Consideration of molecular weight during compound selection in virtual target-based database screening

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Virtual database screening computationally selects compounds based on interactions with biological targets.
• Current methods often favor high molecular weight compounds due to energy scoring biases.

Purpose of the Study:

• To address the bias towards high molecular weight compounds in virtual screening.
• To propose and evaluate normalization strategies for energy-based screening.

Main Methods:

• Developed normalization strategies based on the total number of heavy atoms in screened compounds.
• Analyzed molecular weight distributions of selected compounds compared to the original database.

Main Results:

• The proposed normalization effectively accounts for molecular weight bias.
• Achieved molecular weight distributions that can be lower or similar to the original database.
• Demonstrated computational efficiency of the normalization approach.

Conclusions:

• Normalization strategies can eliminate bias towards higher molecular weight compounds in virtual screening.
• This approach is expected to improve the success rate of computer-aided drug design.
• The method offers flexibility for selecting lead-like or drug-like compounds.