ZDISTANCES CV

[Efstathia Mntz]

Hello, 

I have been using ZDISTANCE CV to calculate the number of contacts between the CAs of a protein and a reference atom in a silica surface. I also use a 2nd CV which is the COM distance of the protein and the same reference silica atom of the surface. In the beginning of the simulation (WT-MetaD), in the higher COM distance i observe an increase in the number of contacts (sharp peak in the 2nd plot in Fig1), which is wrong.

I have checked the CA-Si distances with MDAnalysis for the trj and it seems that for some CAs  the calculated distances are not corresponding (Fig2).

Fig1

Fig2

This is my plumed.dat:

MOLINFO MOLTYPE=protein STRUCTURE=.pdb
WHOLEMOLECULES ENTITY0=1-44600 ENTITY1=45001-47501
INCLUDE FILE=protein_ca.txt

#COM definitions
protein: COM ATOMS=45001-47501

# surface_COM atom from the top layer 39437
d: DISTANCE ATOMS=39437,protein COMPONENTS

ps_contacts: ZDISTANCES GROUPA=39437 GROUPB=protein_ca LESS_THAN={RATIONAL R_0=0.38 D_0=0.0 NN=6 MM=12}

meta: METAD ARG=hb,d.z,ps_contacts.lessthan PACE=2500 HEIGHT=1.2 SIGMA=1.2,0.4,2  FILE=HILLS BIASFACTOR=20 TEMP=303.0 CALC_RCT GRID_MIN=0.0,0.0,0.0 GRID_MAX=160,7,180

uwall: UPPER_WALLS ARG=d.z AT=6.5 KAPPA=300
lwall: LOWER_WALLS ARG=d.z AT=0 KAPPA=150

bias: REWEIGHT_METAD TEMP=303

PRINT ARG=d.z,meta.bias,meta.rbias,meta.rct,bias,uwall.bias,ps_contacts.lessthan STRIDE=500 FILE=colvar 

The printed z components from PLUMED for these CAs at these frames are negative. 

How can this be possible? I have inspected the trajectory in VMD as well and it seems reasonable. I am not sure were this inconsistency comes from. 

Thank you for your help in advance!

Best, 

Efstathia