Modeling Quantum Machine Learning for Genomic Data Analysis

Genomic data analysis is essential for personalized medicine, disease diagnostics, and understanding evolutionary biology, yet remains challenging due to high dimensionality and sequence complexity. Quantum machine learning (QML) offers promising computational advantages for scalable genomic analysis. We explore QML modeling for genomic sequence classification using 100 000 sequences from two transcript classes. Our new framework incorporates PCA dimensionality reduction and evaluates multiple quantum feature maps (ZFeatureMap, ZZFeatureMap, PauliFeatureMap). Results show Pegasos-QSVC achieving 99.12% recall and 67.41% F1 score, while QNNs reach 54.38% training accuracy. Performance varies across feature maps, with some models exhibiting overfitting.