A study by Jason Grealey (Cambridge Baker Systems Initiative, Melbourne) and Loïc Lannelongue (Public Health and Primary Care, Cambridge Baker Systems Initiative), led by Michael Inouye (Public Health and Primary Care, Cambridge Baker Systems Initiative), has quantified the carbon footprint of widely used algorithms in bioinformatics, computational biology and statistical genetics. The study raises awareness, provides easy-to-use metrics, and makes recommendations for greener bioinformatics.
Bioinformatics relies on large-scale computational infrastructures, which have a non-zero carbon footprint; however, no study has quantified the environmental costs of bioinformatic tools and computing strategies. In this work, published in Molecular Biology and Evolution, the researchers used the freely available Green Algorithms calculator (www.green-algorithms.org) to estimate carbon footprints in kilograms of CO2 equivalent units, then contextualised this using distances travelled by car/plane and the corresponding number of months a mature tree would need to sequester a given amount of CO2.
The study assessed genome-wide association studies, RNA sequencing, genome assembly, metagenomics, phylogenetics and molecular simulations, as well as computation strategies, such as parallelisation, CPUs (central processing unit) vs GPUs (graphics processing unit), cloud vs local computing infrastructures and geographies.
Overall, this work provides a framework for green computing in bioinformatics and facilitates carbon reduction in health data science more broadly.
