Patterns (Jun 2021)
Quantum processor-inspired machine learning in the biomedical sciences
Abstract
Summary: Recent advances in high-throughput genomic technologies coupled with exponential increases in computer processing and memory have allowed us to interrogate the complex molecular underpinnings of human disease from a genome-wide perspective. While the deluge of genomic information is expected to increase, a bottleneck in conventional high-performance computing is rapidly approaching. Inspired by recent advances in physical quantum processors, we evaluated several unconventional machine-learning (ML) strategies on actual human tumor data, namely “Ising-type” methods, whose objective function is formulated identical to simulated annealing and quantum annealing. We show the efficacy of multiple Ising-type ML algorithms for classification of multi-omics human cancer data from The Cancer Genome Atlas, comparing these classifiers to a variety of standard ML methods. Our results indicate that Ising-type ML offers superior classification performance with smaller training datasets, thus providing compelling empirical evidence for the potential future application of unconventional computing approaches in the biomedical sciences. The bigger picture: Advances in sequencing technology, leading to an ever-increasing volume and variety of data, present an opportunity to probe the molecular underpinnings of disease. Inspired in part by recent developments in physical quantum processors, we evaluated several Ising-type algorithms, which are relatively unused in the biomedical sciences, on actual human tumor data from The Cancer Genome Atlas. Our results show performance competitive with conventional machine-learning algorithms in classifying human cancer types and associated molecular subtypes when training with all available data; but perhaps more strikingly, the Ising-type algorithms demonstrate superior performance with smaller training datasets. This gain in performance suggests a tantalizing application for rare diseases or other clinical applications where the number of training samples may be quite small. In addition, the features extracted from the Ising-type algorithms have biological relevance.
