Introduction
The global aquaculture industry has experienced rapid growth, now accounting for approximately half of the world’s seafood production (FAO, 2022). Driven by a growing human population and increased demand for seafood, aquaculture production is projected to reach 200 million tons globally by 2040 (Kobayashi et al., Reference Kobayashi, Msangi, Batka, Vannuccini, Dey and Anderson2015). However, this expansion is not without challenges, particularly the need to source large volumes of protein and lipids for aquafeeds. Fishmeal, a commodity derived from wild-caught forage fish such as anchovy and menhaden, has traditionally been the primary animal protein ingredient in aquaculture feeds. While fishmeal is highly effective at promoting growth in farmed fish, its overexploitation has raised environmental and economic concerns (Naylor et al., Reference Naylor, Goldburg, Primavera, Kautsky, Beveridge, Clay, Folke, Lubchenco, Mooney and Troell2000; Tacon and Metian, Reference Tacon and Metian2008). In many regions, inconsistent or weak regulatory frameworks governing forage-fish harvests have contributed to fluctuating supply, rising costs, and practices viewed as ecologically contentious. As a result, identifying alternative, reliable protein sources has become a top priority for the aquaculture industry.
Efforts to replace fishmeal with sustainable alternatives have focused on plant proteins, livestock by-products, yeast, algae, and insect-based meals (Bruce et al., Reference Bruce, Sindelar, Voorhees, Brown and Barnes2017; Øverland and Skrede, Reference Øverland and Skrede2017; van Huis, Reference van Huis2020). However, for many high-value, carnivorous fish species such as rainbow trout (Oncorhynchus mykiss), plant proteins alone may not suffice due to anti-nutritional factors and deficiencies in essential amino acids (Francis, Makkar and Becker, Reference Francis, Makkar and Becker2001; Gatlin et al., Reference Gatlin, Barrows, Brown, Dabrowski, Gaylord, Hardy, Herman, Hu, Krogdahl, Nelson, Overturf, Rust, Sealey, Skonberg, Souza, Stone, Wilson, R. and Wurtele2007). Among alternative protein sources, insect larvae have emerged as a promising solution due to their high nutritional value, environmental sustainability, and ability to be reared on organic waste materials (van Huis et al., Reference van Huis, Van Itterbeeck, Klunder, Mertens, Halloran, Muir and Vantomme2013; Sánchez-Muros, Barroso and Manzano-Agugliaro, Reference Sánchez-Muros, Barroso and Manzano-Agugliaro2014; van Huis, Reference van Huis2020).
Insect-based meals and their production methods, particularly from the black soldier fly (Hermetia illucens), have been extensively studied for use in aquafeeds due to their high protein content and versatility in assimilating diverse substrates (Bondari and Sheppard, Reference Bondari and Sheppard1981, Reference Bondari and Sheppard1987; Kroeckel et al., Reference Kroeckel, Harjes, Roth, Katz, Wuertz, Susenbeth and Schulz2012). Despite their potential, other insect species, such as the common housefly (Musca domestica), have received comparatively little attention. Housefly larvae are highly efficient at reducing organic material, globally ubiquitous, and adaptable to a variety of climates and substrates, making them well-suited for large-scale production in diverse agricultural systems (Barroso et al., Reference Barroso, de Haro, Sánchez-Muros, Venegas, Martínez-Sánchez and Pérez-Bañón2014; Henry et al., Reference Henry, Gasco, Piccolo and Fountoulaki2015; Miranda, Cammack and Tomberlin, Reference Miranda, Cammack and Tomberlin2020; Camperio et al., Reference Camperio, Suarez, Simonton, Paresky, Parodi and Benetti2025). Safe mass production, however, requires addressing key concerns such as fly escapement and the pathogen load associated with certain feed substrates. Even so, where regulations permit, the ability of insects to bioconvert organic wastes, from livestock manure to discarded food, into a high-value feed ingredient offers a meaningful dual benefit of waste remediation and nutrient recovery (Selvaraj and Won, Reference Selvaraj and Won2024; Sibinga, Won and Selvaraj, Reference Sibinga, Won and Selvaraj2025).
Our previous research characterized housefly larvae as a nutrient-rich feedstuff for animal agriculture, demonstrating favorable nutritional profiles with ~60% crude protein and high levels of essential amino acids, surpassing those of soybean meal and even fishmeal (Hussein et al., Reference Hussein, Pillai, Goddard, Park, Kothapalli, Ross, Ketterings, Brenna, Milstein, Marquis, Johnson, Nyrop and Selvaraj2017). These nutritional attributes indicate that housefly larvae have the potential to function as a high-quality protein ingredient for aquaculture feeds.
In this study, we evaluated the potential of housefly larvae meal as an alternative to fishmeal in the diets of rainbow trout, a common carnivorous aquaculture species. Diets were formulated to replace fishmeal with larvae meal at 50% and 100% inclusion levels. Growth performance, body composition, and gut microbiota were assessed over a 16-week feeding trial. This research provides insights into the viability of housefly larvae as a sustainable aquafeed ingredient, offering a practical solution to reduce reliance on fishmeal and enhance the sustainability of aquaculture.
Materials and methods
Larvae production
Adult houseflies (M. domestica) were propagated, and larvae were collected as previously described (Hussein et al., Reference Hussein, Pillai, Goddard, Park, Kothapalli, Ross, Ketterings, Brenna, Milstein, Marquis, Johnson, Nyrop and Selvaraj2017). Cattle manure was sourced biweekly from the Cornell College of Veterinary Medicine’s Teaching Dairy Barn (Ithaca, NY, USA) and stored at 4°C. Each week, 5 ml of semi-solid manure was placed in netted enclosures housing adult flies to serve as a substrate for egg deposition. After 24 h, the manure containing fly eggs was collected and transferred into containers (30 x 30 x 8 cm) filled with fresh cattle manure. These containers were incubated in a humidity-controlled tent at 28 ± 2°C and 65% humidity for 1 week.
