Abstract: Understanding protein function is key to identify potential new drug targets as well as to unravel their links to pathogenic processes, but experimental annotation is challenging and time-consuming. Protein 3D structure can provide key insights into this task, but has been limited to those with regular protein folds. The evolution of protein modelling has provided the ability to structurally explore the entire proteome at great resolution.
Despite significant advances in protein function prediction, tools have primarily used sequence information, with variable efficacy and success. We created LEGO-CSM to fill this gap, is a comprehensive web-based resource that leverages our graph-based signatures and uses protein sequence and/or structure to accurately model key aspects relevant to protein function characterisation: Localization, Enzyme Commision numbers and GO terms.
LEGO-CSM implements predictive models in an easy-to-use interface as well as viz programming APIs to facilitate large-scale processing and incorporation into analytical pipelines.
- This website is free for all users. For policy please click here.