Loss-of-function mutations in the gene coding for filaggrin are the single most important risk factor for development of atopic dermatitis and associated allergic rhinitis and asthma. Filaggrin is enzymatically degraded to natural moisturizing factor (NMF) in the stratum corneum (SC). In vivo Raman spectra of human skin can be used to quantify the NMF concentration in SC and thereby identify carriers of a loss-of-function mutation in the gene coding for filaggrin, which results in decreased NMF content. Here, we demonstrate that strongly reduced Raman spectral information is sufficient to make this differentiation. This is an important step towards development of a dedicated diagnostic device of reduced complexity, size and cost as compared to current state-of-the-art Raman equipment. A genetic algorithm was used to select the spectral regions needed to classify skin based on normal or reduced NMF content in SC. Using the NMF content based on full spectral information as gold standard, only four Raman regions were required to create a linear discriminant analysis model that can differentiate between skin with low NMF and skin with normal NMF with a prediction accuracy of 93 %.