EMBOSS: Project Meeting (Tue 5th May 09) |
Alan has also corrected edialign which crashed if the maximum fragment size was greater than 40. This was a fixed size memory allocation bug in the original dialign code.
Patches are available on the FTP server for all Alan's fixes.
Tim Carver has requested a change to restrict to knockout restriction sites that are methylated by the dcm and dam methylase systems. Other programs needing the same capability are restover, remap and showseq. A methylase data file will be needed with patterns for the methylase target site. Internally, the methylated base is replaced by 'N' which is not accepted by any restriction site. We need to check whether methylation of one strand prevents a cut in both stands at the restriction site.
Alan has updated library data structures to avoid alignment warnings on 64 bit systems.
Alan reported that the latest MACOSX compiler issues 3 warnings for print statements with no format string in plplot and AJAX.
Martin Senger will visit on June 2nd.
The production server will be updated with the namespace changed from embossws to soaplab.ebi.ac.uk with the organizationname substituted by an include file. This should be completed later in the week.
Tests using the RunAndWaitFor method have timed out. The server runs and completes, but the apache gives a timeout to the client. This method could return a report before the timeout to say the service is still running.
Mahmut attended a short reads bioinformatics course at EBI. Topics included:
Mahmut suggested generating a Jemboss client with jar files. This needs a known location for libajax and libnucleus unless the JNI dependencies are removed. We need to check whether some JNI dependencies remain.
Alan has drafted a server specification for Peter to review. The disks are SAS (serially attached SCSI) which are faster than SATA. The cost is about £500/TB. We have a budget of £7.5k for the server and £12.5k for disks although we have felixibility. We should be able to afford about 16TB of disk. The processor specification is 2 quad core processors, preferably (for large indexing jobs and multiple logins) with multiple disk channels. Alteratively we could use a separate smaller machie for the index building tests.
Alan reported that the 3TB disk array failed to survive the data centre power failure. Gavin will remove the failed components and hand over to us for testing. The cause is probably a disk failure. It seems sensible to replace all the aging disks.
Peter reported that the new grant has started on 1st May. We have a Scientific Advisory Board. Peter will contact them with the good news and ask about availability for an SAB meeting in September.
Peter will revise the Developers manual, and Alan will review it.
Peter will review the User manual. Alan has some updates from Lisa for her chapter.
edialign is patched to have the "aligned" attribute set for the output.
Next week is the EMBRACE AGM and review. Peter and Mahmut are at a
meeting on Monday 18th. The next meeting will be on Tuesday 19th May.
5. User queries and answers
Alan has fixed a remap bug reported by Jean Mao
6. AOB
None.
7. Date Of Next Meeting
Future meetings will be weekly at the usual time (10am on Mondays) in
the Hall View Room. Alan will chair the meetings. Long
technical discussions can continue after the meeting.