Cryo-EM: problem with RosettaCM using multiple template models as input

 Hi,

First of all, I am not sure about the category of my post, but the "Phenix/MR rosetta"  category seems to be the most relevant, considering the topic of this post.

I have been folllowing the cryo-EM tutorial from 2015 to produce a model, which would be derived from the fragments of multiple templates and nicely fit to cryo-EM map. However, all produced models are out of density. I understand that the purpose of this protocol is  to perform a more exhaustive sampling, but still several top-scored models are simply out of density, not even close. The problem is with both ccp4 and mrc maps.

Templates are aligned to the density. I checked in both COOT and Chimera, so it is not a problem of headers...I think.

Adding -relax::constrain_relax_to_start_coords flag with different values of of cst improve, but do not solve the problem.

Any suggestions? 

rosetta_scripts.linuxgccrelease \
 -database database \
 -in::file::fasta fasta.fasta \
 -parser::protocol xml.xml \
 -edensity::mapfile mrc.mrc \
 -edensity::mapreso 4.3 \
 -edensity::cryoem_scatterers \
 -relax:minimize_bond_angles \
 -relax:min_type lbfgs_armijo_nonmonotone \
 -relax:jump_move true \
 -relax:default_repeats 2 \
 -default_max_cycles 200 \
 -nstruct 50 \
 -out::suffix _ref2_$1 \

<ROSETTASCRIPTS> 
	<SCOREFXNS> 
		<stage1 weights="score3">
			<Reweight scoretype=atom_pair_constraint weight=0.25/> 
		</stage1>
		<stage2 weights="score4_smooth_cart"> 
			<Reweight scoretype=atom_pair_constraint weight=0.25/> 
		</stage2> 
		<fullatom weights="talaris2013_cart"> 
			<Reweight scoretype=atom_pair_constraint weight=0.25/> 
		</fullatom> 
	</SCOREFXNS> 
	<MOVERS>¨
		<Hybridize name=hybridize stage1_scorefxn=stage1 stage2_scorefxn=stage2 fa_scorefxn=fullatom batch=1>
			<Template pdb="thr_1.pdb" weight=1.0 cst_file="AUTO"/> 
			<Template pdb="thr_2.pdb" weight=1.0 cst_file="AUTO"/>
			<Template pdb="thr_3.pdb" weight=1.0 cst_file="AUTO"/>
			<Template pdb="thr_4.pdb" weight=1.0 cst_file="AUTO"/> 
		</Hybridize> 
	</MOVERS> 
	<PROTOCOLS> 
		<Add mover=hybridize/> 
	</PROTOCOLS> 
</ROSETTASCRIPTS> 

2.) Additionally, I am not sure whether the tutorial itself is up to date.  According to the tutorial, the "AUTO" flag should be "auto", but actually it should not be.

When I change to "auto" I get error = [ERROR] Unable to open constraints file: auto

3.) Finally, which flags specify rescoring againts density? Shouldn't we use any density scoring function? I mean, does the hybridizer follows any density data when building structures? The main script file also does not contain any density function.

thank you

Karl