These are a merged grouping of Anton's, Dan's, and Jen's notes from the breakout. Please feel free to update, comment, correct or add in anything we didn't have a chance to address during our session.
How Galaxy currently facilitates research
- Galaxy Main
- to identify 'vetted' tools
- to experiment with workflows before committing resources to Cloud instances
- to quickly access an instance with tools/reference data set-up and ready to go
- for storage (S3)
- General Comments
- The 'Big Plus' for end users: Amazon allows Galaxy to be used in highly controlled IT environments.
- "Galaxy better than commercial software" (the Galaxy Team says "thank-you"!!)
- Trackster/Sweepster amazed everyone (shout-out to Jeremy!)
- New end user community to be aware of: Mass spec
What would be most helpful going forward
Better methods to identify 'Vetted' and 'Best Practice' tools
- Applies to tools on Main, but in particular tools from Tool Shed
- Community currently relies on Main for this type of information
- Extends to Libraries & other source Data, Workflows, Visualizations
- Ideas (some overlap):
- Best practices as a way to adopt to correct tools
- A need to explain what is needed for particular analysis goals
- Tag tools/workflows based on publications, citations, peer-reviewed status
- DOI would be the ideal tag
- Methods to evaluate tools in the Tool Shed
- Tool Shed: Download counts/trends e.g. downloads this week week
- Tool Shed: Flag bad objects, like craigslist, to alert admin moderator
Data integration between Main and CloudMan
- S3 bucket attachment
- Ease transfer of libraries/datasets/histories/workflows/pages/etc (primary objects) between Main ↔ CloudMan
Tool integration between Tool Shed and CloudMan
- Ease installation of tools on cloud instances
ACTION: For all above - Discuss/Prioritize
- Add a free text "Header tool" that doesn't proliferate/duplicate datasets, or at least keeps to min. Alt actions on form: add line, remove/replace lines. E.g. UCSC custom tracks: track line for bedgraph file, browser line. May be two changes. ACTION: Needs clarification then bitbucket (Dan/Jen)
- Tools to do file conversions to/from GenomeSpace or was this a "Get Data' tool request? ACTION: Team discuss then bitbucket (Jen)
- Genome access was hurdle to Local install - Dan mentioned rsync server & auto genome tool. ACTION: Communicate/wiki (Jen/Dan)
- CloudMan auto-scaling is not working well for long running jobs - e.g. Tophat. ACTION: Team discuss then bitbucket. (Jen)
- Mailing list → Forum
- Galaxy-dev 1st. ACTION: Switch now (Dannon/Jen). Maybe add Google Searches (Dave).
- Galaxy-user 2nd. ACTION: Dual Forum/List ~ 1 year, then transition to forum only. (Dannon/Jen/Dave).
- Others no change.
- Trackster .... any data - rna, small genome region, etc. can be used as 'ref genome' with or without sequence. ACTION: Tutorial. (Jeremy/Jen)
- Multiple versions of tools in UI. ACTION: Tutorial. (Jen)
- Of interest: CloudMan Costs ~ $900 for 1 flow cell Hi-Seq analysis run. ACTION: Useful info to capture in wiki where users could share usage data? (Dave)
Plan of Action
See ACTION items inline above
Others to be added
More to come ...