An Independent Implementation of Quantum Machine Learning Algorithms in Qiskit for Genomic Data
Navneet Singh, Shiva Raj Pokhrel
TL;DR
The study addresses applying quantum machine learning to genomic sequence classification. It implements QSVC, Pegasos-QSVC, VQC, and QNN in Qiskit, using feature maps $ZFeatureMap$, $ZZFeatureMap$, and $PauliFeatureMap$ to encode data into quantum states, enabling a quantum kernel with entries $K_{ij} = \langle \psi(\vec{x}_i) | \psi(\vec{x}_j) \rangle$ and a decision function $f(\vec{x})$. A dataset of 100k genome sequences is preprocessed via text vectorization and PCA, then evaluated across training/test splits, with convergence improvements observed for QNN and VQC across feature-map schemes. Quantitative metrics such as accuracy, precision, recall, F1-score, and AUROC show improvements over prior works (e.g., Gentinetta2024, Abbas2021, Pokhrel2024), and the work provides open-source QML implementations in Qiskit and demonstrates potential for high-dimensional genomic data.
Abstract
In this paper, we explore the power of Quantum Machine Learning as we extend, implement and evaluate algorithms like Quantum Support Vector Classifier (QSVC), Pegasos-QSVC, Variational Quantum Circuits (VQC), and Quantum Neural Networks (QNN) in Qiskit with diverse feature mapping techniques for genomic sequence classification.