Nature Computational Science: Championing computational work across multiple disciplines

By: undefined, Wed Jun 2 2021

Our goal with Nature Computational Science has been to focus on the impressive computational work being developed across multiple disciplines: we have envisioned the journal being a dedicated home for this work, where researchers from different domains can learn from each other in order to advance science. Since we launched the journal in January 2021, we’ve published various examples that champion such a multidisciplinary aspect, some of them being the highest accessed articles in the journal.

Fernando Chirigati

For instance, the work by Chi Chen and colleagues is a great example of how the materials science field can benefit from machine learning: they developed graph neural networks that take advantage of data from varying levels of fidelity to accurately model ordered and disordered materials, an important research problem. Interestingly, their approach can be generalized to any problem that can be described using graph structures, showcasing its potential broad impact. In another highly-accessed manuscript, Peter Bauer and colleagues discuss how to best exploit computing technology advances in the climate sciences domain. These advances range from new algorithmic frameworks to new hardware architectures, requiring different sets of expertise to move weather and climate predictions forward. Thomas Löhr and colleagues advanced the understanding of Alzheimer’s disease by repurposing a method based on deep learning to effectively describe the structural ensemble of the peptide associated with the progression of the disease.

In addition to publishing high-impact research, we want Nature Computational Science to serve as a platform to discuss timely, ethical, and policy-related topics that are of interest to the broad community of computational science. For this reason, we have commissioned pieces from various experts on such topics. On the topic of COVID-19, Rosalind M. Eggo and colleagues discuss why there is no one-size-fits-all model, and which uncertainties must be fully understood when building these critical models that help inform public health officials about potential interventions. In another Comment, Rebecca Knowles and colleagues discuss a fundamental issue in computational science research: the lack of appropriate funding for software infrastructure, which deeply affects the maintenance of important software contributions that are essential to research projects in different fields.

We have heard multiple times from researchers that “it was about time for Nature Portfolio to have a journal dedicated to computational science,” and we couldn’t agree more. Looking ahead, we will continue to publish important advances in computational science, and strive to bring more diversity in topics and more representation from minority groups.

Click here to request more information for your organization.