Download PDFOpen PDF in browserAccelerating Structural Variant Detection with GPU and Machine LearningEasyChair Preprint 1412213 pages•Date: July 25, 2024AbstractStructural variants (SVs) are significant genomic alterations that play a crucial role in genetic diversity, evolution, and various diseases, including cancer. Traditional methods for detecting SVs often face challenges in terms of computational efficiency, accuracy, and scalability, particularly when dealing with large-scale genomic data. In recent years, the advent of Graphics Processing Units (GPUs) and machine learning (ML) has opened new avenues for addressing these challenges. This paper explores the integration of GPU acceleration and ML techniques to enhance the detection and analysis of structural variants. We present a comprehensive framework that leverages deep learning models, optimized for parallel processing on GPUs, to achieve real-time SV detection with high accuracy. Our approach not only reduces the computational burden but also improves the sensitivity and specificity of SV detection compared to conventional methods. Through extensive benchmarking on various genomic datasets, we demonstrate the superior performance of our GPU-accelerated ML framework in terms of speed, accuracy, and scalability. The findings underscore the potential of combining GPU and ML technologies to revolutionize genomic research and pave the way for more efficient and precise structural variant analysis in clinical and research settings. Keyphrases: Graphics Processing Units (GPUs), Machine Learning (ML), Structural variants (SVs)
|