IVART Methodology

IVART Methodology

To be suitable for the IVART analysis, data is annotated with the WWARN tagging system, uploaded, and transformed into a standardised format. The IVART analysis process is detailed below.


Experimental data are converted to a percentage scale based on:

  • Top constraint (100% value): E(C0) defined as the mean effect across all wells on a plate which contain no drug.
  • Bottom constraint (0% value): Emin defined as the average effect over the two concentrations with the lowest mean effect for a particular drug.


Non-linear regression

A sigmoid, 4-parameter concentration-inhibition model is applied to the data based on established approaches (Le Nagard et al., 2011).


IC50 estimates and curve slopes


Core Criteria for summary analyses

Confidence in the IC50 value and slope, determined by linear regression, will be used to define a subset of results with a narrow confidence interval suitable for core analyses and reports according to the following core criteria:

  • If gamma is not 10, ratio of upper : lower 95% confidence intervals for IC50 must be less than 3 (and both confidence intervals must be positive).
  • If gamma is 10 (indicating a fixed, steep slope), ratio of E(C0) : Emin must be greater than 2, indicating acceptable growth (Basco 2007).

Gamma is fixed to 10 when non-linear regression fails to produce an acceptable curve, due to either a very steep slope or noisy data. For these curves, confidence intervals are not sensitive measures and criterion based on signal and noise is applied.

Range warnings

Range warnings are triggered when the range of drug concentrations used are not adequate to give a reliable IC50 for the specific isolate tested.

  • The ‘Range high’ warning is triggered when IC50 < lowest drug concentration.
  • The ‘Range low’ warning is triggered when (E(C0) – ECmax) > (1.05 x (E(C0) – Emin)). This is designed to pick up 'unfinished assays' where there is evidence that further inhibition would have occurred if a higher range of concentrations had been used.    



Analysis details are specified in the In Vitro Data Management and Statistical Analysis Plan


Le Nagard H, Vincent C, Mentre F, Le Bras J. Online analysis of in vitro resistance to antimalarial drugs through nonlinear regression. Comput. Methods Programs Biomed. 2011 Oct;104(1):10-8

Basco LK. Field application of in vitro assays for the sensitivity of human malaria parasites to antimalarial drugs. World Health Organization 2007.