Once larval activity was observed, the larvae and manure mixture was portioned into eight tins (30 x 10 x 5 cm), and additional cattle manure was added. The tins were placed back in the humidity tent on trays with raised edges to prevent larvae from escaping. Water (3–4 ml) was added to each tin daily, and the manure within was gently mixed to maintain moisture and aeration. Over the next 2 weeks, wandering pre-pupae that exited the tins into collection trays were gathered daily and immediately frozen for storage.
Formulation of experimental diets
Three experimental diets were formulated to replace fishmeal with housefly larvae meal as the protein source at 0% (FM100), 50% (FM50LM50), and 100% (LM100) inclusion levels. The formulation procedure was adapted from a feeding guide published by the University of Florida (Royes and Chapman, Reference Royes and Chapman2003). Fish oil inclusion levels were adjusted according to the formulation requirements. The final composition of each diet is detailed in Table 1.
Ingredients and formulation for the different experimental diets: Fishmeal diet (FM100) and Larvae meal diets (FM50LM50 and LM100)
* Formulations based on a dry-matter basis.
Stored frozen larvae were dried at 50°C and powdered before diet formulation to ensure consistency in dry matter content. Dried ingredients (fishmeal, larvae meal, cornstarch, soybean meal, and wheat gluten) were thoroughly mixed. Supplemental ingredients (mineral and vitamin mixes) were then added, and the mixture was homogenized again. Gelatin (15 g per 1.5 kg of final diet) was dissolved in water, boiled and gradually added to the dry mixture in equal parts. The resulting mixture, maintained at approximately 80–90 °C, was extruded into pellets. Pellets were dried overnight at 50 °C, cooled to room temperature, and stored at −20°C until used in the feeding experiments. Commercial diet used in this study was the Finfish Starter 2 mm pellets (50% protein, 15% fat) from Zeigler Bros., Inc. (Gardners, PA, USA).
Diet nutritional content analysis
Nutritional content analysis was conducted on three samples of each formulated diet used in the feed trial, performed at Brookside Labs Inc. (New Bremen, OH, USA). Moisture content was determined by drying each sample at 40.6°C for 14 h. Crude fat content was analyzed using AOAC Method 920.39 (ether extraction), while crude protein was measured using AOAC Method 962.09. Mineral content was assessed via nitric acid/hydrogen peroxide digestion followed by inductively coupled plasma (ICP) analysis. Fiber content was determined using AOAC Methods 973.18 (acid detergent fiber) and 2002.04 (neutral detergent fiber). Values were compared between the formulated diets using ANOVA and post-hoc Tukey’s test to identify nutritional differences.
Rainbow trout feed trial
Rainbow trout fingerlings (15.06 ± 0.49 g) were obtained from the Bath New York State Hatchery and housed in the Aquatic Animal Facilities at Cornell University. All experimental procedures were reviewed and approved by the Cornell Institutional Animal Care and Use Committee. Fish were randomly assigned to 70 L flow-through tanks at 12 ± 2 °C with a stocking rate of 10 fish per tank, except for one tank with 14 fish. Fish were fed a 2.0 mm Finfish Starter diet (50–15 slow sinking; Zeigler, Inc.) at 1% of their body weight once daily during a 3-week acclimation period, followed by the growth performance trial for 16 weeks (Figure 1), allowing them to approximately triple in size.
Schematic of the study design, diet treatments, and assessments. Schematic of the study design, treatments, and assessments. Fingerling rainbow trout (~15 g) were acclimatized for 3 weeks in flow-through 70 l tanks while being fed a commercial diet (COMM; Zeigler Finfish Starter). Following acclimatization, fish were fed one of three test diets: Fishmeal-only (FM100), a mixed diet (FM50LM50), or Larvae meal-only (LM100) for 16 weeks. Tank weights were measured approximately every 4 weeks throughout the study. At termination, fish were analyzed for growth performance, body composition, intestinal health, and gut microbiota.
During the growth experiment, tanks were randomly assigned to one of four dietary groups: [1] FM100 (3 tanks of 10 fish), [2] FM50LM50 (3 tanks of 10 fish), (3) LM100 (3 tanks of 10 fish), or [4] Zeigler commercial diet (1 tank with 14 fish). The commercial diet group was included only as a reference conventional industry feed and was not used in statistical analysis. Fish were batch-weighed by tank approximately every 4 weeks during the trial.
At the end of the experiment, trout were individually euthanized with 300 mg/l tricaine methanesulfonate (MS-222 from Western Chemical, Inc) buffered with sodium bicarbonate, followed by pithing and decapitation. They were then weighed and their length measured. Blood was collected from the caudal vein before the coelomic cavity was opened for evaluation. Fat deposition in the coelomic cavity was scored on a scale of 1 to 4 [modified from (Gjerde, Reference Gjerde1989)], where 1 indicated minimal fat, 2 indicated moderate fat along the intestine without covering it, 3 indicated substantial fat partly covering the intestine, and 4 indicated excessive fat covering the entire intestines. The liver was excised, weighed, and used to calculate the hepatosomatic index (liver weight/body weight ratio). The intestine was removed, its contents extruded using forceps and stored at −80°C for microbiome analysis, and the tissue fixed for histology. The left-side fillet was removed, weighed, and used to calculate the total meat yield (fillet weight/body weight).
PCR sexing
Sex determination was conducted using PCR to detect the male sex-determining gene, sdY, following (Yano et al., Reference Yano, Nicol, Jouanno, Quillet, Fostier, Guyomard and Guiguen2013). Genomic DNA was extracted from abdominal muscle tissue by digesting samples in Lysis Buffer with proteinase K, followed by NaCl precipitation and ethanol-based DNA isolation (Morohaku et al., Reference Morohaku, Pelton, Daugherty, Butler, Deng and Selvaraj2014). DNA was resuspended in DEPC-treated water for downstream use. PCR amplification targeted sdY (Primers: forward 5’-CCCAGCACTGTTTTCTTGTCTCA-3′, reverse 5’-CTGTTGAAGAGCATCACAGGGTC-3′) and the positive control gene 18S using (Primers: forward: 5’GTTCGAAGACGATCAGATACCGT-3′, reverse 5’-CCGCATAACTAGTTAGCATGCCG-3′). Reactions were performed in a 20 μl volume with standard PCR reagents as previously described (Tu et al., Reference Tu, Morohaku, Manna, Pelton, Butler, Stocco and Selvaraj2014). PCR products were visualized on a 1.5% agarose gel.
