gpuZoo: Cost-effective estimation of gene regulatory networks using the Graphics Processing Unit

Abstract

Gene regulatory network inference allows for the modeling of genome-scale regulatory processes that are altered during development, in disease, and in response to perturbations. Our group has developed a collection of tools to model various regulatory processes, including transcriptional (PANDA, SPIDER) and post-transcriptional (PUMA) gene regulation, as well as gene regulation in individual samples (LIONESS). These methods work by postulating a network structure and then optimizing that structure to be consistent with multiple lines of biological evidence through repeated operations on data matrices. Although our methods are widely used, the corresponding computational complexity, and the associated costs and run times, do limit some applications. To improve the cost/time performance of these algorithms, we developed gpuZoo which implements GPU-accelerated calculations, dramatically improving the performance of these algorithms. The runtime of the gpuZoo implementation in MATLAB and Python is up to 61 times faster and 28 times less expensive than multi-core CPU implementation of the same methods. gpuZoo is available in MATLAB through the netZooM package https://github.com/netZoo/netZooM and in Python through the netZooPy package https://github.com/netZoo/netZooPy.

The Network Zoo: a multilingual package for the inference and analysis of biological networks

“…Using a synchronized resource for code development avoids creating parallel branches and gives users access to tested and optimized tools that are up to date with the newest frameworks, particularly for the growing userbase in R and Python, as well as with third-party dependencies. The codebase includes additional helper functions for plotting and analysis, and GPU-accelerated implementations [67] for faster network inference across large numbers of samples. A continuous integration system called ZooKeeper runs unit tests using GitHub actions and a custom server to maintain the integrity of the software and update dependencies to third-party software.…”

The Network Zoo: a multilingual package for the inference and analysis of gene regulatory networks

“…Co-developing methods in various languages while using the same unit tests across them has helped identify inconsistencies for some edge cases and has dramatically improved reproducibility. netZoo implementations were also optimized for runtime and memory usage which included using GPU [41], and wrapping faster implementations to be used in other languages.…”

“…Using a synchronized resource for code development avoids creating parallel branches and gives users access to tested and optimized methods that are up to date with the newest frameworks, particularly for the growing userbase in R and Python, as well as with third-party dependencies. The codebase includes additional helper functions for plotting and analysis, and GPU-accelerated implementations [41] for faster network inference across large numbers of samples. The netZoo codebase is part of a larger ecosystem of online tools, that together support reproducible science.…”