Mehmet Serkan Apaydin wrote:

> Hi, > > I am a new MODELLER user. I am trying to duplicate a result in > literature using v8.1 of MODELLER, and am trying to use 1H8C.pdb as > template to model 1D3Z.pdb structure. I followed the basic outline of > the scripts in the tutorial (and attached the scripts I used). At the > end of modelling, the homology model I obtain is about 9A cRMSD from > the target structure, whereas in the literature an earlier version of > MODELLER (v6.2) was used to achieve a closer structure to the template > (1.9A cRMSD). I was wondering whether there is any simple step I may > be missing.

Are you using the same sequence alignment as the authors of the published work? If the homology is not high there might be some subjective decisions made in the alignment regarding the position of loops/insertions that could significantly affect those parts of the structure. If they didn't publish their alignment you might want to write to them and ask for it.

Also, you might want to try comparing RMS values that exclude regions where the backbone of the template is disordered, since Modeller might generate significantly different conformations for those regions depending on the random number seeds.