\documentclass[a4paper]{article} \usepackage{times,mathptm} \newcommand{\action}[1]{\marginpar{\flushleft \textbf{Action}: #1}} \begin{document} \begin{center} \large\bf MINUTES OF THE WORKING GROUP 2 MEETING OF CCP4 HELD AT BIRKBECK ON 6th SEPTEMBER 1995 \end{center}\vspace{5ex} \noindent{\it Present:} \noindent \begin{tabular}{@{}l@{\hspace*{2em}}l} \\ David Rice (Chairman) & Sheffield University \\ Don Akrigg & Leeds University \\ Peter Brick & Imperial College\\ %Ian Clifton & Oxford University \\ Kevin Cowtan & York University \\ Christine Cardin & Reading\\ Peter Daly & ESRF Grenoble \\ Eleanor Dodson & York University \\ Cameron Dunn & Bristol University \\ %Paul Emsley & Glasgow University \\ Phil Evans & MRC, Cambridge \\ Sheila Gover & Oxford University \\ %Paul Holden & Bristol University \\ Neil Isaacs & Glasgow University \\ Misha Isupov & Exeter University \\ Peter Keller & Bath University \\ Gareth Lewis & Southampton \\ David Love (Acting Secretary) & Daresbury Laboratory \\ %Sandra McLaughlin & Hamburg \\ %Simon Phillips & Leeds University \\ Jim Raftery & Manchester University \\ Adam Ralph & Daresbury Laboratory \\ Tadeusz Skarzynski (Industrial rep) & Glaxo \\ Ian Tickle & Birkbeck College \\ %David Waller & Sheffield University \\ Steve Yewdall & Edinburgh University \\ \end{tabular}\vspace{2ex} \noindent\textit{Apologies: } David Waller, Paul Freemont, Simon Phillips.\vspace{3em} \noindent Items for discussion were taken in the following order, not per the agenda. Neil Isaacs assumed the Chair after David Rice's departure at lunchtime on other business. \section{Minutes of the previous meeting and matters arising} Concerning the minutes of the meeting held on June 7, there were several typographical and editorial corrections, some noted in items below. \section{Study Weekend} The main speakers have been arranged and a provisional programme drawn up, which Eleanor is to finalise.\action{Eleanor} Eleanor appealed for suggestions for suggestions of other speakers to give 20 minute case-study talks. Some suggestions were made. The publicity flyer is in print and the closing date is November 5. It has been stated that the Daresbury conference office will be recharging for their services by that time, but the cost has not yet been determined by the Lab management. We will need to find out the outcome.\action{Adam} This will add to the estimated costs and the extra amount will have to be found.\footnote{It transpires that costs won't be recharged this year, but may well be in future, depending on Lab.\ policy decisions.} \section{New committee members} Adam had written to various unrepresented UK PX groups inviting them to send representatives to Working Group 2 meetings and several had accepted the invitation. The new attendees were welcomed. The meeting is open to representatives of all UK PX groups. A complete list should be assembled and invitations issued to those remaining. If numbers get too large joint representatives might be in order (as already agreed by Reading/ICRF). There was a suggestion that the somewhat isolated Irish groups should be invited. \section{Grant renewal report} The CCP4 grant renewal application had been fully funded (running costs, existing staff salaries and an additional post-doctoral post). The application had apparently been received enthusiastically. The comments in the referees' should be noted regarding suggestions for missing facilities and directions (see item \ref{sect:objectives}). \section{New appointments/prioritisation of objectives}\label{sect:objectives} Neil Isaacs led the discussion on policy issues. It was necessary to consider the separate aspects of managerial issues and decisions on direction for development etc. There was discussion about the r\^ole of the post-doc funded by the grant. It was generally agreed that this post should not be based at Daresbury. Dave Love thought it important that the reporting structure for someone employed by Daresbury but based at a university should be worked out satisfactorily. Practical matters of overhead charges etc.\ need to be discussed with Peter Lindley. There were differing opinions on whether this person should do purely development work or maintenance too; Neil Isaacs said both. David Rice suggested splitting the three year grant into two 18-month blocks for different people but this was not generally thought advisable. It was asked what type of problems people were raising and whether the workload has gone up in proportion to the suite's usage. Dave Love said that there was a wide range of things that had to be daely with and that it wasn't clear how the load was related to usage since the relative usage level isn't known, but that demands were still high. Tadeusz asked whether this is what requires resources. He was asked what support the industrial customers required; the answer is not so much since they tend to be relatively expert. Dave Love concurred with the comment on their expertise. Neil reported that the referees' comments suggested that areas to be addressed included: real space molecular replacement, molecular dynamics (twice) and graphical interfaces and graphical display (a long comment). There was a general feeling that molecular dynamics had not proved its worth. A discussion about graphics systems followed. Tadeusz thought it important to distinguish between proper interactive graphics systems and simple form-filling to input parameters for essentially batch mode working; Dave Love concurred. Phil Evans suggested using Web browser tools for interaction as employed by John Irwin at LMB\@. Neil Isaacs and Dave Love thought this was not sensible. The question was raised about whether \texttt{xloggraph} display is helpful or not and in general whether graphical input or output was required, without clear conclusions. Eleanor contrasted the usefulness of \texttt{denzo}'s interface compared with \texttt{mosflm}'s and reported a general move to \texttt{denzo} for that reason. Neil asked for proposals on graphics and contributions on what people considered would be helpful. Regarding manpower, Dr.