification very challenging along with limited sample size, curse of dimensionality, class imbalance and noise in the data. Deep learning methods have been successful in many other related areas but have low minority class accuracy and suffer from the lack of explainability when used directly with RNA expression features for ALS molecular classification. In this paper, we propose a deep-learning-based molecular ALS classification and interpretation framework. Our framework is based on training a convolution neural network (CNN) on images obtained from converting RNA expression values into pixels based on DeepInsight similarity technique. Then, we employed Shapley additive explanations (SHAP) to extract pixels with higher relevance to ALS classifications. These pixels were mapped back to the genes which made them up. This enabled us to classify ALS samples with high accuracy for a minority class along with identifying genes that might be playing an important role in ALS molecular classifications. Taken together with RNA expression images classified with CNN, our preliminary analysis of the genes identified by SHAP interpretation demonstrate the value of utilizing Machine Learning to perform molecular classification of ALS and uncover disease-associated genes.

Explainable artificial intelligence for omics data: a systematic mapping study

“…For example, Karim et al . [ 98 ] utilize the DeepInsight framework to transform their gene expression data into images, which are then classified with a CNN, allowing for visual post-hoc explanations via SHAP pixel maps. The main advantage of CNNs and image input data is that they allow for a visual explanation, which is possibly the best comprehensible explanation technique to date.…”

Enhancing patient stratification and interpretability through class-contrastive and feature attribution techniques

“…SHapley Additive exPlanations (SHAP) [4] is another state-of-the-art explainability technique. Fast approximations of SHAP have been applied to analyse gene expression data [7,8,9,10,11], such as kernelExplainer, treeExplainer and gradientExplainer [12]. Yu et al use a deep autoencoder [9] to learn gene expression representations, applying treeExplainer SHAP to measure the contributions of genes to each of the latent variables.…”

Gene targeting in amyotrophic lateral sclerosis using causality-based feature selection and machine learning

“…RNA-sequencing data has been used for a wide range of analyses, including weighted gene co-expression network analysis and protein–protein interaction networks analysis (Kotni et al 2016 ; Saris et al 2009 ) as well as unsupervised clustering analysis of gene expression for the molecular classification of ALS samples (Aronica et al 2015 ; Morello et al 2018 ). In this context, a classification pipeline based predominantly on machine learning techniques was recently introduced (Karim et al 2021 ). However, the classification efficiency of machine learning (ML) methods deteriorates when applied to biomedical data; due to " the curse of dimensionality" (Vasilopoulou et al 2020 ), exacerbated by the lack of large, labeled datasets to properly train the ML models .…”

“…High dimensionality is a very common problem in such studies, since datasets include an extremely high number of features with a reduced number of samples (Vasilopoulou et al 2020 ). Recently, an indirect solution to the problem was proposed, based on conversion of RNA expression data into images and use of these images to train a convolution neural network (CNN) (Karim et al 2021 ). The main shortcoming of this approach is that the training data is too small to allow for efficient model building and separation between relevant and irrelevant genes.…”