No. | Method | Procedure | Advantages | Disadvantages | Reference |
---|---|---|---|---|---|

1 | Transducer coefficients [ΔLog(τ/K_{A})] | Fit DR curves to operational model/calculate ratios of Log(τ/K_{A}) within each pathway (Δlog(τ/K_{A}); calculate Log bias as ΔΔLog(τ/K_{A}) across pathways | Sensitive/system independent | Possible difficulties in fitting Black/Leff operational model to DR curves | Kenakin et al. (2012) |

Quantifiable to yield a scale | |||||

High throughput | |||||

Statistically verifiable/bounded | |||||

Accommodates full agonists | |||||

Theoretically sound | |||||

2 | ΔLog (max/EC_{50}) | Calculate Log(max/EC_{50}) of DR curves; calculate ΔLog(max/EC_{50}) within each pathway; calculate ΔΔLog(max/EC_{50}) across pathways for Log bias | Sensitive/system independent | Cannot be used if n << 1 and/or intrinsic activity <30% | Kenakin (2017) |

Simple/quantifiable yielding a scale | |||||

High throughput | |||||

Statistically verifiable/bounded | |||||

Accommodates full agonists | |||||

No difficulties fitting curves to operational model | |||||

Theoretically sound | |||||

3 | Relative efficacy [ΔLog(τ)] | Fit DR curves for both functional pathways with a single measurement of affinity (binding) to obtain efficacy (τ values); express bias as relative Log(τ) values | Sensitive/system independent | Not theoretically sound (ignores possible Δaffinity) | Rajagopal et al. (2011) |

Quantifiable to yield scale | Binding affinity must be used for full agonists | ||||

High throughput | Possible difficulties in fitting Black/Leff operational model to DR curves | ||||

Statistically verifiable/bounded | |||||

4 | RA | Equiactive concentrations from DR curves for two agonists are compared with null methods in each pathway to yield RA values. Bias is then evaluated through Δlog(RA) values | Sensitive/system independent | Not high throughput (need dual agonist simultaneous comparison) | Ehlert (2008) |

Quantifiable to yield a scale | Works best with divergent curves (i.e., full vs. partial agonist) | ||||

Statistically variable/bounded | |||||

No difficulties fitting curves to operational model | |||||

Theoretically sound | |||||

5 | Method of Barlow, Scott, and Stephenson | Double-reciprocal plot of equiactive concentrations of two agonists in a functional system are used to yield a ratio of efficacy and affinity in each pathway. These ratios are then compared across pathways to yield estimate of bias | Sensitive/system independent | Not high throughput (need dual agonist simultaneous comparison) | Barlow et al. (1967) |

Quantifiable to yield scale | Needs divergent curves (i.e., full vs. partial agonist) | ||||

Theoretically sound | Double-reciprocal plots often skewed and yield erroneous parameters | ||||

No difficulties fitting curves to operational model | |||||

6 | Trajectory and rank order | Maximal responses to a range of agonists plotted for each pathway to determine trajectory relationship defining system bias. Outliers (either from plot itself or rank order) identified as biased | Model independent | No quantifiable scale | Onaran et al. (2017) |

Sensitive/system independent | Requires a range of agonist intrinsic activities to define trajectory | ||||

No difficulties fitting curves to operational model | Ignores affinity (may miss bias in some compounds) | ||||

Theoretically sound |

DR, dose response.