Deep learning for predicting disease status using genomic data

Abstract

Predicting disease status for a complex human disease using genomic data is an important, yet challenging, step in personalized medicine. Among many challenges, the so-called curse of dimensionality problem results in unsatisfied performances of many state-of-art machine learning algorithms. A major recent advance in machine learning is the rapid development of deep learning algorithms that can efficiently extract meaningful features from high-dimensional and complex datasets through a stacked and hierarchical learning process. Deep learning has shown breakthrough performance in several areas including image recognition, natural language processing, and speech recognition.However, the performance of deep learning in predicting disease status using genomic datasets is still not well studied. In this article, we performed a review on the four relevant articles that we found through our thorough literature review. All four articles used autoencoders to project high-dimensional genomic data to a low dimensional space and then applied the state-of-the-art machine learning algorithms to predict disease status based on the low-dimensional representations. This deep learning approach outperformed existing prediction approaches, such as prediction based on probe-wise screening and prediction based on principal component analysis. The limitations of the current deep learning approach and possible improvements were also discussed.PeerJ Preprints | https://doi.org/10.7287/peerj.preprints

Risk prediction for malignant intraductal papillary mucinous neoplasm of the pancreas: logistic regression versus machine learning

“…The six variables had a relatively simple structure. Recently, ML techniques have been utilised to develop disease prediction models with high-dimensional omics data, such as the genomics and transcriptomics data, and these approaches outperformed existing prediction methods 25,26 . If the genomics or transcriptomics data on IPMN can be included in the future model development with ML techniques, the performance may be increased.…”

Explainable Deep Learning Approach for Multilabel Classification of Antimicrobial Resistance With Missing Labels

“…As genomic sequence features are one-dimensional data, convolution layer and down sampling should be changed to one dimensional Deep neural networks are trained by estimating the optimal values of the biases and edge weights, minimising the difference between the true and predicted values of the labels. The function used to minimise this difference is termed the loss function [36], and the model's performance is estimated based on the loss value. Cross-entropy is a common choice of loss function for deep neural networks, which measures the difference between two probability values of true labels and predicted labels, as shown in Eq.…”

An energy-efficient in-memory computing architecture for survival data analysis based on resistive switching memories

“…In the last decade, medical researchers have started to extensively rely on machine learning (ML) and artificial neural networks (ANNs) to gain further insights into large amounts of complex and intertwined data (Anaya-Isaza et al, 2021 ; Allegra et al, 2022 ). Records concerning patients' clinical and genetic features, pathologies, interventions, hospitalizations, and follow ups are deeply investigated through survival analysis models, whose goal is to provide ad hoc treatment options and ultimately shed light on the origins of the disease (Wu et al, 2018 ). State-of-the-art data analysis platforms are built on Von Neumann computing architectures that devise bulky and power-hungry central processing units (CPUs), graphic processing units (GPUs), and memory devices embedded in high performance computing (HPC) machines (Bajaj and Ansari, 2021 ).…”

“…hospitalizations, and follow ups are deeply investigated through survival analysis models, whose goal is to provide ad hoc treatment options and ultimately shed light on the origins of the disease (Wu et al, 2018). State-of-the-art data analysis platforms are built on Von Neumann computing architectures that devise bulky and power-hungry central processing units (CPUs), graphic processing units (GPUs), and memory devices embedded in high performance computing (HPC) machines (Bajaj and Ansari, 2021).…”