\ Martyn Winn has been appointed to the advertised post funded by industrial income, to start at the beginning of October. Peter Keller's 6-month contract to work on mmCIF facilities has been extended by another 6 months. \section{Finance} There was no significant change in the financial situation since the previous reportm though note the extra cost associated with te study weekend due to conference office recharging. The income account currently stands at $\sim$\pounds66\,k and non-salary spending at $\sim$\pounds2\,k (not including the cost of a new workstation under tender). \section{Redistribution and networking problems\protect\footnote{This item was mistakenly put on the agenda though it was discussed at the previous meeting; nevertheless, is was discussed.}} There have been various complaints (some from commercial users) about inability to access the ftp server usably from abroad due to the poor availability of international Internet bandwidth. This is particulaly a problem from the US, but possibly also a general one from other places. Dave Love suggested two options: (1) use one or more mirror sites the other side of slow links for redistribution (2) buy extra bandwidth from a service-provider (possibly for paying customers). (1) is currently not an option due to the distribution conditions, but these might be changed. There had been a volunteer from Stanford to host a mirror in if this impediment was removed and San Diego Supercomputing Center was suggested as another possibility worth approaching. There was uncertainty whether the PDB would provide such a service (c.f.\ previous minutes) and, if so, whether this was sensible in view of their heavy network loading. It was agreed that we needed some sort of assurance about the reliability of any such mirror site and, preferably, administrative control. It was reported that the trans-Atlantic bandwidth was due to be doubled shortly.\footnote{Daresbury networkers say this is not so, but there is a promising proposal to upgrade the current 2$\times$2\,Mbaud link to 34\,Mbaud, which would probably suffice and solve our problems in the medium term.} The Working Group chairmen\action{David Rice, Neil Isaacs} would reconsider the current distribution terms in the light of the possible need for mirror sites. \section{mmCIF} Peter Keller asked to clarify remarks attributed to him in the previous minutes: There are errors in the published `specification' of STAR (the basis of CIF) and (mm)CIF itself. The authors assumed that `interpretation' of the definitions was necessary, which seems to defeat the object. There are system-dependencies such as what constitutes a newline and the restriction to ASCII character sets. The \texttt{rwbrook} routines \emph{will} be replaced, but with a unified interface to PDB/CIF, the format on a given stream being chosen at runtime. Programs would be able to put out both formats, selectable by a mechanism such as logical names, keywords or file extension semantics yet to be decided. Peter gave a brief report of his visit to the mmCIF Workshop at the ACA meeting. One salient point to emerge is that there are \emph{no known implementations} of mmCIF-handling software able to process files using the current definition other than Peter's current prototype and (arguably) Hall's general \texttt{starbase} program. The mmCIF definition is still not stable, however. The feasibility of PDB$\rightarrow$mmCIF (subset) and mmCIF$\rightarrow$PDB converters was discussed. Alwyn Jones is apparently unwilling to support mmCIF and so there will be a continuing need to generate PDB files for \texttt{O}, probably with the latter converter. Note that no current programs purporting to do this are known to be compatible with the current mmCIF definition. Peter had e-mailed notes on his current implementation before the meeting and solicited comments. The major point raised was that their needs to be the possibility of having several CIFs open at once\action{Peter}, possibly with the same \verb+data_+ tag, since that isn't envisaged to change during refinements etc. There was debate about the need for a full implementation as opposed to just coordinates as a working format in conjunction with extra software for PDB submission. Peter is producing a general, layered implementation, about which there were doubts expressed. Dave Love supports the current plan as good practice. Peter pointed out that (surprisingly) it is allowed for mmCIF files to be missing such data as cell information and yet have coordinates expressed in fractional coordinates, so it will be necessary to enforce constraints on mmCIf-conformant input files to allow them to be processed satisfactorily with the \texttt{rbrook}-replacement routine. He reported that there is work in progress (separately) on defining a CIF-type format for reflection data. This was not thought likely to be useful other than for PDB submission if required. It was agreed that Peter should proceed with the current plan of producing general mmCIF reading and writing facilities and tracking and attempting to influence the continuing definition of the format. When the mmCIf `toolbox' is usable, the current coordinate-handling programs would then be converted to use it one-by-one so that mmCIF becomes usable in parallel with PDB\@. The first program to be tackled will be \texttt{coordconv}, to provide the interconversion (as far as this is possible without losing information). \section{Missing data treatment} Eleanor had implemented the oft-mooted handling of missing data in MTZ files (represented by arbitrary values, but normally IEEE NaNs or VMS ROps). She explained the implementation, which involves changes to MTZ headers and additional routines in \texttt{mtzlib} which must be called by all MTZ-reading programs. This had been installed at Daresbury and caused serious problems with floating point exceptions due to some programs not having been converted and others needing more work. There was a widespread feeling that the implementation should be altered. Currently the only programs which insert such missing data values are \texttt{unique} and \texttt{cad}, though \texttt{mtzutils} should also. Eleanor expressed the opinion that \texttt{unique} should be used as a matter of course to insert records for all $hkl$s in the resolution range for later use calculating completeness and inserting free $R$ flags. This essentially requires the use of missing data values. Regarding free $R$, various programs had been changed to implement new semantics for the free $R$ column allowing full cross-validation.