Documentation

RIN Generation Method

This method for RIN generation incorporates non-covalent interactions between main and side chains of amino acid residues as well as the strength of these interactions. RINs are generated from a PDB molecular structure file by performing the following steps:

1. All hydrogens are added to the protein structure by the Reduce program [1].
2. The non-covalent interactions are identified using the Probe program [2].
3. A residue interaction network is generated by the RINerator package.

Most of the protein structures (about 90 %) are determined by X-ray crystallography. Unfortunately, this technique is not capable of inferring the coordinates of hydrogen atoms, and they are thus not included in the PDB file of a protein structure. However, hydrogen atoms are crucial for identifying non-covalent interactions between amino acids. Therefore, the well-known program Reduce is used to add the hydrogen atoms to the PDB structure file using local geometry.

The resulting PDB structure file has explicit hydrogen atoms and serves as input for the program Probe. Probe identifies the contacts between amino acids in a protein by evaluating their atomic packing using small-probe contact dot surfaces. A small virtual probe (typically 0.25 Å) is rolled around the van der Waals surface of each atom and an interaction (contact dot) is detected if the probe touches another non-covalently bonded atom (see Figure 1). Overlapping van der Waals shells of non-polar atoms and hydrogen bonds are indicated by spikes, i.e., lines drawn from the dot position to the contact midplane, along the atom radius. The spike length is denoted by lsp and dots without spikes have length 0.

The contact dots and spikes can be used to quantitatively measure the molecular goodness-of-fit of the packing interactions. Hydrogen bonds and other overlaps are quantified by the volume of overlap. The contact scores are evaluated per dot, and are then summed for each atom pair. The non-overlapping van der Waals contacts are quantified by an error-function weighting, which awards close contacts a higher score than distant or significantly overlapping ones, but slight overlaps are still favorable in net effect. The combined score is a weighted sum of the weighted non-overlapping van der Waals contacts, the volume of hydrogen bonds and the volume of overlaps. The Probe program summarizes these scoring data for all parts of an entire structure and can output a file with information for every atom or residue.

The last step of the network generation method is accomplished by the Python package RINerator. It creates an undirected weighted network with multiple edges, as shown in Figure 2. The nodes represent the amino acid residues of the protein, whereas the links between them represent the non-covalent interactions identified by Probe. The edges are labeled with an interaction type and subtype. Possible types are interatomic contact (cnt), hydrogen bond (hbond), overlap (ovl), and combined (combi), while the subtypes indicate interactions between main chains (mc) and side chains (sc) of the amino acids. Each edge is weighted with the respective score for the interacting amino acid residues as computed by Probe, and the weight is proportional to the strength of the interaction. The resulting RINs are stored in a file format that can be loaded into Cytoscape. The file format specifications are described in detail here.