[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[ccp4bb]: [cnsbb] Crystallography on OSX
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
This posting is more of an attempt to summarize my knowledge of the current state of crystallography software for Darwin from a user's standpoint rather than speculate about its direction. It should be noted at the outset that OSX and Darwin are essentially interchangeable as Darwin is the open source kernel upon which OSX has been built.
PHASING AND REFINEMENT:
I must admit that, like many other crystallographers, I am by no means a compiling expert, so I do need help at times. However, I have compiled a bit of information. The most relevant of which is that several claim to have compiled CNS for OSX/Darwin (e.g. <http://dryden.biol.yorku.ca/darwin.html>, <http://gnu-darwin.sourceforge.net/gdc/>) but I have never seen complete directions that may be followed to produce a workable CNS install. The link on the latter page is to the CNS 1.0 page at Yale, which lists Darwin/OSX as niether a supported or unsupported OS. I have gotten pretty far on the CNS install myself, but my fortran compiler, g77 from gcc-2.95.2 retrieved from the now deservedly famous page <http://www-jlc.kek.jp/~fujiik/macosx/10.1.X/>, chokes on a single routine in the entire CNS source (note also that the pricey commercial Absoft fortran compiler chokes at the exact same routine). There is some evidence that this problem has been fixed by the compiler's authors in!
later versions of the g77 compiler, specifically in gcc-3.0, but I can't say for sure as I still haven't got gcc-3.0 to work, despite my attempt at following the instructions at <http://dryden.biol.yorku.ca/darwin.html>. Of course another option is ccp4, but I have not yet tried to port ccp4 to OSX. Some notes for porting CCP4 to OSX/Darwin may be found at <http://gnu-darwin.sourceforge.net/gdc/ccp4.html>.
MODEL BUILDING
I think everyone knows that O version 8 is ported to OSX. It works great! One technical issue, though, it helps to start O from your "O" directory and have the O executable (osx_ono) or a link to it in your O directory. The commands to start O should look something like this:
> cd ~/My_O_Directory/
> ./osx_ono
This will allow the directories in your $ODAT environment variable to be utilized correctly.
DATA REDUCTION
Neither of the commercial programs HKL or d*trek lists OSX/Darwin as an option. I do not think it would be wise to hold one's breath. However, mosflm may be an option. A working version of mosflm is reported at <http://gnu-darwin.sourceforge.net/gdc/>, though no instructions or binaries are provided. The screenshot at this page of mosflm in action contains a picture of HKLview in action, a ccp4 program to view a reciprocal lattice of a data set that has already been processed. The mosflm window is a text window and is somewhat tough to spot. This screenshot does hint at X-windows running under OSX/Darwin. One may easily set this up by following the instructions at <http://www.mrcla.com/XonX/> and the links therein. I have not yet tried to compile mosflm for Darwin/OSX.
PRESENTATION
<http://dryden.biol.yorku.ca/darwin.html> lists molmol as working and even has a downloadable installer. I have not tried this install yet. PyMol is reported to work at both the PyMol home page <http://pymol.sourceforge.net/> and at <http://gnu-darwin.sourceforge.net/gdc/>, though good instructions are not given at either and I have not tried. The PyMol home page does claim that OSX will be supported in future versions. I have made some nice figures using the MacOS 9.1 with PovRay and SwissPDBViewer, but that is beside the point.
In summary (from a non-computer science expert standpoint):
CNS: Cant make it work with anyone's instructions
CCP4: Might work
HKL: Nope
D*Trek: Nope
Mosflm: No good instructions on making this work anywhere
O: Something cool actually works!
molmol: Might work
PyMol: Probably works now, will probably work soon
As I have said, I am not an expert, so if I am wrong in this summary, please let me know. But before you make some grandiose claim about how much crystallography software you have running on your OSX box, provide some thorough, reproducible instructions for the rest of us! Isn't that how we do it in the scientific journals?
Without getting too deep into the merits of operating systems, I wouldn't blow a lot of money on Mac OSX boxes just yet as much software is not supported for them and is often painful to get set up. CNS alone has eaten many days of my non-expert time. So Linux or, heaven forbid, windows, may be a more feasible option for low cost crystallography stations at this time.
James
--
--------------------------------------------------------
James C. Stroud
Department of Chemistry and Biochemistry
University of Colorado at Boulder
Boulder, CO 80309
Tel: 303-492-4503 Fax: 303-735-1347
--------------------------------------------------------