network types based on specific residue-residue interaction networks. Since long-range communication can happen through less
specific contacts, however, we also included a more general way to
generate and analyze the PSN which is based on distances between
centers of mass (cmPSN).
The primary function of PyInteraph is to compute, on conformational ensembles, three types of intramolecular interactions
which are thought to be the most important non-covalent interactions in the protein structure: hydrophobic contacts, hydrogen
bonds, and salt bridges. In PyInteraph, a hydrophobic contact is
identified when the center of mass of the side chain of two hydrophobic residues is found within a given distance cut-off, which is
5 A ˚ by default.
For salt bridges, the program is able to derive which charged
groups belonging to side chains and main chain are present,
depending on the topology. Charged groups are defined as those
groups of charged residues (as Asp, Glu, Lys, Arg, and His, plus
main-chain N- and C-termini) with a certain protonation state.
Considering one pair of charged groups at the time, all the distances between atom pairs belonging to them are calculated, and
they are selected as taking part in a salt bridge if at least one pair of
atoms is found at a distance shorter than 4.5 A ˚ by default.
A hydrogen bond is identified when a hydrogen and an acceptor atom are within 3.5 A ˚ and the donor-hydrogen-acceptor angle
is greater than 120
.
Each type of interaction is calculated independently and considered as a separate network at first. For each ensemble conformation, an interaction graph is calculated and two nodes are
connected by an edge if at least one interaction of the specified
type is found between the residues. The final network is constructed by counting, for each residue pair, the number of ensemble
conformations in which two residues were connected by an edge
over the total number of frames, times 100 to obtain a persistence
value. In this way, one network per interaction type is collected, in
which edges are weighted by the persistence of the given interaction
in the ensemble. Two residues are connected by an edge in the final
network if this value is above 0.0 or above a chosen significance
cut-off.
Merging these networks to account for the interactions of each
type is just a matter of keeping those edges that are present in at
least one network, thus generating the so-called intramolecular
interaction network (IIN).
PyInteraph also implements a knowledge-based potential [17]
for the calculation of energy networks. Briefly, it is based on a fourdistance description of interactions between specific atoms of side
chains, which were chosen as those having the largest number of
contacts in a dataset of high-resolution protein structures. The
potential is calculated as follows:
156
Matteo Lambrughi et al.
specific contacts, however, we also included a more general way to
generate and analyze the PSN which is based on distances between
centers of mass (cmPSN).
The primary function of PyInteraph is to compute, on conformational ensembles, three types of intramolecular interactions
which are thought to be the most important non-covalent interactions in the protein structure: hydrophobic contacts, hydrogen
bonds, and salt bridges. In PyInteraph, a hydrophobic contact is
identified when the center of mass of the side chain of two hydrophobic residues is found within a given distance cut-off, which is
5 A ˚ by default.
For salt bridges, the program is able to derive which charged
groups belonging to side chains and main chain are present,
depending on the topology. Charged groups are defined as those
groups of charged residues (as Asp, Glu, Lys, Arg, and His, plus
main-chain N- and C-termini) with a certain protonation state.
Considering one pair of charged groups at the time, all the distances between atom pairs belonging to them are calculated, and
they are selected as taking part in a salt bridge if at least one pair of
atoms is found at a distance shorter than 4.5 A ˚ by default.
A hydrogen bond is identified when a hydrogen and an acceptor atom are within 3.5 A ˚ and the donor-hydrogen-acceptor angle
is greater than 120
.
Each type of interaction is calculated independently and considered as a separate network at first. For each ensemble conformation, an interaction graph is calculated and two nodes are
connected by an edge if at least one interaction of the specified
type is found between the residues. The final network is constructed by counting, for each residue pair, the number of ensemble
conformations in which two residues were connected by an edge
over the total number of frames, times 100 to obtain a persistence
value. In this way, one network per interaction type is collected, in
which edges are weighted by the persistence of the given interaction
in the ensemble. Two residues are connected by an edge in the final
network if this value is above 0.0 or above a chosen significance
cut-off.
Merging these networks to account for the interactions of each
type is just a matter of keeping those edges that are present in at
least one network, thus generating the so-called intramolecular
interaction network (IIN).
PyInteraph also implements a knowledge-based potential [17]
for the calculation of energy networks. Briefly, it is based on a fourdistance description of interactions between specific atoms of side
chains, which were chosen as those having the largest number of
contacts in a dataset of high-resolution protein structures. The
potential is calculated as follows:
156
Matteo Lambrughi et al.