Key methodologies
1. Proteome annotation
The vision for the combined workflow for proteome annotation is shown in Figure 2:
2. DASTY front end for proteome annotation
- DAS decouples the databases from how the data is displayed in the e-protein portal, allowing the annotations to be easily accessed using other clients.
- DAS is used by many other projects allowing seamless integration of our annotation with theirs.
- The DASTY viewer allows the user to combine and compare annotation from the component databases on one page.
3. Job Yield Distribution Environment (JYDE)
The team at UCL have developed the meta-scheduler JYDE to manage the combined used of computational power at different sites at UCL and at Imperial. It was designed with the
specific aim of supporting a high volume production system for a user community with limited interest in the technical aspects of scheduling. Although it can provide user with information about the status and
location of their jobs, it does not require users to have any knowledge about processing details. Neither does it require user to know the resource requirements of their requests. This is achieved by providing
a selection of services for the user to choose between. They will accept the user request and fragment it into suitable tasks for distribution among available resources.
A demonstration of the system was undertaken recently when it was used to combine three clusters with just over 500 processors for the purpose of annotating the human proteome.
The system managed to accomplish this in around 24 hours, maintaining a steady throughput of around 1450 tasks per hour.
4. GridSAM
Imperial have developed GridSam that enables orchestration of different components of the proteome annotation to be performed on different computers in a seamless manner.
| 
