Biomass consists of a complex mixture of different components, of which protein potentially has a high added value for biorefinery. In this study, protein extractability of different types of biomass, mostly by-products, was analyzed. Protein yield obtained from a three step extraction using alkali was correlated to biomass chemical composition through Partial Least Square (PLS) regression. The results showed that protein extractability depended crucially on the type of biomass used. Protein from cereals and legumes were highly extracted, compared to other materials. High protein extractability coincides with the biological function of protein as a storage protein, as opposed to functional protein. Protein extraction was furthermore correlated to the composition of the biomass. Especially cellulose and oil hamper extractability of protein whereas lignin has no significant influence, suggesting that alkaline treatment removed lignin sufficiently.