Year: 2026
Type: preprint
Abstract: Computational methods are a growing component of every scientific discipline, but the pace of training and expertise for using these methods has not kept up with their widening usage. Researchers expect state-of-the-art methods to be able to independently explore their results, and do so dynamically and interactively. Producing such an application which implements a new algorithm, even though it would access a much wider audience, is typically beyond the expertise and time constraints of computational method developers. Here, we show our Sparrowhawk framework can rapidly create an interactive web application to run interactive analysis on a user's machine, across a wide range of algorithms. Focusing first on genomic epidemiology, a domain with a specific need for such methods, we created novel lightweight methods for genome assembly, antimicrobial resistance gene annotation, and taxonomic identification. Our results exceed performance of previous lightweight methods, and produce results that reproduce or closely match previous analysis pipelines. By using WebAssembly, our framework benefits from not requiring installation, command line knowledge, or transferring sensitive data over an internet connection. We show how algorithms from other researchers, and spanning fields from machine learning to physics, can rapidly be ported to web tools using our framework. We demonstrate an immediate impact on microbial bioinformatics, but expect that uptake of Sparrowhawk by developers of other tools broadly will help other researchers, and create new tools for teaching and public engagement. Thanks to the broad availability and familiarity with web browser tools across society, Sparrowhawk has the potential to significantly lower the barrier to entry to advanced computational methods.
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7236791540
