| 摘要 |
This invention is intended to perform the screening of the compound for efficiency to win reasonably. It is solved by the data processing method that the first space expressing coordinates of the first chemical substance group and the second space expressing coordinates of the second chemical substance group are defined, and the first chemical substance group is characterized by the first characteristic amount, and the second chemical substance group is characterized by the second characteristic amount, the coordinates in the said first space and the coordinates in the said second space are map-transformed by technique chosen from multivariate analysis technique, mechanic learning method and methods equivalent for those. To be concrete, interaction forms of a protein group and compound group are studied by machine learing as a statictic pattern of protein characteristic amount (biological information such as arrangement or the expression information) and compound charcteristic amound ((chemical compound information such as a structure, the physical property) of the compound, and it is the system which realized a interaction prediction based on it, and the hereby problems mentioned above were solved. |
