Ethics statement

All animal experiments were performed in accordance with the Ethical Principles and Guidelines for Using Animals for Scientific Purposes established by the National Research Council of Thailand and were approved by the Suranaree University of Technology Animal Care and Use Committee under the animal use protocol number SUT-IACUC-016/2024.

Insect rearing and experimental setup

BSF (Hermetia illucens L.) eggs were obtained from a colony maintained at the pilot greenhouse for industrial insect research and production (Industrial Insect Division, Khon Kaen University, Thailand). The eggs were incubated at 28.0 °C ± 1.0 °C and 70–80% relative humidity for 24 h and subsequently transferred to a hatching chamber under the same environmental conditions. On day 3 after oviposition, when hatching was optimal and larvae were of uniform age, the larvae were used for subsequent experiments.

Two feed types were evaluated:

1. (1) Modified feed (MF): In each feed production, a mixture of commercial DRB and DSSCs from a single production batch was prepared at a 9:1 (w/w) ratio.
2. (2) Organic waste (OW): Mixture of expired fruits (cantaloupe, watermelon, dragon fruit, Thai melon, papaya and avocado) and vegetables (onion, coriander, water spinach, cabbage, broccoli, kale, pumpkin, cucumber, tomato and corn), collected from local supermarkets in Khon Kaen, Thailand.

During the pre-experimental phase, 150 g of each feed substrate was placed in plastic containers (12 × 17 × 5.5 cm) and moisture content was adjusted to 70–80% using distilled water. One gram of 1-day-old larvae (60 000 larvae) was introduced onto the surface of a nutritious substrate containing a mixture of soybean milk residue and commercial chicken feed containing 14% protein in a 1:1 ratio to ensure optimal growth and development. After 5 days of growth, the larvae were used to initiate the main feeding experiments with the two rearing substrates (MF vs OW).

Growth performance was evaluated using a completely randomised design involving the two feed treatments, with five replicates per treatment. Two hundred larvae were allocated to each replicate and counted manually. Larvae were individually transferred using a soft brush under a stereomicroscope into a Petri dish to ensure accurate counting and minimise handling damage. All counts were performed by the same trained personnel to reduce human error and maintain consistency across replicates. Once 200 larvae were confirmed, they were immediately transferred to large plastic containers (20 × 31 × 8 cm) containing 200 g of the assigned feed. Moisture content was maintained at 70–80%, and feeds were replenished every 2–3 days as required.

The growth parameters measured included larval weight, prepupal weight, development time (from the larval to the prepupal and adult stages) and adult sex ratio. Five random samples from each replicate on day 5 of the larval stage (the experimental onset) were collected daily until the prepupal stage for weight assessment. Samples were collected during key developmental stages (i.e. larval, prepupal and pupal). The samples were cleaned and stored at −18 °C for 3 days before being transferred to Suranaree University of Technology (Thailand) for further analysis.

Proximate composition

The proximate composition of BSF samples was determined following the standard procedures established by the Association of Official Analytical Chemists (AOAC, 2016), including moisture (method 934.01), crude protein (method 976.05), crude fat (method 920.39) and ash (method 942.05) analyses. Moisture content was measured by drying the samples in an oven at 105 °C until a constant weight was achieved. Crude protein was quantified using the Kjeldahl method, with a nitrogen-to-protein conversion factor of 6.25. The total lipid content was measured using the Folch method (Folch et al. 1957). Ash content was measured by incinerating the samples in a muffle furnace at 550 °C. Total dietary fibre was quantified using the AOAC method 985.29 (AOAC, 2023).

Mineral and heavy metal detection

Mineral and heavy metal concentrations were measured using inductively coupled plasma mass spectrometry (ICP-MS) as described by the AOAC. Arsenic (As), lead (Pb) and mercury (Hg) contents were measured using the AOAC (2019) method 2015.01. Cadmium (Cd) content was measured using the AOAC (2023) method 2015.01. Calcium (Ca), chromium (Cr), copper (Cu), magnesium (Mg), nickel (Ni), potassium (K) and sodium (Na) contents were measured using the AOAC (2019) method 999.10. Iron (Fe), manganese (Mn), phosphorus (P) and zinc (Zn) contents were measured using the AOAC (2023) method 999.10.

Amino acid analysis

The amino acid composition of BSF samples was determined using the AOAC Official Methods 994.12 and 988.15 (AOAC, 2000). Briefly, 50 mg of each sample was hydrolysed with 6 M hydrochloric acid (HCl) at 110 °C for 24 h under a nitrogen atmosphere to prevent oxidation.

Following hydrolysis, amino acids were separated using an amino acid analyser (Biochrom 30+, Biochrom, Cambridge, UK) equipped with a high-resolution cation-exchange column (u-3183; 200-mm bed length, 4.6-mm internal diameter). Mobile phases consisted of a lithium citrate buffer (pH 2.80–3.55) and a lithium hydroxide buffer (pH 14.0). Post-column derivatisation was performed using ninhydrin and amino acids were detected at 570 nm. To detect sulfur-containing amino acids (methionine and cysteine), performic acid oxidation was performed before acid hydrolysis. Tryptophan was quantified separately through alkaline hydrolysis. Amino acid concentrations were expressed as grams per 100 grams of wet sample (g/100 g wet sample).

Fatty acid composition analysis

The fatty acid composition of BSF samples was determined using the AOAC (2023) method 996.06. Briefly, total lipids were extracted using a chloroform–methanol solvent system as described by Folch et al. (1957). The extracted lipids were subjected to transesterification to obtain fatty acid methyl esters as described in the AOAC method.

