Molecular docking method and apparatus based on coherent ising machine

A method for screening a lead drug molecule compound based on a molecular docking simulation performed by a coherent Ising machine (CIM), comprising:

• selecting a drug molecule from a drug molecule library to be screened, and constructing a three-dimensional (3D) molecular structure diagram based on the selected drug molecule to obtain a ligand graph;
• selecting a receptor from a receptor library, and constructing an internal pseudo-atom point diagram of a receptor target based on the selected receptor molecule to obtain a receptor graph;
• constructing a ligand-receptor similarity graph based on the ligand graph and the receptor graph, wherein the ligand-receptor similarity graph comprises a plurality of vertices and edges, any vertex of the plurality of vertices comprises a point in the ligand graph and a point in the receptor graph, and there is an edge between any two vertices;
• and
• determining whether each edge exists by determining whether there is an edge between any two vertices of the plurality of vertices by a distance between the two vertices comprising
• determining a first distance between a point in a ligand graph at a first vertex and a point in a ligand graph at a second vertex;
• determining a second distance between a point in a receptor graph at the first vertex and a point in a receptor graph at the second vertex;
• and
• determining that, when the first distance and the second distance are less than a preset distance, the edge between any two vertices is added;
• constructing a pharmacophore model based on the ligand-receptor similarity graph, wherein the pharmacophore model is performed by a CIM configured to calculate a maximum weight clique between the selected drug molecule and the selected receptor to screen a drug molecule compound, wherein the maximum weight clique between the selected drug molecule and the selected receptor is determined according to a following equation:
• ∑
• i
• =
• 1
• m
• ∑
• j
• =
• i
• n
• w
• (
• i
• ,
• j
• )
• ⁢
• x
• (
• i
• ,
• j
• )
• -
• K
• 1
• ⁢
• ∑
• i
• ,
• k
• =
• 1
• m
• ∑
• j
• ,
• l
• =
• 1
• n
• w
• (
• i
• ,
• j
• )
• ,
• (
• k
• ,
• l
• )
• ⁢
• x
• (
• i
• ,
• j
• )
• wherein w(i,j) denotes the weight of a vertex x(i,j) in the ligand-receptor similarity graph, depending on a type of an atom/group corresponding to i,j;
• w(i,j),(k,l) denotes the weight of an edge between x(i,j) and x(k,l) in the ligand-receptor similarity graph;
• and K1 is coefficient of the maximum weight clique;
• and
• constructing a 3D structure from the pharmacophore model to determine whether the selected drug molecule could inhibit the activity of the selected receptor, wherein the 3D structure is the 3D pose of the ligand-receptor structure in the bound state.