Determining the nutrient digestibility of a new ingredient is considered the first step to evaluate its suitability for use in aquaculture feeds. Thraustochytrids are considered a promising alternative to the partial replacement of fish oil. These microorganisms have a high long-chain fatty acid content, mainly docosahexaenoic acid (22:6-n-3). In this study, the apparent digestibility coefficient (ADC) of macronutrients provided by the biomass of whole and disrupted cells of Aurantiochytrium acetophilum in diets for Atlantic salmon (Salmo salar) were evaluated. A total of 44 fish per tank (34.4 ± 0.5 g) were used; the fish were randomly assigned to fifteen 100-L tanks. The apparent nutrient digestibility of the two forms of A. acetophilum biomass (whole and disrupted cells) was calculated by comparison with a reference diet apparent nutrient digestibility and in two levels of combination (10 and 20%) with each ingredient to be tested. The fish were fed the reference diet and the test ingredient diets for 25 days. Fish faeces were collected using decantation columns and analyzed to determine the ADC of dry matter, protein, lipids and gross energy. Two-way factorial analysis of variance indicated that the digestibility of dry matter, lipids and energy were significantly affected (P < 0.05) by cell treatment, inclusion percentage and the interaction between both factors. Compared to the whole-cell format, the physical treatment applied to rupture the cell walls favoured the ADC of dry matter, energy and lipids in juvenile Atlantic salmon when 20% of the ingredient was incorporated. On the contrary, inclusion of whole-cell biomass at 20% contributed to a significant decrease (P < 0.05) in the ADC of all the nutrients evaluated. The results suggest that cellular disruption of biomass from thraustochytrid with hard cell walls such as A. acetophilum will be advisable before inclusion in a feed formula. This aspect should be specially considered in non-extruded feeds for fish