Background:  Growth competition assays have been developed to quantify the relative fitness of HIV-1 mutants.

Methods:  In this study we develop mathematical models to describe viral/cellular dynamic interactions in the assay experiment from which new competitive fitness indices or parameters are defined. These indices include the log fitness ratio (LFR), the log relative fitness (LRF), and the production rate ratio (PRR).

Results:  From the population genetics perspective, we clarify the confusion and correct the inconsistency in the definition of relative fitness in the literature of HIV-1 viral fitness. The indices, LFR and LRF, are easier to estimate from our assay experimental data than the PRR, which was misleadingly defined as the relative fitness in recent HIV-1 research literature. Calculation and estimation methods based on 2 data points and multiple data points were proposed and were carefully studied.

Conclusions:  In particular, we suggest using both standard linear regression (the least squares method) and a measurement error model approach for more accurate estimates of competitive fitness parameters from multiple data points. The methods developed are generally applicable to similar growth competitive assays. A user-friendly computational tool is also developed and publicly available on the web at http://www.urmc.rochester.edu/bstools/vfitness/virusfitness.htm.