- Sequencing: turn sample in test tube into digital sequence ACTGGACT....
- Sequencing is expensive, but lots of biology labs need it
- So, universities establish central sequencing centre
- You pay sequencing centre from your lab budget
- You give them sample, they give you back the digital sequence
- The researcher
- Likely an overworked, underpaid grad student
- He doesn't want to deal with software
- He just wants his sequence back so he can get on with his research
- The sequencing tech
- Her job is to take submitted samples, sequence them, then return these data to the researcher
- Again: busy, underpaid, not a programmer -- wants software to get out of the way, be unobtrusive
- The sequencing administrator
- aka "the bean counter"
- Wants to know which labs are using sequencing, how much they're paying, what costs of sequencing centre are
- MVP will be our first release
- Anything less than MVP is useless to the centre -- must support at least the researcher and the sequencing tech
- Any software will be big improvement on current pen/paper system
- All high priority features will be implemented for MVP. Low/med come later
- User permissions -- we must present different systems to researcher and sequencing tech (med priority)
- We don't need this for MVP -- we're in university setting, so we can trust people to be who they say they are
- Monitor status of sequencing jobs -- is it done yet? (med priority)
- Explicit listing not necessary -- if files not available for network-based access, then they're not ready
- E-mail notifications of sequencing status (low priority)
- Initial information entry before sequencing -- name, sample ID, etc. (high priority)
- Network-based access to sequence results (high priority)
- Idea: can we use existing issue tracker like JIRA?
- JIRA: you have tasks, users, status of task, notifications ... this sounds a lot like our domain
- JIRA has an API, so we could later build our own frontend, but continue using JIRA as our backend
- Needs access to initial data accompanying sample (high priority)
- Must be able to link output on computer attached to sequencer to the customer record (high priority)
- Barcodes on test tubes could ease information retrieval (med priority)
- Build desktop app for local users (sequencing tech), web app for customers
- Customers could also enter data in computer in sequencing centre, when dropping off their samples
- We need tons of storage for sequencing results (multi-GB text files), but not much computation
- Can delete sequencing results ater they're downloaded -- but what if user doesn't download them for a month?
- Can we build this in three months?
- We have ~10 programmers working 40 hours/week
- Consensus: yes, definitely -- we'll follow an agile process