You are here

BluePrintBDRMover set_atom_type error

4 posts / 0 new
Last post
BluePrintBDRMover set_atom_type error

Hello Everyone,


I am trying to get the BluePrintBDR mover to work properly, my script so far is as follows:

#Here is generate a starting structure that contains only 1 VAL
temp = open('temp.pdb' , 'w')
temp.write('ATOM      1  N   VAL A  1       25.945   4.358  33.648  1.00 22.51           N  \nATOM      2  CA  VAL A  1       26.375   4.305  35.016  1.00 24.82           C  \nATOM      3  C   VAL A  1       27.860   4.146  35.064  1.00 17.93           C  \nATOM      4  O   VAL A  1       28.451   3.503  34.206  1.00 27.99           O  \nATOM      5  CB  VAL A  1       25.647   3.121  35.836  1.00 38.86           C  \nATOM      6  CG1 VAL A  1       24.936   2.161  34.876  1.00 40.73           C  \nATOM      7  CG2 VAL A  1       26.659   2.335  36.692  1.00 39.90           C  ')
pose = pose_from_pdb('temp.pdb')
#Here I attempt to build a full structure on the starting structure
scorefxn = pyrosetta.rosetta.core.scoring.ScoreFunctionFactory.create_score_function('ref2015_cst')
mover = pyrosetta.rosetta.protocols.fldsgn.BluePrintBDR()


Here are samples of my Blueprint file and constraints file:

#Constraints file sample
AtomPair CA 1 CA 2 GAUSSIANFUNC 3.83519 2.0 TAG
AtomPair CA 1 CA 3 GAUSSIANFUNC 6.91223 2.0 TAG
AtomPair CA 1 CA 4 GAUSSIANFUNC 10.1876 2.0 TAG

#Blueprint file sample
0 V EX R
0 V EX R
0 V EX R
0 V EX R
0 V LX R
0 V LX R
0 V LX R
0 V LX R


My first issue is that I get this error:

ERROR: Illegal attempt to score with non-identical atom set between pose and etable
	pose   atom_type_set: 'centroid'
	etable atom_type_set: 'fa_standard'

ERROR:: Exit from: /home/benchmark/T/rosetta.Glass/_commits_/main/source/src/core/scoring/etable/ line: 297


I tried to change the atom_set_type of  my pose to become fa_standard through pose.is_fullatom() but that did not work, I tried to find other ways to change the pose's atom type but I could not find a forum entry with an issue similar to mine. How can I solve this error?


The other thing, is my BDR mover setup OK? can you see majore problems with it that I should fix?


Thank you for your continuous help.

Post Situation: 
Sat, 2017-12-09 02:16

SInce the Blueprint mover is going through large backbone rearrangements, it's likely going through a centroid stage, rather than being fullatom. That is, even if you pass it a fullatom protein, it will convert the Pose to a centroid one.

So what you want to do is to use a centroid scorefunction to match the centroid structure that the pose is being converted to internally. The default scorefunction for this mover is `fldsgn_cen`, and unless you have reason to change it, I'd recommend leaving it as-is.

(Note that fldsgn_cen doesn't have any constraints terms turned on by default. If you're using constraints during the proceedure, you may need to turn them on. But I'd recommend turning them on from a base of fldsgn_cen.

Sat, 2017-12-16 10:43

Oh ok , that makes a lot of sense now.


The constraints are an essential part of the protocol, without them the mover does not result in any meaningful topology, so how can i turn on the constraints for the fldsgn_cen score function?

Sat, 2017-12-16 12:06

Simply set the relevant scoreterms to non-zero in your scorefunction prior to passing it to the mover. For example:

scorefxn.set_weight(rosetta.core.scoring.atom_pair_constraint, 1.0)


Sat, 2017-12-16 13:29