In the cold regions of protostars, atoms and molecules in the gas phase can stick on dust grains and form theso-called ice mantle. Understanding the chemical composition of the mantle and its diversity from source to source requires a systematic comparison with laboratory-measured ice spectra. In fact, telescope observations analyzed in light of experimental data have allowed concluding that H2O[1,2] ice is the major component of the interstellar ices and also that molecules can be mixed in ices (e.g., CO:CH3OH[3]). Much more information about the composition, morphology and structure of interstellar and circumstellar ices will be obtained in the next decades with the James Webb Space Telescope and the ground-based Extremely Large Telescope. While a large variety of laboratory data is needed to interpret these observations, the question that remains is which data best fit the observations. Previous works[4,5] report high degeneracy in the fits of astronomical observations with ice spectra. In this context, we created the ENIIGMA fitting tool[6], which is a public Python package to search for the global minimum solution by combining laboratory-measured ice spectra. The code handles a large amount of laboratory-measured ice spectra and uses genetic modelling algorithms to fit astronomical data. The solutions are assessed using two-dimensional χ2 maps, recurrence plots and histogram analysis. This statistical analysis allows us to quantify the degeneracy of the solutions and to derive robust ice column densities. By using ENIIGMA, we have fitted the broad-band spectrum of the protostar Elias 29 between 2.5 and 20 μm. We have derived the ice column densities for the major ice components, and we have found evidence for the presence of ethanol (CH3CH2OH) in the ices.