On Sun, 26 May 2002 kkli...@gate.sinica.edu.tw wrote: > Now the problem is, when I tried to load some of the PDB files > into a movie, pymol will link the water molecules wrongly.
> Some oxygen atoms are linked to more than two hydrogens. > What's more, I tested the data both on the Windows version > and the LINUX (redhat) version (both are the latest just downloaded > and installed last Thursday), there are many more mistakes in > the LINUX version than in the Windows version, > although both versions make mistakes. > I dig out the test data file of water (water.pdb) from the > package, which uses HETATM instead of ATOM for the atom records, > and again I tried it on both versions. That's the right idea - PyMOL doesn't currently use residue dictionaries for connectivity except for HETATM waters (in latest windows version I think). Residue-based connectivity is planned as a option for the future... > This time, there is no mistakes in the Windows version, Good. That means we've made some progress... > but still there are several mistakes in the LINUX version. Which is an older version, if you're using the binary (RPM, etc). > I was wondering whether there is anything we can do > to prevent pymol from linking the atoms wildly? If you compile from the CVS source code, you can get PyMOL working under linux right away with the HETATM water solution. Otherwise, just wait for the next linux binary release. > In my datafile, there are TER records after the protein > and after each molecule. > I also tried to add CONNECT records for waters, > but these do not provent pymol from making extra links. It should if the waters are HETATMs *and* you provide CONECT records (even on the older versions). To be absolutely sure that PyMOL can handle your structures, you might want to send me (war...@delanoscientific.com) a tar archive with 5-10 frames of your trajectory... Cheers, Warren