Fatty acid analysis of muscle
Five individual muscle samples (~200 mg of caudal fillet) from each group were homogenized, and lipids were extracted (Hara and Radin, Reference Hara and Radin1978) to determine fatty acid composition of the consumable filet. Extracted lipids were methylated overnight at 40°C in 1% methanolic sulfuric acid (Christie, Reference Christie1982), and subsequently transmethylated using a modified protocol as described (Chouinard et al., Reference Chouinard, Corneau, Saebø and Bauman1999). The resulting fatty acid methyl esters (FAME) were dissolved in heptane and stored at −20°C until analyses. FAME were quantitatively analyzed using a HP 5890 series II GC-FID with a BPX 70 column (length: 60 m, inner diameter: 0.32 mm, film: 0.25 μm; Hewlett Packard, Palo Alto, CA, USA). Hydrogen was used as a carrier gas. Structural identification was performed using gas chromatography-covalent adduct chemical ionization tandem mass spectrometry (GC-CACI-MS/MS) as previously described (Park et al., Reference Park, Kothapalli, Reardon, Lawrence, Qian and Brenna2012). Response factors were calculated using an equal-weight FAME mixture standard (68A; Nu-Chek Prep, Inc., Elysian, MN, USA), and all GC analyses were performed in triplicate.
Blood chemistry panel
Blood samples (five randomly selected fish/group) were collected from rainbow trout at the end of the feeding trial to assess blood chemistry parameters. Blood was drawn from the caudal vein and collected in heparinized tubes. The samples were immediately placed on ice until processing. Plasma was separated by centrifugation at 1500 × g for 10 min and stored at −80°C until analysis. Biochemical analytes were measured using a fully automated biochemical analyzer (Animal Health Diagnostic Center, Cornell University College of Veterinary Medicine) (Dawson et al., Reference Dawson, DeFrancisco, Mix and Stokol2011). The following parameters were analyzed: sodium, potassium, chloride, calcium, phosphate, glucose, total protein, aspartate aminotransferase (AST). Triglyceride concentration was estimated via levels of turbidity indicated as lipemia. Reference intervals for each parameter were determined based on previous studies (Manera and Britti, Reference Manera and Britti2006).
Statistical analyses of growth parameters
Statistical analyses were performed using a linear mixed-effects model to evaluate the effects of diet on measured outcomes. Tank was treated as a random effect to account for potential within-tank variation, while diet was included as a fixed effect to test for differences between treatment groups. This approach ensures that tank-level clustering does not bias the results and that variability across tanks is appropriately controlled. Weights were recorded as tank averages (n = 3) for each of the diet groups at weeks 0, 4, 8, 11, and 14 weeks, and measured as individual weights at termination (week 16). Sample sizes were: FM100, n = 25; FM50LM50, n = 24; LM100, n = 27; COMM, n = 14. Based on a 12% coefficient of variation in final body weight of genetically uniform stocks, a priori power calculations indicated that approximately 20 fish per diet group, distributed across replicate tanks, would be sufficient to achieve at least 80% power to detect ~10% diet-related differences in final weight at α = 0.05. Statistical significance was assessed using analysis of variance (ANOVA) on the model outputs, with post-hoc pairwise comparisons performed where applicable. All analyses were conducted in R (version 4.1.2) using the lme4 and lmerTest packages. Results are presented as means ± standard error of the mean (SEM), with statistical significance set at P < 0.05.
Intestinal histology
The intestine samples (five randomly selected fish/group) were collected and processed as scrolls prior to fixation in 4% buffered formaldehyde for histology (Jimenez et al., Reference Jimenez, Stilin, Morohaku, Hussein, Koganti and Selvaraj2022). Fixed samples were paraffin embedded and thin sectioned (4 μm) prepared on glass slides. Sections were stained with hematoxylin and eosin for histomorphological examination. Slides were scanned to a digital format (Aperio Scanscope, Leica), and examined for any histopathological changes. Each slide was evaluated for signs of inflammatory reactions in the intestinal tract using a semi-quantitative scoring system. The following parameters were assessed: (a) general morphology of the mucosal folds; (b) microvilli structure; (c) position of the nucleus within epithelial cells; (d) position and shape of cytoplasmic vacuoles; (e) presence and quantification of Goblet cells (average per mucosal fold); and (f) morphology and thickness of the lamina propria. Each parameter was assigned a score from 1 to 4, with 1 representing the lowest and 4 the highest inflammatory response in enteritis as established for fish (Urán et al., Reference Urán, Schrama, Rombout, Taverne-Thiele, Obach, Koppe and Verreth2009; Penn et al., Reference Penn, Bendiksen, Campbell and Krogdahl2011).
Analysis of intestinal microbiota
For microbiota analysis (six randomly selected fish/group), distal intestinal content samples (hind gut region) were processed for bacterial DNA extraction using the MoBio PowerSoil DNA Isolation Kit (MoBio, Carlsbad, CA, USA) and quantified with the Qubit dsDNA BR Assay Kit (Life Technologies, Grand Island, NY, USA) and Qubit 3.0 Fluorometer. Equal molar amounts of DNA from each sample were pooled to construct metagenomic libraries.
Fusion primers with Illumina adaptor sequences and gene-specific primers targeting the V3/V4 region of the 16S rRNA gene (342F/806R) were used for PCR amplification. No template negative controls were included for each run. PCR products (~532 bp) were verified on a 1% agarose gel, excised, and purified with the QIAquick Gel Extraction Kit and AMPure XP magnetic beads. Illumina indexes were added using the Nextera XT Index Kit, and PCR was performed with KAPA HiFi HotStart ReadyMix. Purified products were quantified, diluted to 1 ng/μl, and pooled for sequencing on the Illumina MiSeq platform (2 × 300 bp paired-end protocol).
