The Minutes of the CCP4 WG2 meeting at Birkbeck ----------------------------------------------- Weds. Sept. 27th 2000 at 11.30 am --------------------------------- Present: Phil Evans (Chairman) MRC, Cambridge Eleanor Dodson York Jan White Sheffield Alexei Vagin York Misha Isupov Exeter Ian Tickle Birkbeck Sheila Gover Oxford Paul Fyfe Bristol Jim Raftery Manchester Garib Murshudov York Christine Cardin Reading Harry Powell Cambridge Jim Naismith St Andrews Sue Bailey Daresbury/CCP4 Alun Ashton Daresbury/CCP4 Peter Briggs Daresbury/CCP4 Martyn Winn Daresbury/CCP4 David Brown Daresbury/CCP4 Tadeusz Skavzyuski Glaxow Peter Keller EBI/Cambridge Kim Henrick EBI/Cambridge Rasmus Fogh CCPn/Cambridge Michael Roberts StGeorges Med Sch Quan Hao De Montfort Liz Potterton York Eugene Krissmel EBI/Cambridge Kevin Cowtan York Appologies: Pierre Rizkallah, James Nicholson, Peter Moody, Neil Isaacs 1) Minutes of meeting of 3-5-00 and matters arising. Chairman ------------------------------------------------------------------------ Jan White : Any progress on SOLVE accepting MTZ unmerged Data? Phil explained that although Tom Terwilliger had been approached little had actually happened. Scala has an option OUTPUT POLISH? that allows scala output to be read into solve. Eleanor D. : Eleanor wanted to know if all the relevant people were aware of the progress with imageCIF or more correctly now CBF file format? Harry was happy to inform us that work on mosflm has been undertaken to allow it to read CBF though progress was slow on the detector manufacturers to output CBF. Detector software is expected and promised to output CBF soon. 2) Study Weekend 2001 Jim/Kevin ------------------------------------------------------------------------- Jim made an appeal for suggestions to fill in spaces in the timetable left empty to allow Hot Topics to be included and one slot probably being empty due to sickness. the topics are: Case Studies Fragment Fitting New approaches some suggestions were: Peter Artymiuk (Sheffield) David Jones (THREADER / Brunel) DALI (possible Liisa Holmes, EBI?) Pfam (Alex Bateman, Sanger) A decision was taken to keep the registration fee at 50ukp. David thanked those who had responded to email's to WG1 and WG2 offering assistance to help/host the visiting CCP4 scholarship students from India and China. However, more volunteers are needed. Pierre Rizkallah (DL) had approached CCP4 asking if it would be possible to hold the UK PX SRS users meeting just before the CCP4 study weekend. The meeting was unanimous that this was a good idea and suggested it be held on the Thursday afternoon so as not to clash with any of the study weekend activities. 3) Update on CCP11 ------------------------------------------------------------------------- David Moss could not be found 4) Introduction of Eugene Krissinel - CCP4/EBI post on new Martyn coordinate format ------------------------------------------------------------------------- Eugene Krissinel is a new CCP4 programmer employed to work at/alongside the EBI in Hinxton, Cambridge. Eugene is to work on a replacement coordinate format - with the eventual aim of replacing PDB as a working coordinate format. Kim H. questioned if the format will definitely be a binary format. Martyn pointed out that this was the decision of the special interest group, and no objections were raised in the meeting. 6) Report on Stanford Meeting Eleanor ------------------------------------------------------------------------- Summary of points raised: Attached to the Crystallography School organised by SSRL was a developers one day workshop. Those present in the workshop included: Representatives from the BLUE-Ice project (Beam-line control and monitoring software) Graphics Experts Representatives from the PHENIX Project [detailed list in EJD document] Much conversation then ensued on the PHENIX project and what it could mean to CCP4 points: PHENIX project is to be developed in Python. Data structures were/have been developed to optimise performance on modern computers. Statistical analysis tools have been developed. Symmetry tools have been/are being developed. coordinate and map 'formats' not yet investigated. The project still needs parser and core xtallography programs. This maybe done by existing software. PHENIX want to collaborate closely with CCP4. Questions were asked as to why there is a need to shift software development and working file formats from the traditional FORTRAN procedures. It was pointed out that software development is more productive in modern languages and programming techniques are used. [see elsewhere for definitions and explanations on OO programming e.g. I did a quick search on the web and found this: http://www.sunderland.ac.uk/~wu0gmc/What_is_Object_Orientated_Computing.htm ] Kevin gave an example of how development of DMMULTI has been frozen for the last few years due to major difficulties with the current data format. The problem is mainly due to the code being written in FORTRAN and Fortran is not recognised to have good data structures. He is currently working on code written in C++. Kevin has recently shown how a binary XDR data block when tarred together with an XML header could be used as an alternative reflection and map format. It was suggested to Kevin that Pickle could also be used as a data storage which he will now investigate. Eleanor raised the point that future CCP4 development should be conducted using modern programming languages and techniques [discussion restarted at end of meeting - details below] Kim pointed out that the correct nomenclature was Data structures. Data formats are for users ie a reporting tool. Kim was asked what would be the best possible option in a perfect world. Though Kim was realistic about the prospect of such a development he advocated the use of DATABASES. Much discussion then ensued on the merits (including speed, log history etc) and pitfalls (including the need to use proprietary software) of using databases. A decision was taken to hold a special interest group in York possibly during November. 