Efficient Variant Calling in Genomics Using Machine Learning and GPU Acceleration

EasyChair Preprint no. 13940

Abstract

The advent of high-throughput sequencing technologies has revolutionized genomics by enabling the rapid and cost-effective generation of vast amounts of genetic data. A critical task in genomics is variant calling, the process of identifying genetic variants from sequencing data, which is essential for understanding genetic diversity and its implications for health and disease. Traditional variant calling methods, while accurate, often suffer from significant computational bottlenecks due to the massive scale of genomic datasets. This paper explores the integration of machine learning techniques and GPU acceleration to enhance the efficiency and accuracy of variant calling. By leveraging the parallel processing capabilities of GPUs, we aim to significantly reduce the computational time required for variant detection while maintaining high precision and recall rates. Our approach employs a deep learning-based model trained on annotated genomic datasets to predict variants, which is then optimized using GPU acceleration to handle large-scale data processing. Experimental results demonstrate that our method outperforms conventional CPU-based variant calling pipelines in both speed and accuracy, highlighting its potential for real-time genomic analysis in clinical and research settings. This study underscores the transformative impact of combining machine learning and GPU acceleration in genomics, paving the way for more efficient and scalable solutions in personalized medicine and genetic research.

Keyphrases: Accelerated sequence analysis, Bioinformatic algorithms, Computational genomics, Computational Proteomics, Deep learning in bioinformatics, Genomic data processing, GPU-accelerated machine learning, GPU-based bioinformatics, High Performance Computing, Machine learning in computational biology

@Booklet{EasyChair:13940,
  author = {Abi Cit},
  title = {Efficient Variant Calling in Genomics Using Machine Learning and GPU Acceleration},
  howpublished = {EasyChair Preprint no. 13940},

  year = {EasyChair, 2024}}