The 16S rRNA sequences were pre-processed using MICCA v1.5.0. (Albanese et al., Reference Albanese, Fontana, De Filippo, Cavalieri and Donati2015). Reads were assembled with FLASH (Magoč and Salzberg, Reference Magoč and Salzberg2011), primers were trimmed with Cutadapt (Martin, Reference Martin2011), and sequences were quality filtered based on expected error rates, length, and ambiguous bases. USEARCH grouped sequences into OTUs at 97% similarity and removed chimeras. Taxonomy was assigned using the RDP classifier with Greengenes and UNITE fungal databases (Kõljalg et al., Reference Kõljalg, Larsson, Abarenkov, Nilsson, Alexander, Eberhardt, Erland, Høiland, Kjøller, Larsson, Pennanen, Sen, Taylor, Tedersoo, Vrålstad and Ursing2005; DeSantis et al., Reference DeSantis, Hugenholtz, Larsen, Rojas, Brodie, Keller, Huber, Dalevi, Hu and Andersen2006b). Multiple sequence alignment was performed with NAST (DeSantis et al., Reference DeSantis, Hugenholtz, Keller, Brodie, Larsen, Piceno, Phan and Andersen2006a), and a phylogenetic tree was constructed.
Alpha diversity (PD whole tree, Chao1 index, Shannon entropy) and beta diversity (weighted and unweighted UniFrac) were calculated, with principal coordinates analysis visualized using the EMPeror tool (Vázquez-Baeza et al., Reference Vázquez-Baeza, Pirrung, Gonzalez and Knight2013). OTU tables and metadata were compiled in BIOM format for downstream analysis in QIIME v1.9.1. (Caporaso et al., Reference Caporaso, Kuczynski, Stombaugh, Bittinger, Bushman, Costello, Fierer, Peña, Goodrich, Gordon, Huttley, Kelley, Knights, Koenig, Ley, Lozupone, McDonald, Muegge, Pirrung, Reeder, Sevinsky, Turnbaugh, Walters, Widmann, Yatsunenko, Zaneveld and Knight2010).
Results
Nutritional composition of experimental diets
The experimental diets FM100, FM50LM50, and LM100 were formulated to be isonitrogenous and isocaloric, differing only in their protein sources (Table 1). Analysis after production revealed that the formulated diets were comparable with low levels of moisture FM100 (3.32 ± 0.12%), LM100 (4.21 ± 0.08%) and FM50LM50 (3.71 ± 0.06%), and consistently high in dry matter content. The nutritional composition of the experimental diets (Table 2) showed both similarities and differences across the fishmeal-only (FM100), larvae meal-only (LM100), and mixed diets (FM50LM50). Crude protein content was comparable between FM100 and FM50LM50, with LM100 slightly lower but still within acceptable ranges for fish feed (FAO, 2025). Fat content was similar across all diets ensuring sufficient energy density.
Nutritional content analysis for the different experimental diets: Fishmeal diet (FM100), and Larvae meal diets (FM50LM50 and LM100)
* Data are presented as mean ± SEM on dry-matter basis (n = 3/experimental diet).
Note: a,b,c Different alphabets in superscripts indicate significant differences between diets (p < 0.05).
Differences were observed in several nutritional categories (Table 2). LM100 had significantly higher fiber content compared to FM100 and FM50LM50 (p < 0.05). FM100 had the highest levels of calcium and phosphorus, however, all groups remained within the recommended ranges for aquafeeds. Potassium was higher in LM100 compared to FM100 and FM50LM50). Trace minerals, such as zinc and manganese, were elevated in LM100 compared to the other diets. Importantly, while there were statistically significant differences between diets, all nutrient levels fell within the dietary requirements for trout (FAO, 2025).
Growth performance
Mean total biomass between treatments remained comparable across sampling intervals (three tanks/diet). At the conclusion of the trial (16 weeks), the FM100 diet group had the highest mean weight ± SEM (53.30 ± 3.46 g; n = 27), followed by the FM50LM50 diet group (45.04 ± 5.08 g; n = 24) and the COMM diet group (43.54 ± 3.45 g; n = 14), which were nearly identical. The LM100 had the lowest mean weight (38.13 ± 2.98 g; n = 27), and was significantly different from FM100, but not the other groups (Figure 2a). Body length measurements at the conclusion of the trial for the FM100, LM100, FM50LM50 and COMM diet groups were not different (Figure 2b).
Growth of Rainbow trout fingerlings in fishmeal (FM100) Larvae meal (FM50LM50 and LM100) and commercial (COMM) diets. (a) Weights were recorded as average tank weights for each of the diet groups at weeks 0, 4, 8, 11, and 14 weeks, and precisely measured as individual weights at termination (week 16). Graph shows weight increases for each of the diet group until termination. No significant differences in weight were observed between groups at any of the intermediary time points. At termination (week 16), FM100 was significantly different from LM100 (p < 0.05; indicated by different alphabets). (b) Body length indicative of skeletal development was measured for individual fish at termination (week 16). Graph shows average body length for each of the diet groups. No significant differences in body length were observed between groups. Sample sizes were: FM100, n = 25; FM50LM50, n = 24; LM100, n = 27; COMM, n = 14; Data are presented as mean ± SEM.
Body composition analysis
The growth performance and body composition of rainbow trout fed the different diets were evaluated based on muscle ratio (fillet yield), hepatosomatic index (HSI), and fat score (Figure 3). The muscle ratio, expressed as the percentage of fillet weight relative to body weight, showed no significant differences across the dietary groups (Figure 3a). All groups exhibited similar muscle growth efficiency, with muscle ratios clustering around 40%. The HSI was also comparable across different diet groups (Figure 3b). HSI values ranged between 1% and 1.2% on average, with no statistically significant differences observed. Fat deposition in the coelomic cavity, evaluated using a fat score scale (1–4), also showed no significant differences among the diet groups (Figure 3c). Similar distributions were observed for the different diets, with most fish scoring between 2 and 3, indicative of moderate fat deposition.
