Torch-eCpG: A fast and scalable eQTM mapper for thousands of molecular phenotypes with graphical processing units

Area of Science:

• Genomics
• Epigenetics
• Computational Biology

Background:

• Gene expression is influenced by DNA methylation in cis, distal, and trans regulatory regions.
• Expression quantitative trait methylation (eQTM) loci mapping, also known as expression associated CpG loci (eCpG), is crucial for understanding DNA methylation's role in gene expression.
• Existing methods lack open-source tools and face computational challenges due to the large number of comparisons required for eQTM mapping.

Approach:

• Introduced Torch-eCpG, an open-source software tool designed for efficient eQTM mapping.
• Implemented an optimized computational approach utilizing graphical processing units (GPUs) to significantly reduce analysis runtime.
• Ensured the tool's scalability to handle large datasets with numerous methylation loci.

Key Points:

• Torch-eCpG provides reproducible eQTM mapping analyses.
• GPU acceleration with Torch-eCpG achieves up to an 18x speed increase compared to traditional methods.
• The tool demonstrates linear scalability with an increasing number of methylation loci.

Conclusions:

• Torch-eCpG offers a fast, reliable, and scalable solution for eQTM mapping.
• The availability of Torch-eCpG as open-source software facilitates broader research in epigenetics and gene regulation.
• Source code is accessible at https://github.com/kordk/torch-ecpg.