Distances measured using Type Ia supernovae (SNIa) are an excellent way of constraining the equation of state of the dark energy. However, the brightnesses of these SNIa have variabilities due to internal mechanisms or environmental effects that have to be accounted for. Empirical models of the SNIa spectral energy distribution can be employed to standardize these objects. A new model called SUpernova Generator And Reconstructor (SUGAR) was proposed by Léget et al. (submitted), who showed with the help of the data from the Nearby Supernova Factory, that two new sources of variability have to be taken into account. In this presentation, we aim to check the impact of these new variabilities on cosmological analysis by fitting supernovae light curves with SUGAR and by using these parameters for standardization. Furthermore, we also intend to analyse the impact of SUGAR in the correlation between Hubble residuals and host galaxy properties.