Table 2 Open in new tab Several...

Year	Method	Target(s)/inhibitor(s)	Prediction accuracy,^a%	Sample set, mutants	Ref
De novo prediction approaches (structure-based prediction approaches)
2008	MD with free energy/variability value	HIV protease/5 inhibitors	75–100			33	[43]
2016	Alpha shape model with MD simulation	Epidermal growth factor receptor/1 inhibitor		90		30	[44]
2016	Homology modeling, docking, MD simulation and MM/GBSA calculation	Topoisomerase I/7 inhibitors			91.6^b	72	[45]
2017	Docking with Molecular Mechanics Generalized Born Surface Area (MM/GBSA) calculation	ABL kinase/5 inhibitor	70			55	[46]
2018	AIMMS	HIV protease/5 inhibitors HIV reverse transcriptase/4 inhibitors Influenza virus neuraminidase/1 inhibitor bc1 complex/1 inhibitor, DHFR-TS/2 inhibitors Aldose reductase/1 inhibitor, kinase/3 inhibitors	79–100	83–97		311
Data-driven and comprehensive prediction approaches
2005	Linear regression, docking with dynamics, sequence consensus	HIV protease/6 inhibitors	73–86			1792	[47]
2008	Docking and multivariate statistical procedures	HIV reverse transcriptase/55 inhibitors from NIAID HIV protease/51 inhibitors from NIAID	65–80			530	[47]
2013	Logistic regression/random forests	HIV reverse transcriptase/5 inhibitors	67.2–84.0			3133	[48]
2016	Biophysics-based fitness	Dihydrofolate reductase/1 inhibitor			85	21	[49]

Year

Method

Target(s)/inhibitor(s)

Prediction accuracy,^a%

Sample set, mutants

Ref

Criterion 1

Criterion 2

Criterion 3

De novo prediction approaches (structure-based prediction approaches)

2008

MD with free energy/variability value

HIV protease/5 inhibitors

75–100

[43]

2016

Alpha shape model with MD simulation

Epidermal growth factor receptor/1 inhibitor

[44]

2016

Homology modeling, docking, MD simulation and MM/GBSA calculation

Topoisomerase I/7 inhibitors

91.6^b

[45]

2017

Docking with Molecular Mechanics Generalized Born Surface Area (MM/GBSA) calculation

ABL kinase/5 inhibitor

[46]

2018

AIMMS

HIV protease/5 inhibitors
HIV reverse transcriptase/4 inhibitors
Influenza virus neuraminidase/1 inhibitor
bc1 complex/1 inhibitor, DHFR-TS/2 inhibitors
Aldose reductase/1 inhibitor, kinase/3 inhibitors

79–100

83–97

311

Data-driven and comprehensive prediction approaches

2005

Linear regression, docking with dynamics, sequence consensus

HIV protease/6 inhibitors

73–86

1792

[47]

2008

Docking and multivariate statistical procedures

HIV reverse transcriptase/55 inhibitors from NIAID
HIV protease/51 inhibitors from NIAID

65–80

530

[47]

2013

Logistic regression/random forests

HIV reverse transcriptase/5 inhibitors

67.2–84.0

3133

[48]

2016

Biophysics-based fitness

Dihydrofolate reductase/1 inhibitor

[49]

Year	Method	Target(s)/inhibitor(s)	Prediction accuracy,^a%	Sample set, mutants	Ref
De novo prediction approaches (structure-based prediction approaches)
2008	MD with free energy/variability value	HIV protease/5 inhibitors	75–100			33	[43]
2016	Alpha shape model with MD simulation	Epidermal growth factor receptor/1 inhibitor		90		30	[44]
2016	Homology modeling, docking, MD simulation and MM/GBSA calculation	Topoisomerase I/7 inhibitors			91.6^b	72	[45]
2017	Docking with Molecular Mechanics Generalized Born Surface Area (MM/GBSA) calculation	ABL kinase/5 inhibitor	70			55	[46]
2018	AIMMS	HIV protease/5 inhibitors HIV reverse transcriptase/4 inhibitors Influenza virus neuraminidase/1 inhibitor bc1 complex/1 inhibitor, DHFR-TS/2 inhibitors Aldose reductase/1 inhibitor, kinase/3 inhibitors	79–100	83–97		311
Data-driven and comprehensive prediction approaches
2005	Linear regression, docking with dynamics, sequence consensus	HIV protease/6 inhibitors	73–86			1792	[47]
2008	Docking and multivariate statistical procedures	HIV reverse transcriptase/55 inhibitors from NIAID HIV protease/51 inhibitors from NIAID	65–80			530	[47]
2013	Logistic regression/random forests	HIV reverse transcriptase/5 inhibitors	67.2–84.0			3133	[48]
2016	Biophysics-based fitness	Dihydrofolate reductase/1 inhibitor			85	21	[49]

