Hi Paul, Thanks a lot, I'll try to work with the original map as suggested.
Best, Qiuye On Mon, Sep 21, 2020 at 10:54 AM Paul Emsley <[email protected]> wrote: > > On 21/09/2020 14:49, Qiuye Li wrote: > > > Yes, I extract one single subunit just to quickly access the shape or the > sequence of the ordered region; next I would like to model 3~5 subunits. > The reason I think segmentation is necessary is, many subunits at the ends > of the filament were half chopped off (by masking in relion). So, to have a > map consisting of entirely full subunits, I tried to clean up the ends. > However, even doing so ended up replicate display problems in Coot (just > tested a "clean" 5mer in a 125*109*59 box). It seems that as long as I use > segment->select->save selected region function in Segger, it will cause > such problems. So I guess the question now is, how should one extract > a subregion of a map without causing display problems in Coot? > > > 1) It is not yet clear to me that there is any problem with Coot. > > 2) I am unconvinced that extraction of a sub-region is needed. If I wanted > to see the shape of the density or side chains, I'd look down the z axis > and narrow down the clipping. > > 3) If there are sub-units at the ends of the filament that are chopped by > masking, then build in a region where the masking hasn't chopped the map. I > don't think that you need a map consisting entirely of full sub-units. You > just need a few - maybe only one. > > 4) If this is like the tau filaments built by my colleagues over the last > couple of years it should be straight-forward, shove in a few beta strands, > link them and extend them as needed... add the symmetry-related copies as I > mentioned yesterday. The only non-trivial problems (IIRC) were the sequence > assignment and wondering if the strand went up or down for the next rung at > the corner with bad/missing density. > > > Paul. > > > Any information/background/suggestion are welcome! > > Best, > Qiuye > > On Sun, Sep 20, 2020 at 10:41 PM Paul Emsley <[email protected]> > wrote: > >> >> On 20/09/2020 19:27, Qiuye Li wrote: >> >> >> The protein is a filament and the reconstructed map is in a 512*512*512 >> box. I segmented this reconstructed map in Chimera using Segger, selected a >> single asymmetrical unit, and saved it as a new map (which was referred to >> as the segmented map). This new map automatically changed to a box size of >> 94*84*26. >> >> >> That does seem a bit thin in z. Does segger know about helical symmetry? >> If not, I don't think that this is a winning strategy. >> >> But it does occur to me to wonder why you even need to segment the map. >> Relion tells you the helical symmetry and you should (it seems to me) just >> apply that to the model. >> >> if you even need to do that... I would build a model into the set of >> tubes in the middle, and then when I wanted to be concerned about symmetry, >> just try cut paste rotate translate (just shove it up a rung and use jiggle >> fit) and let the NCS chain copy tool modify the other chains as I go. From >> what I can see of your map, one chain either side of the central chain >> should be enough. >> >> >> Paul. >> >> >> BTW, I tried to display from another direction and it looks like these >> copies do lie along the axis, see the attachment. >> >> Best, >> Qiuye >> >> >> On Sun, Sep 20, 2020 at 1:29 PM Paul Emsley <[email protected]> >> wrote: >> >>> >>> On 20/09/2020 17:29, Qiuye Li wrote: >>> > >>> > Thanks for the suggestion, it looks map segmentation and coot do not >>> > work well with each other, see the tests I did below: >>> > >>> > 1. .mrc map directly from relion: >>> > in Chimera: OK >>> > in Coot: OK >>> > >>> > 2. segmented map generated from Chimera: >>> >>> >>> what do you mean by "segmented"? >>> >>> >>> > in Chimera: OK >>> > in Coot: displayed as replicates >>> > >>> > When I segment the map, the box dimensions and map threshold level >>> > were changed, would these cause problems? >>> > >>> >>> I am not really sure what you mean by the map threshold level. It's >>> something of a surprise that the box (cell) dimensions have changed. >>> Maybe that's a result of reboxing? But I doubt that that would cause >>> problems to Coot. >>> >>> >>> It looks like the map is interspersed with sections that have zero >>> value. However, from the orientation axes top left it doesn't seem that >>> the sections lie along the direction of a cell axis. >>> >>> >>> Paul. >>> >>> >>> ######################################################################## To unsubscribe from the COOT list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=COOT&A=1 This message was issued to members of www.jiscmail.ac.uk/COOT, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
