If you want to have this newsletter mailed to you or you want to make comments/suggestions about the format/content then send an email to acedb@sanger.ac.uk.
This month saw the first Acedb Workshop since 1997, which took place in Vancouver. Thanks go to Richard Bruskiewich and Steve Jones (a former Sanger person, who now works at the Genome Sequence Centre, Vancouver) for their efforts in organising the workshop. This issue will be about the conference and what happened there.
Apologies for no Acedb Users meeting this month, in the end there just was not time to organise one.
There will be a meeting in July, if you have issues you want to raise before then, please come and see the acedb developers in our new room (D2-12) which is just down from reception, near personnel/accounts etc.
Partly as a result of requests at the conference and partly because of independent requests from acedb users at other sites, this newsletter is now being added to the acedb newsgroup each month and is also being sent out to users at other sites. I hope to solicit some useful articles from contacts made at the conference.
I'm not going to cover everything that happened at Vancouver, the following sections deal with subjects most relevant to the Sanger Centre. A fuller description of the conference will appear on the acedb web site in due course.
Gudmund Thorission from Iceland gave a short presentation on how he hopes to use the grid models/display for micro-array data. The main points for discussion were:
This conversation mirrors several that have happened in the Sanger Centre over at least the last year (e.g. with Dan Lawson and Adam Butler). What's required now is an initial specification which includes:
Richard Durbin spent much of his time at Vancouver working on a new storage layer for acedb. This is the part of acedb that stores/retrieves objects to/from disk. The idea here would to allow a choice of "backends", perhaps including free (e.g. mySQL, BerkeleyDB, the original acedb system) or commercial (e.g. Oracle, Sysbase) systems to be used. The backend would not "know" about acedb objects, it would only be used to store/retrieve them and provide other services such as locking.
The rationale for doing this is that the backend system could be used to allow truly concurrent reading/writing via threading and more reliability via support for transactions. This would be a great advance for acedb and would supply features requested by Sanger and external users.
This work is currently only at the "broken prototype" stage, much work needs to be done on further prototyping, in particular to assess whether performance would be good enough with a different backend.
The session on the format/usability of the acedb web site produced one main conclusion: it's not very user friendly. Fortunately there were also a lot of suggestions for improvements including:
and a host of other good suggestions.
We plan to substantially reorganise the web site over the next few weeks.
There were a number of talks/tutorials on how to use/set up the new sockets-based version of the acedb server/client (available now in Acedb 4_8c in ~acedb/RELEASE.DEVELOPMENT/bin.XXX directories). If you would like more information or a tutorial on the new server/client system (how to set up user passwords etc.) then contact Ed Griffiths (edgrif@sanger.ac.uk).
Other important developments related to the socket server included:
Note that you can also interface to the socket server from Java: (http://www.biojava.org/docs/api/.
There has been a lot of discussion at Sanger over how to construct larger sequences up to whole chromosome level. There was a whole session devoted to this with an introductory talk by Richard Bruskiewich, followed by a talk by Richard Durbin which gave an overview of the proposed model for SMAP and a model for exons.
A fair summary would probably be that the SMAP model should solve a lot of the problems that were experienced with chromosome 22 and with large sequence construction in general. The difficult bit of the code still needs writing but that should happen over the next few weeks.
What to do about building haplotype sequences remains unsolved however.
There was a lot of interest in AQL as a more powerful replacement for the Query language and TableMaker. We would welcome more reports about how people find AQL to use, any performance problems etc.
A major requirement for AQL is some kind of query builder and also more tutorial documentation of the "how to..." type. Anyone fancy writing an article on how to use AQL ??
There were a significant number of users who would like to see a more up to date copy of acedb available on the Mac. Simon Kelley has made some headway with obtaining a software environment in which to build acedb on the Mac. The basic problem here is that we do not have the resources to maintain code that requires a completely different build environment, graphics routines etc.
Things look hopeful but no more than that.