Growth performance evaluated through production characteristics in fishmeal (FM100) Larvae meal (FM50LM50 and LM100) and commercial (COMM) diets. Each characteristic was measured for individual fish at termination (week 16) of the feeding study. (a) Fillet percent yield (muscle ratio %) indicative of muscle growth efficiency was similar with no significant differences between the different diet groups. (b) The hepatosomatic index (HSI) indicative of energy metabolism and health was similar with no significant differences between all the diet groups. (c) Violin plots showing fat scores indicative of energy storage was also similar with no significant differences between the different diet groups. Representative images show the fat score scale (1–4) showing the reference levels of stored fat in the coelomic cavity. Data are presented as mean ± SEM.
Fatty acid analysis of muscle tissue revealed that major fatty acids were consistent across diets (Table 3). Saturated fatty acid levels, including 14:0 and 16:0, were comparable among groups, with slight increases in FM50LM50 diet compared to FM100 and LM100 diets. Unsaturated fatty acids exhibited more pronounced differences; the n-6 polyunsaturated linoleic acid (18:2n-6) was significantly higher in FM100 compared to LM100 and FM50LM50 diets (p < 0.05). The proportion of EPA (20:5n-3) and DHA (22:6n-3) showed an increasing trend in LM100 and FM50LM50 diets, that is perhaps an impact of increased menhaden fish oil levels in the larvae meal-containing diet formulations.
Muscle fatty acid composition for the different experimental diet groups: Fishmeal diet (FM100), and Larvae meal diets (FM50LM50 and LM100)
* Fatty acid analysis was conducted on 5 representative samples of muscle fillets from each group. Data are presented as mean ± SEM.
Note: a,b,c Different alphabets in superscripts indicate significant differences between diets (p < 0.05).
Gender influence on growth parameters
Male fish were identified through PCR amplification of the male-specific sdY gene using target primers. Agarose gel analysis of the PCR products revealed 56 males and 31 females, distributed randomly across treatments (Figure 4a). Gender was subsequently incorporated into terminal growth analysis for the different diets to assess potential biases, but no significant effects of gender on final body weight (Figure 4b) or body length (Figure 4c) were detected for the age and time period under study. During euthanasia, gonadal maturation was not observed in either sex.
Influence of end point morphometrics by the sex of rainbow trout. (a) Sex of individual rainbow trout was determined based on PCR detection of the male-specific sdY gene. Representative DNA electrophoresis gel showing amplifications specific to male fish for sdY. Amplification of the 18S rRNA as internal control for all fish DNA samples is shown in a separate gel with positive amplification in both male and female samples. (b) Graph showing the sex ratio within the randomly assigned fish in the different diet groups. (c) Violin plots showing the distribution and probability density for terminal body weights (week 16) in the different diet groups compared between males and females. No significant differences were detected, suggesting minimal influence of fish sex within the studied age group. (d) Violin plots for terminal body length (week 16) in the different diet groups compared between males and females. No significant differences were detected, suggesting minimal influence of fish sex within the studied age group.
Blood biochemistry analysis
Blood plasma analysis indicated that most biochemical parameters were consistent across dietary treatments, reflecting stable fish health (Table 4). Sodium, potassium, and chloride levels remained within physiological ranges, with no significant differences between groups. Phosphate levels were significantly lower in LM100 compared to FM100 and FM50LM50 diets, suggesting potential differences in phosphorus metabolism or bioavailability might occur at higher levels of larvae meal inclusion. Calcium levels were slightly reduced in LM100 compared to FM100 and FM50LM50 diets, though they remained within normal physiological limits. The other parameters, including total protein, glucose, and AST levels, showed no significant differences between groups. Additionally, the blood lipid levels were similar across groups. Lipemia values were marginally lower in the larvae meal-containing diets compared to FM100 diet, though the differences were not significant.
Blood chemistry analysis of fish for the different experimental diet groups: Fishmeal diet (FM100), and Larvae meal diets (FM50LM50 and LM100)
* Data are presented as mean ± SEM (n = 17–23/group).
Note: a,b,c Different alphabets in superscripts indicate significant differences between diets (p < 0.05).
Intestinal health
Features of intestinal tissue histological sections, including intact mucosal folds, villi structures, and epithelial cell nuclei, were scored on a semi-quantitative scale on its entire length (Figure 5). The average scores were 1.2 for LM100 diet, 1.1 for FM50LM50, FM100 and COMM diets, with no notable differences observed between groups. All histopathological parameters assessed were within expected normal ranges across the dietary groups. General morphology of the mucosal folds appeared intact, with no structural abnormalities observed in any groups. Villi structures were well-preserved, with uniform length and organization. The position of the nucleus within epithelial cells was consistently basal, as typical for healthy intestinal epithelium. Goblet cells, responsible for mucosal secretion, were evenly distributed and assessed to be at similar levels per mucosal fold across all groups. The lamina propria showed normal morphology and thickness, with no signs of inflammation, edema, or other pathological alterations.
Histological assessment of Intestinal health in fishmeal (FM100) Larvae meal (FM50LM50 and LM100) and commercial (COMM) diets. Representative histological sections of anterior and posterior intestinal tissue from fish fed the four experimental diets (FM100, LM100, FM50LM50, and COMM) are shown. Key features across all diets include intact mucosal folds, well-preserved villi structures, and consistent basal positioning of epithelial cell nuclei across all dietary groups. Goblet cells were evenly distributed, indicating uniform mucus production and barrier integrity. The lamina propria exhibited normal morphology and thickness, with no signs of inflammation, edema, or other pathological alterations. These findings highlight the absence of diet-induced adverse effects on intestinal health.
Intestinal microbiota
The gut microbiome analysis revealed significant differences in microbial composition among the dietary groups, with distinct clustering patterns observed in hierarchical analysis (Figure 6a). The larvae meal-containing diets (LM100 and FM50LM50) clustered together, whereas the fishmeal-only (FM100) and commercial diet (COMM) groups exhibited a more divergent and separate alignment. The heatmap analysis (Figure 6b) confirmed the dominance of Clostridia in larvae meal diets and the COMM diet, while the FM100 demonstrated greater microbial diversity, including contributions from Actinobacteria and Bacilli.