\footnote{For information on this, consult a Br\"unger preprint for the \textsl{Methods in Enzymology} volume available on the Web under \texttt{http:/\slash xplor.csb.yale.edu\slash group\slash papers\slash papers.html}} There was considerable discussion of these points and the question was raised of what were appropriate columns for \texttt{truncate} to put out\action{Eleanor, Phil} ($I$ and $\sigma I$ and/or $F$ and $\sigma F$). \section{Programs} \begin{description} \item[\texttt{solomon}] Dave Love asked for clarification about Jan-Pieter Abrahams \texttt{solomon} program---exactly how it differes from \texttt{dm} and whether we really want it since \texttt{dm} now implements the `Solomon' technique for density modification. Kevin Cowtan explained that its advantage was for more difficult problems and experimentation, since it `uncanned' compared with \texttt{dm} and thus more flexible. This was thought to be desirable, despite the previous policy to support just \texttt{dm}. \texttt{solomon} is written in C and may present some portability problems, depending on how the interface to MTZ-reading works. It also needs programs from the RAVE system, although Kevin may provide alternatives in the future. Dave Love\action{Dave Love} is to contact Jan-Pieter to confirm that it can be included and then work on doing so. \item[\texttt{amore}] The item concerning \texttt{amore} in the previous minutes concerns cutting out and manipulating bits of maps. \item[New programs] Dave Love reported that the Birkbeck refinement program \texttt{restrain} was being added along with \texttt{crossec} for deriving scattering factors (with permission from Don Cromer) and \texttt{hgen} (from Eleanor, originally Simon Phillips) to generate hydrogen atom positions. \item[\texttt{watersort}] Phil reported that \texttt{watersort}\action{Dave Love, Phil} is broken and should be replaced by his rewrite (to be obtained by DL). \item[Anisotropic scaling] Phil reported that he had a copy of a Sherriff and Hendrikson program for anisotropic data scaling, currently called \texttt{compar}. There was some dissent about the merits of such a technique. It was agreed the program should be included anyway.\action{Dave Love, Phil} \item[Obsolete programs] Dave Love asked why obsolete programs were to be moved to `unsupported' (per previous minutes) rather than removed to lessen maintenance and user-confusion problems. It was agreed that the ones whose functionality is definitely elsewhere could be removed,\action{Dave Love} but not \texttt{almn} (which has functionality not in \texttt{amore}) or \texttt{flatmap} (not yet replaced by Kevin Cowtan's utilities). \item[\texttt{mosflm}] The status of \texttt{mosflm} and related things in CCP4 such as \texttt{abscale}, \texttt{oscgen} (which was actually removed some time ago) and \texttt{graflib} was also queried by Dave Love, who was asked to\action{Dave Love} liaise with Andrew Leslie about what the current and next \texttt{mosflm} versions. \end{description} \section{Any other business and date of next meeting} Adam reported that there had been 22 positive responses to the previously-mooted idea of a CCP4 `workshop'. It was thought this was probably an underestimate of interest in view of the fraction of users probably reading the CCP4BB list where interest was solicited. Some US respondents had requested a `virtual workshop' (presumably along the lines of the Internet protein structure course rather than on paper\dots). This was though to be even more work than a real one. The problem is still (as previously) finding organisers\slash tutors for such an event. This is one of the sort of things for which Keith Wilson apparently has funds and he should be contacted about the possibilities.\action{Eleanor} It would still be helpful to have more\slash imporved tutorial material in the distribution to benefit everyone rather than just attendees at a specific event. Dave Love asked what the `MR tutorial' mentioned in the previous minutes was. This has been done by Paul Hempstead, but it is not clear how it relates to Ian Tickle's existing material from the Bath summer school.\footnote{Copies of Ian's material could probably be arranged through Daresbury if required---this hasn't yet been included in the distribution\slash manual.} A (further) copy should be obtained as the location original, apparently sent to Sue, isn't known.\action{Dave Love} The situation regarding maintenance of the FAQ isn't clear; Dave Love has hacked the version put up in June, but is still unhappy with it. Eleanor reported that she had been asked to provide (urgently) an article on CCP4 for a fortcoming \textsl{Methods in Enzymology} volume which was going to contain items on other packages. She asked for volunteers to do it or help, and none were forthcoming.\action{Eleanor} She suggested that Sue Bailey might be willing to work on it at home during her current maternity leave and would contact Sue. The next meeting was set for Wednesday December 6, at York. \end{document}