gCUP: rapid GPU-based HIV-1 co-receptor usage prediction for next-generation sequencing.

Area of Science:

• Computational biology
• Bioinformatics
• Medical diagnostics

Background:

• Next-generation sequencing (NGS) offers significant potential for HIV diagnostics.
• Existing genotypic prediction models for HIV tropism are accurate but computationally inefficient.
• Clinical application of NGS in HIV diagnostics is hindered by computational limitations.

Purpose of the Study:

• To enhance the computational efficiency of HIV tropism prediction models.
• To leverage graphics processing units (GPUs) for faster sequence analysis.
• To improve the clinical utility of NGS in HIV diagnostics.

Main Methods:

• Development of a novel computational model (gCUP) optimized for GPU parallelization.
• Implementation of gCUP using NVIDIA GeForce GTX 460 for performance testing.
• Evaluation of the model's accuracy and classification speed.

Main Results:

• The gCUP model demonstrates high accuracy in HIV tropism prediction.
• gCUP achieves a classification speed exceeding 175,000 sequences per second.
• The GPU-accelerated approach significantly improves computational efficiency.

Conclusions:

• GPU acceleration is a key step towards realizing the clinical significance of NGS in HIV diagnostics.
• The gCUP model's efficiency enables faster and more scalable HIV tropism prediction.
• The developed approach is adaptable for other applications, such as HIV drug resistance prediction.