The DDBJ Pipeline, which is a web application based on the SaaS model of cloud computing, assists in the submission of analysed results to DDBJ databases by automatically formatting data files and facilitates the web-based operation of NIG supercomputers for high-throughput data analysis.
To maximize performance, it uses grid-based computing and the macd queue submission manager.
SkateBLAST is derived from ViroBLAST (14), with a customized interface for job submissions to a UD high-performance computing cluster during heavy computational loads such as during Annotation Workshops.
In this way, simultaneous submissions, receptions, and computing are allowed in each node.
Individual task blocks (e.g. read mapping) are packaged in bash script modules, which can be executed locally or on subsets to test module code, submission parameters and compute node environment in stages.
Submission and monitoring of compute jobs, as well as the retrieval of postprocessed results are realized through a web based portal.
Multiple chains are run on independent threads, or on independent processors in a cluster computing environment (for which a torque job submission script is also available).
A database is used to store user account information, along with information about the computing jobs each user has submitted, such the submission and end times, the number of parallel jobs used and CPU hours used.
On HPC systems and clusters, usually batch submission systems are utilized for allocating the computing resources.
