Microbial oil, also known as single-cell oil (SCO), is a promising sustainable alternative for plant- and animal-derived oils and fats. Oleaginous micro-organisms can produce SCO amounts of more than 20% of their dry mass. This study aimed to evaluate if it is possible to predict the composition of a yeast biomass based on its infrared spectrum. The oleaginous yeast Yarrowia lipolytica was cultured in batch and fed-batch bioreactors to compose a collection of samples of varying oil and protein contents. Yeast samples were characterized using standard analytical techniques, e.g., the Soxhlet method for yeast oil determination and the Kjeldahl method for protein measurements. A Perkin Elmer System 2000 instrument was used to register the FT-IR spectra of yeast, as was a TQ Analyst and GRAMS IA software. The selected region was 4,000–450 cm⁻¹ with a resolution of 2 cm⁻¹. Ten scans were taken for the background spectrum and each sample. Within the scope of this work, two steps of experiments were carried out. The first step covered the recording of FT-IR spectra for freeze-dried defeated yeast samples enriched with microbial oil. The samples contained 0, 10, 20, 30 , 40 and 50% m/m oil. Registering and analyzing the spectra allowed for the preparation of discriminant models that used spectral data alone, with a high correlation coefficient. To validate the created discriminant model, an RMSEP parameter was calculated and reached 0.98. For the second step, the FT-IR spectra for different yeast samples containing varied amounts of storage lipids and proteins were registered. The samples fit the created model and led to the promising conclusion that the technique can be used for protein determination in yeast as well. Studies demonstrated FT- IR to be a fast, low-cost and reliable analytical technique to monitor the composition of yeast biomass.