**Background:** Overfishing and the need to feed a growing population demand better utilization of marine by-catches, by-products, and undervalued fish species. Converting these into protein powders is a sustainable approach, but consumer acceptance is limited by undesirable sensory properties such as fishy odor and taste. While prior studies have examined fish proteins, they have not combined sensory analysis with gas-chromatography–mass spectrometry-olfactometry (GC-MS/O) to identify odor-active compounds, nor have they focused on commercial-scale products. This study aimed to characterize the chemical, functional, and sensory properties of commercial fish proteins to compare their suitability for human consumption and identify factors causing sensory limitations.
**Methods:** Six commercial fish protein powders were procured from European manufacturers: three fish concentrates (CONC1–3, from cod, saithe, haddock, and cod by-products), two hydrolysates (HYDR1 from whole blue whiting, HYDR2 from salmon filleting by-products), and one hydrolyzed fish collagen (COLL1). Proximate composition (protein, lipid, moisture, ash) was determined using Dumas combustion, gravimetric extraction, and muffle furnace methods. Amino acid profiles (total and free) were analyzed via IC-UV (ISO 13903:2005). Peptide molecular weight distribution was assessed by SDS-PAGE and size-exclusion chromatography (SEC). Fatty acid composition was measured by GC-FID after transesterification. Lipid oxidation was quantified via peroxide values (PV, ferric thiocyanate method) and thiobarbituric acid reactive substances (TBARS, as malondialdehyde, by UHPLC). Functional properties included nitrogen solubility, foaming capacity/stability, water holding capacity, fat binding capacity, and heat-induced gelation. Sensory profiling used generic descriptive analysis (GDA) with nine trained panelists evaluating 15 attributes on 0–10 unstructured line scales. Odor-active volatile compounds were identified by HS-SPME-GC-MS/O with detection frequency method and quantified as relative concentrations. Statistical analysis used one-way and two-way ANOVA with Tukey's post hoc (p<0.05).
**Key Results:** All samples were high in protein, but hydrolysates had the highest protein content. Concentrates contained more total lipids (e.g., CONC1 had higher lipid content than CONC2/3) and ash, while HYDR2 and COLL1 had remarkably low ash. Total amino acid content was similar across concentrates and hydrolysates, with adequate essential amino acids; COLL1 lacked tryptophan and had low essential amino acids except lysine. Free amino acids were highest in HYDR1. SEC showed concentrates had >10 kDa as the most prominent fraction; hydrolysates were dominated by 1–10 kDa peptides, with HYDR1 having the highest proportion of <1 kDa peptides. COLL1 had a distinct SEC profile with no compounds <1 kDa. Fatty acid analysis of concentrates showed n-3 proportions of 29.2% (CONC1), 34.2% (CONC2), and 40.9% (CONC3); DHA content was inversely related to oxidation. PVs were highest in CONC1 (15.3 meq/kg oil) and lowest in HYDR1; TBARS ranged from 100.5 nmol MDA/g (HYDR1) to lowest in COLL1. Functionally, hydrolysates had superior solubility (>90%) and foaming capacity (>200%), while concentrates had better water holding capacity. No samples formed gels at 15% concentration. Sensory evaluation clustered samples by processing method: hydrolysates were more bitter and metallic; concentrates were fishier and more sea/seaweed-like; COLL1 was the mildest and whitest. Bitterness correlated with free hydrophobic amino acids and small peptides (<10 kDa). Fishiness correlated with lipid oxidation-derived volatiles (hexanal, heptanal, (Z)-4-heptenal, 2,3-butanedione) and trimethylamine (TMA). GC-MS/O identified 33 compounds (20 identified); 46 odor-active compounds were considered including literature-derived ones. Key odorants included TMA, methional, 3-methylbutanal, and lipid oxidation products. Methanethiol, hydroxyacetone, and 2-heptanone were reported in fish proteins for the first time.
**Clinical Implications:** While this is a food science study without direct clinical outcomes, the findings have nutritional relevance. Fish protein concentrates are rich in essential amino acids and long-chain n-3 PUFAs (EPA/DHA), which are linked to cardiovascular, metabolic, and cognitive health benefits. However, the high levels of lipid oxidation products (TBARS, PVs) in some samples may reduce the nutritional quality of n-3 fatty acids and could contribute to oxidative stress if consumed in large quantities. The identification of processing methods that minimize lipid oxidation and off-flavor formation is critical for developing palatable, high-quality fish protein ingredients that can serve as sustainable protein sources for human nutrition, potentially addressing protein-energy malnutrition and reducing reliance on overfished species.