Astuti Aninda, Shih Pin-Keng, Lee Shan-Chih, Mekala Venugopala Reddy, Wijaya Ezra B, Ng Ka-Lok
Department of Bioinformatics and Medical Engineering, Asia University, Taichung, Taiwan.
School of Medicine, China Medical University, Taichung, Taiwan.
PLoS One. 2025 Jul 17;20(7):e0327928. doi: 10.1371/journal.pone.0327928. eCollection 2025.
Quantum machine learning (QML) combines quantum computing with machine learning, offering potential for solving intricate problems. Our research delves into QML's application in identifying gene expression biomarkers for clear cell renal cell carcinoma (ccRCC) metastasis. ccRCC, the primary renal cancer subtype, poses significant challenges due to its high lethality and complex metastasis process. Despite extensive research, understanding the mechanisms of cancer cell dissemination and establishment in distant sites remains elusive. Identifying metastasis biomarkers is a daunting task in machine learning. Our study addresses the need for improved execution time and accuracy in QSVC and QNN algorithms compared to SVC and NN for binary classification. Drawing inspiration from the Neural Quantum Embedding (NQE) method, we propose a two-stage approach for the binary classification problem. We aim to assess if integrating NQE with QSVC/QNN enhances performance compared to NQE with SVC/NN across diverse biomedical datasets, demonstrating the effectiveness and generalizability of the approach.
量子机器学习（QML）将量子计算与机器学习相结合，为解决复杂问题提供了潜力。我们的研究深入探讨了QML在识别透明细胞肾细胞癌（ccRCC）转移的基因表达生物标志物方面的应用。ccRCC是主要的肾癌亚型，因其高致死率和复杂的转移过程而带来重大挑战。尽管进行了广泛研究，但了解癌细胞在远处部位扩散和定植的机制仍然难以捉摸。在机器学习中识别转移生物标志物是一项艰巨的任务。我们的研究解决了与用于二元分类的支持向量机（SVC）和神经网络（NN）相比，量子支持向量分类器（QSVC）和量子神经网络（QNN）算法在执行时间和准确性方面需要改进的问题。从神经量子嵌入（NQE）方法中获得灵感，我们针对二元分类问题提出了一种两阶段方法。我们旨在评估与将NQE与SVC/NN相比，将NQE与QSVC/QNN集成是否能在各种生物医学数据集中提高性能，从而证明该方法的有效性和通用性。
