Defatted grape seed flour (DGSF), a byproduct of the grape seed oil production, was investigated as a potential source of protein-enriched food ingredients with functional properties. The operation conditions of the alkaline extraction followed by the acid precipitation method were studied by response surface methodology, aiming to maximize the protein extraction yield. These conditions were temperature 50 °C, solvent/meal ratio 1:15, pH 11, and 150 min extraction time in two sequential extraction cycles, achieving a product with 32.63 ± 0.07% (w/w) d.m. of protein content. A second extract with 38.17 ± 1.34% (w/w) d.m. of protein was obtained under the same conditions, but using an additional centrifugation step before the acid precipitation. The amino acid profile, solubility, water and oil retention capacity, and emulsifying and foaming properties of the DGSF and the obtained extracts were compared. Both extracts could be used as food ingredients with different nutritional and technological purposes.

Fish by-catches, along with other fish side-streams, were previously used as raw material for the production of fishmeal and fish oil but appropriate handling allows their use in more valuable options. The aim of this research was to valorize undersized hake (Merluccius merluccius) as a model of using fish by-catch from the Bay of Biscay to produce protein hydrolysates with bioactivities. Six enzymes, with different proteolytic activities (endo- or exoproteases) and specificities, were tested to produce protein hydrolysates. Products obtained with an endoprotease of serine resulted in the most promising results in terms of protein extraction yield (68%), with an average molecular weight of 2.5 kDa, and bioactivity yield (antioxidant activity = 88.5 mg TE antioxidant capacity/g fish protein; antihypertensive activity = 47% inhibition at 1 mg/mL). Then, process conditions for the use of this enzyme to produce bioactive products were optimized using Box–Behnken design. The most favorable process conditions (time = 2 h, solids = 50% and enzyme/substrate = 2% with respect to protein) were scaled up (from 0.5 L to 150 L reactor) to confirm laboratory scale and model forecasts. The results obtained in the pilot-scale testing matched the outcomes predicted by the model, confirming the technical viability of the proposed process.