Automated screening of enzyme variants

A system comprising:

• one or more processors, system memory, and a nucleic acid synthesizer;
• wherein the one or more processors are configured to:
• (a) provide a structural model for each protein variant of a plurality of protein variants, wherein each structural model contains a three-dimensional computational representation of an active site of a protein variant, and wherein the plurality of protein variants comprises at least ten different variants;
• (b) for each protein variant of the plurality of protein variants, dock a computational representation of the ligand to the three-dimensional computational representation of the active site of the protein variant, wherein docking (i) generates a plurality of poses of the ligand in the active site, wherein a pose comprises a position or orientation of the ligand with respect to the active site of the protein variant, and (ii) identifies energetically favorable poses of the ligand in the active site, wherein an energetically favorable pose is a pose having an energy that is favorable for binding between the ligand and the protein variant;
• and
• (c) for at least one energetically favorable pose, determine that the pose meets one or more constraints for the ligand to undergo a particular interaction with the protein variant, the one or more constraints comprise one or more ranges of position, distance, angle, and torsion of the ligand with respect to the active site of the protein variant, a cofactor of the protein variant, or an ideally positioned native ligand in the active site of the protein variant, and thereby determine that the at least one energetically favorable pose is an active pose,
• wherein the nucleic acid synthesizer is configured to:
• synthesize at least one polynucleotide encoding at least a portion of at least one protein variant of the plurality of protein variants, wherein the at least one protein variant has an active site in which the ligand has the active pose, and wherein the at least one protein variant has desired activity and/or selectivity in its interaction with the ligand.

A computer program product comprising one or more computer-readable non-transitory storage media having stored thereon computer-executable instructions that, when executed by one or more processors of a computer system, cause the computer system to implement a method for virtual screening of enzyme variants, the instructions comprising:

• (a) providing a structural model for each protein variant of a plurality of protein variants, wherein each structural model contains a three-dimensional computational representation of an active site of a protein variant, and wherein the plurality of protein variants comprises at least ten different variants;
• (b) for each protein variant of the plurality of protein variants, docking a computational representation of the ligand to the three-dimensional computational representation of the active site of the protein variant, wherein docking (i) generates a plurality of poses of the ligand in the active site, wherein a pose comprises a position or orientation of the ligand with respect to the active site of the protein variant, and (ii) identifies energetically favorable poses of the ligand in the active site, wherein an energetically favorable pose is a pose having an energy that is favorable for binding between the ligand and the protein variant;
• (c) for at least one energetically favorable pose, determining that the pose meets one or more constraints for the ligand to undergo a particular interaction with the protein variant, the one or more constraints comprise one or more ranges of position, distance, angle, and torsion of the ligand with respect to the active site of the protein variant, a cofactor of the protein variant, or an ideally positioned native ligand in the active site of the protein variant, and thereby determining that the at least one energetically favorable pose is an active pose;
• and
• (d) controlling a nucleic acid synthesizer to synthesize one or more polynucleotides encoding at least one protein variant or portions thereof, wherein the at least one protein variant has an active site in which the ligand has the active pose, and the at least one protein variant has desired activity and/or selectivity in its interaction with the ligand.