Intestinal microbiome composition across different dietary treatments: fishmeal (FM100) and larvae meal (FM50LM50 and LM100). (a) Hierarchical clustering of microbiome profiles by the unweighted pair group method with arithmetic mean (UPGMA) algorithm shows distinct clustering patterns. FM100 and COMM diets exhibit closer alignment, while LM100 and FM50LM50 diets cluster separately. Scale bar indicates 0.04 distance threshold. (b) Heatmap showing the relative abundance of bacterial taxa across the dietary groups. The larvae meal diets (LM100 and FM50LM50) are enriched with Clostridia, while the FM100 diet show greater diversity, including contributions from Actinobacteria and Bacilli. The COMM diet shows intermediary diversity between the FM100 and larvae meal diets. (c) Summary table of the relative abundance of dominant microbial taxa at the phylum, order, and family levels for each dietary group. The phylum Firmicutes dominates in the LM100, FM50LM50 and COMM diet groups, while Actinobacteria is more prominent in FM100 groups. At the order level, Clostridiales are notably higher in LM100, FM50LM50 and COMM diets, whereas Bacillales and Lactobacillales are variable across different diets.
At the phylum level (Figure 6c), Firmicutes dominated in the LM100 (95%), FM50LM50 (97%) and COMM (89%) groups, while Actinobacteria contributed a greater proportion in FM100 (24%). At the order level, Clostridiales showed a marked increase in LM100 (56%), FM50LM50 (58%) and COMM (53%) diets, compared to FM100 (15%) diet. In contrast, Bacillales were most abundant in FM100 (29%) and less prominent in the larvae meal and COMM groups. Similarly, Lactobacillales, known for their role in gut health and immune modulation, varied across diets but were highest in the COMM group (30%). At the family level, Peptostreptococcaceae, dominated the LM100 (47%) and FM50LM50 (51%) groups. Bacillaceae were consistently present across all diets but showed the greatest relative abundance in LM100 diet; Enterococcaceae were more prominent in the FM50LM50 diet. Overall, compared to the microbiota diversity in the fishmeal diet, the larvae meal and COMM diets seem to exhibit a shift toward bacterial groups involved in protein and lipid fermentation.
Discussion
The aquaculture industry is under increasing pressure to identify sustainable and renewable feed ingredients to support the growing demand for fish production, driven by the limitations of wild fisheries used to make fishmeal and the environmental impacts associated with intensive reduction fisheries (Tacon and Metian, Reference Tacon and Metian2008; Naylor et al., Reference Naylor, Hardy, Buschmann, Bush, Cao, Klinger, Little, Lubchenco, Shumway and Troell2021). This study evaluated housefly larvae meal as an alternative to conventional fishmeal in rainbow trout diets, focusing on its effects on growth performance, body composition, and the gut microbiome. The collective findings indicate that larvae meal supported physiological performance, tissue composition, and gut health comparable to that achieved using a typical commercial diet, underscoring its viability as a practical component of aquafeeds.
Moreover, housefly larvae have a high protein content, favorable amino acid profiles, and can be reared on a wide range of organic substrates, including agricultural and food waste streams (Sánchez-Muros, Barroso and Manzano-Agugliaro, Reference Sánchez-Muros, Barroso and Manzano-Agugliaro2014; Hussein et al., Reference Hussein, Pillai, Goddard, Park, Kothapalli, Ross, Ketterings, Brenna, Milstein, Marquis, Johnson, Nyrop and Selvaraj2017; van Huis and Oonincx, Reference van Huis and Oonincx2017). Using industrial or agricultural waste streams as rearing substrate presents dual benefits by reducing organic waste and providing an environmentally friendly protein source for aquaculture. The efficient conversion of low-value substrates into high-quality protein, combined with the larvae’s scalability and adaptability, underscores its potential as a sustainable alternative to fishmeal (Henry et al., Reference Henry, Gasco, Piccolo and Fountoulaki2015; Lock, Arsiwalla and Waagbø, Reference Lock, Arsiwalla and Waagbø2016; Selvaraj and Won, Reference Selvaraj and Won2024; Sibinga, Won and Selvaraj, Reference Sibinga, Won and Selvaraj2025).
Diet justification and caveats
Insects represent a natural part of the diet for many fish species, including salmonids, which consume arthropods in their freshwater and oceanic phases (Kawaguchi et al., Reference Kawaguchi, Miyasaka, Genkai-Kato, Taniguchi and Nakano2007; Rodger et al., Reference Rodger, Wolf, Starks, Burroughs and Brewer2021). The use of insect meal in aquafeeds therefore closely mimics the natural diet of common cultivars like trout (Henry et al., Reference Henry, Gasco, Piccolo and Fountoulaki2015). The nutritional profile of insect meal depends on the quality of the substrate used for larval rearing, however. While larvae raised on nutrient-rich byproducts like food waste can potentially enhance lipid and amino acid content (St-Hilaire et al., Reference St-Hilaire, Cranfill, McGuire, Mosley, Tomberlin, Newton, Sealey, Sheppard and Irving2007), even those reared on lower quality substrates like cattle manure offer the dual benefit of livestock waste mitigation and protein production. Nutritional deficiencies in such larvae, such as low omega-3 PUFA levels, can be addressed through supplementation, as demonstrated by the use of menhaden oil in the larva meal diet (Hussein et al., Reference Hussein, Pillai, Goddard, Park, Kothapalli, Ross, Ketterings, Brenna, Milstein, Marquis, Johnson, Nyrop and Selvaraj2017).
All diets could therefore be nutritionally formulated to meet salmonid requirements (Cho, Reference Cho1992; FAO, 2025), with crude protein and lipid levels adjusted to account for differences between the protein sources. Interestingly, the FM100 diet outperformed the commercial Zeigler diet used as a reference, underscoring that the fishmeal-based diet in this study is a higher benchmark than what might typically be used in modern commercial aquafeeds that typically use plant-based proteins like soy to offset fishmeal. This observation highlights the need for careful consideration when comparing larvae meal diets to fishmeal-based controls, as the assumed fishmeal benchmarks in this study seemed to surpass the performance standard of the commercial diet. It should therefore be noted that the FM100 treatment was included to provide a clear experimental contrast between fish-based and larvae-based protein sources, and does not reflect a nutritionally or economically practical aquafeed formulation. Although the exact proportion of fishmeal and plant protein fillers in different feed brands is variable, future assessments of larvae meal use in commercial production settings should arguably focus on replacing the fraction of animal protein in the diet coming specifically from fishmeal, while also including a plant protein component to more accurately represent modern aquafeeds.
