Data intensive analysis tasks such as analyzing genomic data increasingly call for more and more sophisticated levels of programming proficiency in researchers. Desktop and internet based applications have been designed to create workflows which will help address this problem. However many are unintuitive and have limited access to the high performance computing (HPC) infrastructure needed for large datasets.
In a recently published article published in Source Code for Biology and Medicine, Hunter et al. present Yabi; an open source software system designed to provide transparent access to high performance computing. Hunter and colleagues describe their new application as ‘a workflow engine that solves the problem of workflow deployment across disparate legacy HPC resources’. Designed to provide a more intuitive workflow environment, Yabi is a web-based user service intended to target an audience which may not have specialized programming skills.
Yabi employs a three tier architecture; a front end user interface, ‘middleware’ responsible for process management and ‘resource manager’ which computes resources to middleware. This structure allows the system to interface with a variety of different storage and computer systems facilitating easier access to storage resources and preventing ‘data lock-in’.
One of Yabi’s most interesting features is that it allows
users to design their own workflows by selecting each tool they wish to use and
then completing that tool’s mandatory inputs. The user can then name and save
the workflow they have just created which is then stored as part of an audit
trail and may be reused at a later date. Already being used by a number of
research communities in fields like genomics, transcriptomics and proteomics,
Yabi is ‘domain agnostic’ and can be used to service many different life
With the introduction of Yabi, Hunter et al. hope to move the complexity involved in accessing HPC away
from the individual researcher and enable scientists without specialized
computer knowledge access to HPC power. To download Yabi for yourself please
click the following link to the authors website; https://ccg.murdoch.edu.au/yabi/login/?next=/yabi/.
You can also follow the Yabi project on Twitter.