8) Forthcoming release, including NT port Martyn/Pete/Alun ------------------------------------------------------------------------- Pete summarised the changes and contents of release 4.1 of the suite. The deadline for new or major alterations to programs is the end of October and the release date will be mid-November. New Programs: MOSFLM 6.10 (distributed with CCP4), CAVENV (calculate macromolecular cavities and envelopes), COMBAT (rotaprep replacement), D*TREK2MTZ, FFFEAR (fitting fragments in ED), ROTGEN New Versions: REFMAC5, RSPS, DM, MAPMASK, MAPROT, CCP4i, OASIS, MOLREP Other significant: CCP4 on NT, MTZ datasets have associated cell and wavelengths, Modlib updates to include BLAS and LAPACK, summary tags used more extensively. 5) Development of 3D graphical viewer Liz ------------------------------------------------------------------------- CCP4i - new version will include Sketcher which produces lib files on small molecules for refmac. tutorials on new refmac and sketcher on: http://www.dl.ac.uk/CCP/CCP4/martyn/tutorial2000/ The tutorials will grow and eventually be included in the release. 3d viewer Liz outlined how the package will need to be visualised as being in two halves. One half will be the core program including the graphics system and the second half will be application programs that may be written in any language. It was noted that work on data models is tied in with other CCP4 developments to be used throughout CCP4. In May it was agreed that 2 people could be recruited to work on the initiative. Liz also outlined that she was attempting to create links with other disciplines especially bioinformatics. Liz has been approached by several developers who already have graphical building/viewing packages expressing interest in the project. 2 new offers of software were given/outlined in the meeting. 1] IanT outlined how Astex have developed a program written in Java that can display CCP4 maps and molecules. The Program runs in a browser without any hardware acceleration. The program has some limited molecule manipulation tools. Ian was not sure if the source code could be made available. CCP4 would be able to distribute to industrials and academics alike. It was suggested that in the first instance the viewer was made available from the Astex web site. 2] Whilst in Stanford Duncan McRee had approached Phil and Eleanor about the possibility of CCP4 being able to distribute the Xtalview package. It was noted that much of the xtalview package was obsolete but the xfit program was worth investigating. For CCP4 to distribute xfit it was firstly thought that the widget set used for buttons etc needed to be updated and the program needed to be able to read mtz files. Kevin agreed to investigate the possibility of updating the toolkit before attempting to change the program to read mtz. It is unclear at this stage what McRee would want in return for xfit but it was thought that it would be unsuitable for CCP4 staff to support the package. 6) Progress with REFMAC Garib Refinement workshop / developers' meeting - Jan. 2001 ------------------------------------------------------------------------- Garib appealed for users of Refmac5 to report problems to him as this was part of the debugging process. He has already started working on 5.1! Jim pointed out that as a result of Refmac5 his laboratory has now stopped using CNS. Eleanor enquired (to Martyn and Kim) if there had been any progress on the way to record TLS in mmCIF. Kim pointed out that official progress was slow though an accepted mechanism was in place. Garib added that there was an option in Refmac5 that could be activated to allow it to output TLS parameters. He is encouraged to do this. 9) GRID - what might this mean for CCP4? Sue ------------------------------------------------------------------------- It was thought that people should be aware of the potential funding benefits that could be available to the community from an initiative known as GRID - a distributed computing project being developed by other fields, mainly physics. 10) Date of next meeting ------------------------------------------------------------------------- 24th January 2001, York. 11) AOB ------------------------------------------------------------------------- Mosflm. Harry reported that Mosflm will be distributed with the CCP4 package in future. A new version will soon be available. A mosflm tutorial was now available that had been developed with the assistance of E.Garman. It was suggested that maybe the existing mosflm documentation could be integrated/linked better and some assistance could be sought from Maria. Harry then outlined how there is funding for one person for a year to work on a mosflm interface. the funds were from an EU framework initiative. A schematic diagram of how mosflm is proposed to work in future is below: mosflm <-------> server <---------> graphics the server is to be developed by SSRL based on their BLUE-ICE software. The server is responsible for controlling the graphics output and what control the graphics application has on the mosflm job ie if there are 2 graphics jobs viewing one mosflm process, only one should be able to answer questions and influence the process. If mosflm requires input the request to the user is sent via the server to the graphics.