Ok - small oversight - I can see now that phenix has the atoms placed/saved during cutout stage - so how best to use PHASER output to shift this file?
Andy From: Andrew Lovering (School of Biosciences) Sent: 10 January 2019 10:48 To: 'ccp4bb@jiscmail.ac.uk' Subject: RE: complex multi-crystal averaging Thanks everyone for the pointers. I should clarify - my issue here is the multiple stage "disconnect" between PDB in xtal1 -> cutout density (into a "interim cell" just for this stage) -> phaser MR with density search model into xtal2 Hence, do I just simply pop the model from (1) into the "interim cell" (display cutout density, move molecule here in coot, bit of manual shifting / rigid body, save interim PDB) then use PDBSET on saved PDB with angles and shifts from phaser output? To use clear language, "what's the easiest way to get PDB xtal1 to match placed density of MR in xtal2" :) Andy From: Andrew Lovering (School of Biosciences) Sent: 09 January 2019 16:08 To: 'ccp4bb@jiscmail.ac.uk' Subject: complex multi-crystal averaging Dear All, I suspect the way out of this is a new crystal (!) but interested to hear any advice. I have two crystal forms of a 500aa protein, vaguely tube-shaped 1=P3121, diffracts to 4.1 ang, 3 copies in asu, 86% solvent; map indicates it is a relative of other folds (but those are not so close that they'd be a good guide to sequence register). I have selenomet SAD phases which helps identify Met positions. The 3-fold and high solvent give a great map but you wouldn't want to build it and buccanner and phenix think similar 2=I222, 2 copies per asu, 66% solvent. This has a cell that gives wildly anisotropic diffraction - ~3, 3.5, 4.3 down different axes. Not really rectified by staraniso. No phases So I can cut the density out of form 1 map (using secondary structure elements of a rough PDB as a mask), and use phaser with this density as search model to find the two copies with a TFZ of about 10. The phaser map shows a bit of detail and the solution has placed the protomers such that they agree with a self-rotation function. Phenix find_ncs on phaser map similarly agrees (as I would expect). Now...I have an eye on multi-crystal averaging between the two forms. BUT the phaser map isn't good enough to manually place the PDB used in cut-out density, and I can't see a straightforward way of using the angle info in phaser sol to perform a co-ordinate transformation (but I think ccp4bb users will put me right on this - I'd imagine placing the PDB into the cutout density mtz then using PDBset?). I tried converting the phaser mtz to a map using FFT then using phased TF in Molrep to place the PDB but this didn't work, complaining about the grid used. One last caveat - I have multiple sets for the I222 that intriguingly differ by 12 angstrom down a 240 ang axis: doing multi-crystal averaging between these two forms achieves little when I would expect otherwise (all the NCS correlations are good, high initial agreement) Like I said...fingers crossed for a new crystal form! Andy ######################################################################## To unsubscribe from the CCP4BB list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB&A=1