Table 2

Open in new tab

Several computational drug resistance prediction methods compared with AIMMS

Year	Method	Target(s)/inhibitor(s)	Prediction accuracy,^a%			Sample set, mutants	Ref
Year	Method	Target(s)/inhibitor(s)	Criterion 1	Criterion 2	Criterion 3	Sample set, mutants	Ref
De novo prediction approaches (structure-based prediction approaches)
2008	MD with free energy/variability value	HIV protease/5 inhibitors	75–100			33	[43]
2016	Alpha shape model with MD simulation	Epidermal growth factor receptor/1 inhibitor		90		30	[44]
2016	Homology modeling, docking, MD simulation and MM/GBSA calculation	Topoisomerase I/7 inhibitors			91.6^b	72	[45]
2017	Docking with Molecular Mechanics Generalized Born Surface Area (MM/GBSA) calculation	ABL kinase/5 inhibitor	70			55	[46]
2018	AIMMS	HIV protease/5 inhibitors HIV reverse transcriptase/4 inhibitors Influenza virus neuraminidase/1 inhibitor bc1 complex/1 inhibitor, DHFR-TS/2 inhibitors Aldose reductase/1 inhibitor, kinase/3 inhibitors	79–100	83–97		311
Data-driven and comprehensive prediction approaches
2005	Linear regression, docking with dynamics, sequence consensus	HIV protease/6 inhibitors	73–86			1792	[47]
2008	Docking and multivariate statistical procedures	HIV reverse transcriptase/55 inhibitors from NIAID HIV protease/51 inhibitors from NIAID	65–80			530	[47]
2013	Logistic regression/random forests	HIV reverse transcriptase/5 inhibitors	67.2–84.0			3133	[48]
2016	Biophysics-based fitness	Dihydrofolate reductase/1 inhibitor			85	21	[49]

Year	Method	Target(s)/inhibitor(s)	Prediction accuracy,^a%			Sample set, mutants	Ref
Year	Method	Target(s)/inhibitor(s)	Criterion 1	Criterion 2	Criterion 3	Sample set, mutants	Ref
De novo prediction approaches (structure-based prediction approaches)
2008	MD with free energy/variability value	HIV protease/5 inhibitors	75–100			33	[43]
2016	Alpha shape model with MD simulation	Epidermal growth factor receptor/1 inhibitor		90		30	[44]
2016	Homology modeling, docking, MD simulation and MM/GBSA calculation	Topoisomerase I/7 inhibitors			91.6^b	72	[45]
2017	Docking with Molecular Mechanics Generalized Born Surface Area (MM/GBSA) calculation	ABL kinase/5 inhibitor	70			55	[46]
2018	AIMMS	HIV protease/5 inhibitors HIV reverse transcriptase/4 inhibitors Influenza virus neuraminidase/1 inhibitor bc1 complex/1 inhibitor, DHFR-TS/2 inhibitors Aldose reductase/1 inhibitor, kinase/3 inhibitors	79–100	83–97		311
Data-driven and comprehensive prediction approaches
2005	Linear regression, docking with dynamics, sequence consensus	HIV protease/6 inhibitors	73–86			1792	[47]
2008	Docking and multivariate statistical procedures	HIV reverse transcriptase/55 inhibitors from NIAID HIV protease/51 inhibitors from NIAID	65–80			530	[47]
2013	Logistic regression/random forests	HIV reverse transcriptase/5 inhibitors	67.2–84.0			3133	[48]
2016	Biophysics-based fitness	Dihydrofolate reductase/1 inhibitor			85	21	[49]

^aVarious criteria used to represent the prediction accuracy in the references cited. Criterion 1: percentage of the correctly predicted resistance into two categories. Criterion 2: percentage of the correctly predicted drug resistance into multiple levels. Criterion 3: correlation coefficient (R²) for the linear correction between the computational predictive result and the corresponding experimentally derived result. ^bOnly six mutant samples with experimental activity value were used here to calculate the accuracy.

^cNIAID, National Institute of Allergy and Infectious Diseases.

This Feature Is Available To Subscribers Only