Nathaniel Echols wrote:

On Tue, Sep 13, 2011 at 3:37 PM, Yuri wrote:

...

This may be a silly idea after all, nonetheless: Would it be possible (worhtwhile) to have a tool that lets the user input a molecule of a certain type (probably rather small size, i.e, an amino acid or cofactor), and then input a space group. Phenix could then generate a relection file that would be displayed using either data2d or 3d viewer. Obviously there are many other factors to be considered but just an idea...

This is very easy (basically just running eLBOW + phenix.fmodel), and it's actually relevant to something else I wanted to try, so I'll add it when I get a chance. (You'll have to wait until after the next release, however.)

In the meantime if you have ccp4 installed, sfall will calculate the structure factors, which you could display as above or with ccp4. You'll have to give unit cell as well as space group, and it won't add bulk solvent which may not be a problem for your use. something like: (or unit cell and SG can be on CRYST1 card in model.pdb) sfall XYZIN model.pdb HKLOUT model.mtz << END-sfall MODE SFCALC XYZIN CELL 15 20 25 90 90 90 SYMM P212121 RESO 30 2.1 LABO FC=FC PHIC=PHIC END-sfall

Nathaniel Echols wrote:

On Tue, Sep 13, 2011 at 4:31 PM, Edward A. Berry wrote:

...

In the meantime if you have ccp4 installed, sfall will calculate the structure factors, which you could display as above or with ccp4.

phenix.fmodel does this too - it will also include bulk solvent. I'm not sure if this is important for the intended use; I wouldn't expect reflections from a fake small-molecule crystal to be visually similar to those of real macromolecular crystals anyway.

This kind of exercise i good for educational purposes. For example make a model with 5 atoms, say alanine arbitrarily placed in a p1 cell. Now make an "omit" model whier one atom is missing. Calculate Fc and PhiC from both and combine them to make a map in various ways, using resolution 0.5 so each atom makes a peak. Does the omitted atom really show upat half-height in the Freal-PhiC and Fc-Phi-real maps? full height in the 2Fo-Fc map? Or switch it and say 4 atoms is the real structure, 5 is the model- does the extra atom disappear in the 2Fo-Fc map and show up negative in the Fo-Fc? (OK, the last two are obvious and don't need demonstrating) Or explore the limits of resolution imposed by the smoothness of atomic orbitals: calculate Fc at successively higher and higher resolution, use sfcheck to calculate the optical resolution, and explain why it doesn't go below about 0.5 A, and what happens if you give all the atoms B-factors around 80.

On 14/09/11 01:17, Pavel Afonine wrote:

That's exactly how I did most of the slides (that show maps) here: http://www.phenix-online.org/presentations/latest/pavel_maps.pdf

Pavel, there are some slides here that I would like to reproduce (with credit of course). Maybe I do so? Paul.

sfall XYZIN model.pdb HKLOUT model.mtz << END-sfall MODE SFCALC XYZIN CELL 15 20 25 90 90 90 SYMM P212121 RESO 30 2.1 LABO FC=FC PHIC=PHIC END-sfall

Here is another one: iotbx.pdb.as_xray_structure model.pdb --fake=2.5 Faster than phenix.fmodel, but no bulk-solvent correction. Ralf