graph_analysis allows calculating connected components on
the network by using opion -c. We specified the topology
with the -r option to be able to map the graph nodes and
edges onto the structure. Using the -ub option, we wrote a
PDB file containing the reference structure along with the
degree for each hub residue in the b-factor column. Likewise,
option -cb allows to write a PDB file containg the connected
component number for each residue in the b-factor column.
4. Calculation of shortest communication paths. We used the
graph_analysis program to identify pathways of structural communication between a specific pair of residues. By default,
graph_analysis calculates all simple paths between a given pair
of residues up to a certain length. We first ran graph_analysis
with the -r and -a options alone so that it would print a list of
node names, which are used to specify which residues we will
calculate paths between. We then ran graph_analysis again with
the -p option so that it would try to calculate paths between the
residues specified by options -r1 and -r2 (see below for details),
using the node names as detailed above. By default, the maximum path length is 3. No paths were present with this length
and the program reported a minimum path length of 5, which
corresponds to the length of the shortest paths. Finally, we ran
the same command line this time specifying a length of 5 with
option -l. The found paths were printed to the standard output.
We used option -d to save each path as an independent subgraph (i.e., adjacency matrix).
5. Visualization of connected components on the protein structure. We used the xPyder plug-in to visualize the connected
components on the reference structure, as can be seen in
Fig. 3b. This was performed by loading the topology PDB
structure in PyMOL, loading the adjacency matrix in the xPyder plug-in, then generating the graph and calculating the
connected components in the Graph analysis tab, and finally
plotting them one by one changing the plotting color in the
main tab every time.
6. Visualization of hubs on the protein structure. The hubs were
subsequently visualized by loading the PDB file obtained with
the -ub option into the PyMOL software. We used the putty
b-factor in PyMOL representation that changes the thickness
and color of residues according to the values in the b-factor
column. Similarly, the color scale ranges from yellow to red as
the value increases.
By performing the graph analysis on the center of mass PSN in
CypA, we found 15 connected components, the biggest of which
were composed by 58, 19, and 11 residues, with the first connected
component comprising most of the secondary structure elements in
the protein and representing a large part of the protein core
Interaction Networks with PyInteraph
167
the network by using opion -c. We specified the topology
with the -r option to be able to map the graph nodes and
edges onto the structure. Using the -ub option, we wrote a
PDB file containing the reference structure along with the
degree for each hub residue in the b-factor column. Likewise,
option -cb allows to write a PDB file containg the connected
component number for each residue in the b-factor column.
4. Calculation of shortest communication paths. We used the
graph_analysis program to identify pathways of structural communication between a specific pair of residues. By default,
graph_analysis calculates all simple paths between a given pair
of residues up to a certain length. We first ran graph_analysis
with the -r and -a options alone so that it would print a list of
node names, which are used to specify which residues we will
calculate paths between. We then ran graph_analysis again with
the -p option so that it would try to calculate paths between the
residues specified by options -r1 and -r2 (see below for details),
using the node names as detailed above. By default, the maximum path length is 3. No paths were present with this length
and the program reported a minimum path length of 5, which
corresponds to the length of the shortest paths. Finally, we ran
the same command line this time specifying a length of 5 with
option -l. The found paths were printed to the standard output.
We used option -d to save each path as an independent subgraph (i.e., adjacency matrix).
5. Visualization of connected components on the protein structure. We used the xPyder plug-in to visualize the connected
components on the reference structure, as can be seen in
Fig. 3b. This was performed by loading the topology PDB
structure in PyMOL, loading the adjacency matrix in the xPyder plug-in, then generating the graph and calculating the
connected components in the Graph analysis tab, and finally
plotting them one by one changing the plotting color in the
main tab every time.
6. Visualization of hubs on the protein structure. The hubs were
subsequently visualized by loading the PDB file obtained with
the -ub option into the PyMOL software. We used the putty
b-factor in PyMOL representation that changes the thickness
and color of residues according to the values in the b-factor
column. Similarly, the color scale ranges from yellow to red as
the value increases.
By performing the graph analysis on the center of mass PSN in
CypA, we found 15 connected components, the biggest of which
were composed by 58, 19, and 11 residues, with the first connected
component comprising most of the secondary structure elements in
the protein and representing a large part of the protein core
Interaction Networks with PyInteraph
167