Role of moving average analysis for development of multi-target (Q)SAR models

Area of Science:

• Computational chemistry
• Medicinal chemistry
• Drug discovery

Background:

• Multi-target quantitative structure-activity relationship (mt-QSAR) approaches are crucial for in-silico drug design.
• Moving Average Analysis (MAA) is a machine learning technique known for its high predictive accuracy in QSAR modeling.

Purpose of the Study:

• To review the role of MAA in developing QSAR models for single, dual, or multi-target activities.
• To develop and validate novel mt-QSAR models for predicting anti-Plasmodium falciparum and anti-Trypanosoma brucei rhodesiense activities using MAA.

Main Methods:

• Utilized Moving Average Analysis (MAA) for developing mt-QSAR models.
• Assessed statistical significance using intercorrelation analysis, sensitivity, specificity, and Matthew's correlation coefficient.
• Validated the developed models using a test set.

Main Results:

• Successfully developed MAA-based mt-QSAR models for benzyl phenyl ether derivatives.
• Achieved high predictability ranging from 80% to 95% for anti-protozoal activity.
• Demonstrated model safety through high selectivity index values.

Conclusions:

• MAA is a justified approach for developing realistic and accurate mt-QSAR models.
• The developed models can aid in designing novel, potent, and safe anti-protozoal drugs.
• The models offer improved profiles for both anti-Plasmodium falciparum and anti-Trypanosoma brucei rhodesiense activities.