Growth and body composition
Protein is the most expensive ingredient and a critical nutrient in fish diets, playing a key role in promoting optimal growth and maintaining health. Over the 16-week growth trial, juvenile rainbow trout receiving larvae meal-based proteins exhibited comparable growth metrics to those fed fishmeal or commercial diets. At termination, body length, fillet yields and fat scores did not differ significantly among diet groups. Final body weight was statistically lower in the LM100 group relative to FM100, but FM50LM50 and the commercial diet did not differ from FM100, indicating that partial replacement of fishmeal with larvae meal supported equivalent growth to this experimental benchmark. No significant differences in growth metrics were observed between male and female trout, likely due to the absence of gonadal maturation during the juvenile growth phase (Kause et al., Reference Kause, Ritola, Paananen, Mäntysaari and Eskelinen2003).
These outcomes underscore the feasibility of replacing fishmeal with larvae meal without compromising juvenile trout performance. Although fishmeal-only diets may confer a long-term growth advantage, such formulations are no longer practical in commercial aquaculture; FM100 was included here solely as a high-biological value experimental benchmark without interference from plant proteins commonly used in modern feeds. The FM50LM50 diet produced slightly higher weight gain than the LM100, but was not statistically significant. A longer production-cycle trial would be required to determine whether these trends persist. Beyond protein contributions, dietary lipid composition also influences fish growth. Carnivorous fish, including trout, require preformed long-chain omega-3 PUFAs because of limited Δ5-desaturase activity and a poor capacity to convert α-linolenic acid (Hagve, Christophersen and Dannevig, Reference Hagve, Christophersen and Dannevig1986). Because housefly larvae reared on manure substrates are naturally low in omega-3 fatty acids (Hussein et al., Reference Hussein, Pillai, Goddard, Park, Kothapalli, Ross, Ketterings, Brenna, Milstein, Marquis, Johnson, Nyrop and Selvaraj2017), fish oil supplementation was necessary to ensure adequate PUFA levels. This supplementation likely accounted for most of the EPA and DHA deposition observed in fillets. Future sustainable formulations may mitigate this dependence by incorporating algae-derived oils (Santigosa et al., Reference Santigosa, Constant, Prudence, Wahli and Verlhac-Trichet2020) or by enriching omega-3 content through substrate manipulation (St-Hilaire et al., Reference St-Hilaire, Cranfill, McGuire, Mosley, Tomberlin, Newton, Sealey, Sheppard and Irving2007; Hussein et al., Reference Hussein, Pillai, Goddard, Park, Kothapalli, Ross, Ketterings, Brenna, Milstein, Marquis, Johnson, Nyrop and Selvaraj2017). Fillet yield, hepatosomatic index (HSI), and peritoneal fat score did not differ among diet groups, indicating that growth partitioning and energy deposition were comparable. Similar HSI and fat scores, metrics of energetic status in fish (Won and Borski, Reference Won and Borski2013), suggest consistent nutrient utilization between diets. No excess deposition of visceral fat or liver mass was observed, indicating efficient allocation of dietary protein and energy toward somatic growth. Collectively, the growth and body composition data indicate that larvae meal provides sufficient nutritional quality to sustain normal somatic development and energy allocation in juvenile trout.
Intestinal health
Given the known sensitivity of the trout digestive system to plant proteins, the impact of larvae meal on intestinal health and homeostasis was evaluated. Intestinal villi showed uniform length and organization across all diet groups, with no signs of diet-induced morphological alterations or inflammation. Epithelial cell nuclei were consistently basal in position, as expected for a healthy gastrointestinal tract. Goblet cells, responsible for mucosal secretion, were present at similar levels among mucosal folds in all groups, indicating consistent mucus production and barrier integrity. The lamina propria also exhibited normal morphology and was free from inflammation, edema, or other pathological changes. The comparable intestinal health across diet groups in this study underscores the potential of larvae meal as a alternative protein source in aquafeeds. The lack of abnormalities in intestinal tissue contrasts with the well-documented sensitivity of trout to plant-derived proteins, which can lead to enteritis or other digestive disorders due to anti-nutritional factors and structural components such as non-starch polysaccharides (Francis, Makkar and Becker, Reference Francis, Makkar and Becker2001; Krogdahl et al., Reference Krogdahl, Penn, Thorsen, Refstie and Bakke2010). In distinction, the high digestibility and protein content of housefly larvae meal likely contribute to its compatibility with the trout gastrointestinal tract, without inflammatory responses or morphological disruptions.
Blood chemistry
Key electrolytes, including sodium, potassium, and chloride, remained consistent across diets, indicating stable osmoregulatory function. Similarly, levels of total protein, lipids, glucose, and AST were unaffected, suggesting that larvae meal diets support normal metabolic and hepatic functions associated with digestive physiology. The observed decrease in phosphate levels in the FM50LM50 group compared to the FM100 and LM100 groups is noteworthy, as it may reflect differences in phosphorus bioavailability or utilization between the mixed vs the single-protein diets. Nevertheless, indicators of skeletal development showed no signs of being compromised compared to the other treatments. In line with the favorable growth results, blood chemistry suggests that the substitution of fishmeal with larvae meal as a dietary protein source in rainbow trout does not have a negative physiological impact under the conditions of this study. While the calcium levels were slightly lower in the FM50LM50 group, they remained within normal physiological ranges, suggesting no deficiencies or adverse effects on calcium metabolism.
