Accelerating Genome Annotation Pipelines with GPU-Accelerated Machine Learning

EasyChair Preprint 14128

Abstract

Genome annotation is a fundamental step in genomics that involves identifying and labeling functional elements within a genome. Traditional genome annotation pipelines are often constrained by computational limitations, resulting in lengthy processing times and suboptimal scalability. This paper presents an innovative approach to accelerating genome annotation pipelines using GPU-accelerated machine learning techniques. By harnessing the parallel processing power of GPUs, we enhance the efficiency and speed of key annotation tasks, including gene prediction, functional annotation, and sequence alignment. We propose a GPU-accelerated framework that integrates deep learning models, such as convolutional neural networks and transformers, to improve accuracy and processing speed. Our results demonstrate a significant reduction in computational time and an increase in annotation accuracy compared to conventional CPU-based methods. This advancement not only expedites genome annotation but also enables the analysis of larger and more complex genomic datasets, facilitating breakthroughs in genomics research and personalized medicine. The integration of GPU-accelerated machine learning into genome annotation pipelines represents a transformative step forward, offering a scalable and efficient solution to meet the growing demands of genomic research.

Keyphrases: Central Processing Units (CPU), Convolutional Neural Networks (CNNs), Graphics Processing Units (GPUs)

@booklet{EasyChair:14128,
  author    = {Abi Litty},
  title     = {Accelerating Genome Annotation Pipelines with GPU-Accelerated Machine Learning},
  howpublished = {EasyChair Preprint 14128},
  year      = {EasyChair, 2024}}