DeepTRIAGE: interpretable and individualised biomarker scores using attention mechanism for the classification of breast cancer sub-types

Abstract

Background: Breast cancer is a collection of multiple tissue pathologies, each with a distinct molecular signature that correlates with patient prognosis and response to therapy. Accurately differentiating between breast cancer sub-types is an important part of clinical decision-making. Although this problem has been addressed using machine learning methods in the past, there remains unexplained heterogeneity within the established sub-types that cannot be resolved by the commonly used classification algorithms. Methods: In this paper, we propose a novel deep learning architecture, called DeepTRIAGE (Deep learning for the TRactable Individualised Analysis of Gene Expression), which uses an attention mechanism to obtain personalised biomarker scores that describe how important each gene is in predicting the cancer sub-type for each sample. We then perform a principal component analysis of these biomarker scores to visualise the sample heterogeneity, and use a linear model to test whether the major principal axes associate with known clinical phenotypes. Results: Our model not only classifies cancer sub-types with good accuracy, but simultaneously assigns each patient their own set of interpretable and individualised biomarker scores. These personalised scores describe how important each feature is in the classification of any patient, and can be analysed post-hoc to generate new hypotheses about latent heterogeneity. Conclusions: We apply the DeepTRIAGE framework to classify the gene expression signatures of luminal A and luminal B breast cancer sub-types, and illustrate its use for genes as well as the GO and KEGG gene sets. Using DeepTRIAGE, we calculate personalised biomarker scores that describe the most important features for classifying an individual patient as luminal A or luminal B. In doing so, DeepTRIAGE simultaneously reveals heterogeneity within the luminal A biomarker scores that significantly associate with tumour stage, placing all luminal samples along a continuum of severity.

BrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis

“…Intuitively, interpretation is trying to understand how a model makes a right decision rather than a wrong one when learning from a good teacher. We take the model with the highest accuracy for the best predictive performance [8,6]. However, in some other applications, such as understanding the general pattern or mechanism associated with a cognitive task or disease, group-level biomarkers which highlight consistent explanations across individuals are important [3,56,50].…”

A Field Guide to Scientific XAI: Transparent and Interpretable Deep Learning for Bioinformatics Research

“…• Domain knowledge-driven engineering makes use of prior domain knowledge to synthesize new features. For example, one could convert gene expression signatures into a functional pathway score by adding up the expression levels of all genes belonging to each pathway [13].…”

A Neural Model of RNA Splicing: Learning Motif Distances with Self-Attention and Toeplitz Max Pooling

“…In contrast, [Bahdanau et al, 2016] uses the attention mechanism to allow a model to automatically search for the important parts of an input. Attention, being related to the concept of self-explanation [Alvarez-Melis and Jaakkola, 2018b], can assign saliency from within the model, and has been previously used in genomics research to produce interpetable sample-specific importance scores [Beykikhoshk et al, 2020]. Attention is now an important part of transformer architectures [Vaswani et al, 2017], a light-weight alternative to CNNs for big data like whole genomes [Clauwaert and Waegeman, 2020].…”