Gut microbiota
The dietary protein source is known to influence the intestinal microbiota in fish, which plays a critical role in digestion, nutrient absorption, immune function, and overall health (Ringø and Gatesoupe, Reference Ringø and Gatesoupe1998; Llewellyn et al., Reference Llewellyn, Boutin, Hoseinifar and Derome2014; Ringø et al., Reference Ringø, Hoseinifar, Ghosh, Doan, Beck and Song2018). In this study, replacing fishmeal with larvae meal produced clear shifts in microbial composition at the phylum, order, and family levels, indicating that larvae meal substantively alters the gut microbial community. In addition to nutritional drivers, emerging evidence suggests that insect meals may contain bioactive immune components, such as antimicrobial peptides, chitin and bioactive lipids that can selectively influence microbial taxa (Nogales-Mérida et al., Reference Nogales-Mérida, Gobbi, Józefiak, Mazurkiewicz, Dudek, Rawski, Kierończyk and Józefiak2019; Maulu et al., Reference Maulu, Langi, Hasimuna, Missinhoun, Munganga, Hampuwo, Gabriel, Elsabagh, Van Doan, Abdul Kari and Dawood2022). This raises the possibility that both nutrient composition and innate elements of insect tissues contributed to the observed patterns. Although the larvae meal diets generated distinct microbiota profiles compared to the fishmeal diet, the functional consequences of these changes remain to be determined.
At the phylum level, Firmicutes dominated the microbiota in fish fed the mixed (FM50LM50) and larvae meal-only (LM100) diets. In contrast, the fishmeal-only (FM100) group exhibited a more balanced profile, with Firmicutes at 49% and Actinobacteria contributing 24%. Firmicutes are generally associated with efficient energy extraction and metabolic flexibility, which may be advantageous in fish consuming larvae meal diets (Flint et al., Reference Flint, Scott, Duncan, Louis and Forano2012). The reduced presence of Actinobacteria in larvae meal-fed groups could indicate a dietary shift away from substrates typically supporting this phylum, such as those rich in carbohydrates or other complex plant-derived components (Brown, Sadarangani and Finlay, Reference Brown, Sadarangani and Finlay2013).
At the order level, Clostridiales dominated the microbiota of the larvae meal containing groups. Members of Clostridiales are known for their ability to ferment diverse substrates and produce short-chain fatty acids (SCFAs), which support gut health and epithelial integrity (Louis et al., Reference Louis, Hold and Flint2014). The higher abundance of Clostridiales in larvae meal-fed fish may indicate enhanced fermentation activity, potentially reflecting the protein- and lipid-rich composition of larvae meal. Other orders, such as Bacillales and Lactobacillales, were consistently present across all groups, albeit with varying abundances. Bacillales, known for their enzymatic capabilities and potential probiotic effects (Ringø and Gatesoupe, Reference Ringø and Gatesoupe1998), were more prevalent in the FM100 group, while Lactobacillales, associated with antimicrobial properties and gut immune modulation (Ringø and Gatesoupe, Reference Ringø and Gatesoupe1998), were higher in the larvae meal substituted groups.
At the family level, larvae meal-fed groups showed dominance by Peptostreptococcaceae and Bacillaceae. These families include taxa that are metabolically versatile and capable of fermenting high-protein diets (Fan et al., Reference Fan, Liu, Song, Chen and Ma2017). In contrast, the FM100 group exhibited a more diverse microbiota and may reflect a more balanced nutrient profile in fishmeal. Reduced microbial diversity in larvae meal-fed fish could potentially reflect the predominance of certain substrates favoring specific taxa. However, the reduced abundance of Actinobacteria, which include some beneficial genera like Bifidobacterium, warrants further investigation to determine whether their depletion has functional consequences.
Conclusion
This study demonstrates that housefly larvae meal can effectively replace a substantial portion of fishmeal in rainbow trout diets without compromising growth performance, body composition, intestinal morphology, blood physiology, or gut microbiota composition. Fish fed larvae meal–based diets performed comparably to those receiving fishmeal and commercial feeds, and no diet-induced intestinal pathology was observed. These results indicate that housefly larvae meal is a nutritionally viable and biologically compatible alternative protein source for aquafeeds. In line with emerging data that insect meals may even confer benefits to gut health and immune function (Ido et al., Reference Ido, Iwai, Ito, Ohta, Mizushige, Kishida, Miura and Miura2015), the distinct but physiologically consistent microbiota profiles observed here further support the compatibility of larvae meal with trout gastrointestinal homeostasis. Because housefly larvae are naturally low in long-chain omega-3 fatty acids, larvae meal–based formulations will still require complementary lipid supplementation (e.g., fish oil or algae-derived oils) to meet salmonid PUFA requirements.
Given the continued pressures on global fishmeal resources and the need to diversify protein inputs for aquaculture (Tacon and Metian, Reference Tacon and Metian2008), insect-derived proteins offer a promising pathway for reducing reliance on forage fisheries. However, the practical adoption of insect meals, including those derived from housefly larvae, will ultimately depend on regulatory frameworks, production scalability, and market considerations (Selvaraj and Won, Reference Selvaraj and Won2024). Within this broader context, the present findings establish strong biological justification for evaluating housefly larvae at commercial scale. Future work should assess long-term grow-out performance, economic feasibility, and optimized formulations to support the integration of larvae meal into modern aquafeed production.
Data availability statement
All data supporting this study are contained within the manuscript.
Acknowledgements
The authors gratefully acknowledge Dr. Patricia Johnson, Chair of the Department of Animal Science, for providing facilities and resources essential to this study. We also extend our sincere thanks to Dr. J. Thomas Brenna and Dr. Kumar S. Kothapalli of Cornell University for their invaluable expertise and assistance in conducting the fatty acid analysis, and Dr. Michael J. Maguire for expertise and assistance with the gut microbiota analysis.
Funding statement
This study was funded by USDA Hatch Grants (2021-22-184 and 2023-24-144) to V.S. and E.T.W., an academic venture fund from the Cornell Atkinson Center for Sustainability (2013-AVF) to V.S., the Charitable Trust Research Grant from the College of Agriculture and Life Sciences at Cornell University to J.M.G., and funds from the Department of Microbiology and Immunology Aquatic Animal Health Program, Cornell University College of Veterinary Medicine to H.M.
Competing interests
The authors declare none.
Open access