Transcriptomic analysis

Total RNA extraction: BSF individuals reared on MF and OW were collected at the larval, prepupal and pupal stages. Total RNA was extracted from the pooled samples of three individuals using a Tissue Total RNA Purification Mini Kit (Favorgen, Pintung, Taiwan) according to the manufacturer’s instructions. RNA quality was assessed using agarose gel electrophoresis, with RedSafe Nucleic Acid Staining Solution (iNtRON) used for detection. RNA quantity was measured using a Nanodrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA).

Sequencing analysis: Six libraries were prepared using 2 μg of total RNA extracted from larvae, prepupae and pupae. Sequencing was performed by Biomarker Technologies (Beijing, P.R. China). mRNA was purified from total RNA using poly-T oligo-attached magnetic beads. Following fragmentation, first-strand cDNA was synthesised using random hexamer primers and second-strand cDNA was subsequently synthesised using either dUTP for directional libraries or dTTP for nondirectional libraries. The double-stranded cDNA was purified using AMPure XP beads and fragments within the size range of 200–500 bp were selected using the same beads. Thereafter, a cDNA library was constructed by amplifying the cDNA fragments multiple times through polymerase chain reaction (PCR) and the qualified library was sequenced on a high-throughput platform in the PE150 mode. Clean data were obtained by filtering raw reads to remove adapter sequences and low-quality reads. The filtered reads were aligned to the Hermetia illucens L. reference genome (GCA_905115235.1). Differentially expressed genes (DEGs) were identified using the Benjamini–Hochberg false discovery rate (FDR) correction method. Genes with an adjusted p-value (FDR) < 0.05 and an absolute ${log}_2(\mathrm{fold}\mathrm{change})\mathrm{\ge }1$ were considered significantly differentially expressed. The DEGs were annotated using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) to determine their functional roles and pathway associations. We examined three types of GO annotation systems, including the functional annotation of gene products. The KEGG database was used to identify genes within pathways and elucidate high-level functions and utilities of biological systems, such as cells, ecosystems and molecules.

Quantitative reverse transcription PCR for gene expression analysis: One microgram of total RNA extracted from larvae, prepupae and pupae from each treatment group was used as a template for cDNA synthesis using the iScript™ cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA). Quantitative PCR was performed on a CFX96 Touch Real-Time PCR System (Bio-Rad) using the Thunderbird Next SYBR qPCR Mix (Toyobo) and specific primers against target DEGs (Table A1 in the Appendix). PCR conditions were set as follows: initial denaturation at 95 °C for 3 min, followed by 40 cycles of denaturation at 95 °C for 30 s, annealing at 60 °C for 30 s and extension at 72 °C for 30 s. Relative gene expression was analysed from three biological replicates using the 2^{−ΔΔ Ct} method (Livak and Schmittgen, 2001) and normalised against the expression of the 18S rRNA gene (internal control) (Wakuda et al., 2024). The correlation between qRT-PCR and RNA-seq data was analysed using Pearson’s correlation coefficient (Everaert et al., 2017; Ai et al., 2022).

Statistical analysis

All data are presented as mean ± SD. Differences between feed types (DRB vs DSS and OW vs MF) were analysed using an independent samples t-test. The effects of feed type (OW and MF) and developmental stage (larval, prepupal and pupal) were evaluated using a two-way analysis of variance (2-way ANOVA) with 2 main effects (feed type and developmental stage) without interaction term. The main effects were treated as fixed factor. Duncan’s multiple range test was used for multiple comparisons after significant differences were detected by the ANOVA. Prior to the analyses, the response variables were tested for normality and homogeneity of variances by Shapiro–Wilk test and Levene’s test, respectively. For datasets that did not meet the normality and/or variance homogeneity assumptions, the data sets were log10-transformed prior to the t-test and the ANOVA. All statistical tests were conducted at 95% confidence level using IBM SPSS Statistics (Version 26.0) (IBM, 2019).

Chemical composition of BSF feeds

DRB is rich in protein, dietary fibre and ash content, whereas DSSCs contain high water and protein contents (Table 1). Protein, fat, dietary fibre, vitamins, minerals and phytochemicals are abundant in rice bran. A majority of the dietary fibre in DRB is insoluble and includes cellulose, hemicellulose and arabinoxylans. A small amount of soluble dietary fibre in DRB includes pectin and β-glucan (Sapwarobol et al., 2021). DSSCs are primarily composed of fish myofibrillar proteins with small amounts of egg white protein and starch, which is consistent with the observed proximate composition of DSSCs (Table 1). MF (DRB+DSSCs) showed relatively high protein and ash contents and a low fat content. Conversely, OW, which contained expired fruits and vegetables, showed the highest crude fat and crude fibre content ($p\mathrm{<}0.05$).

View Full Size

Chemical composition of feed ingredients and formulated feeds for BSF rearing (wet weight basis)

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

The concentrations of minerals and heavy metals in the two types of feed are shown in Table 2. Large variations were observed between different lots of feed ingredients for both DRB and DSSCs. DRB exhibited a high P concentration, which is consistent with previous reports (Sharif et al., 2014). In addition, K was another predominant mineral. On the contrary, DSSCs exhibited a high Na concentration owing to the use of salt (sodium chloride [NaCl]) in the formulation to solubilise fish myofibrillar proteins for gelation. MF contained higher Mg, Na and P than OW ($p\mathrm{<}0.05$). MF offered a better-balanced mineral profile, exhibiting intermediate levels of Na, P and Mg compared to those in individual substrates.

View Full Size

Mineral and heavy metal concentrations in feed ingredients and feed substrates for BSF rearing (mg/kg wet weight basis)

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

The total arsenic (As) content was higher in DRB than in DSSCs ($p\mathrm{<}$ 0.05). However, MF exhibited a total arsenic content comparable to that of OW ($p\mathrm{>}0.05$). Previous studies reported that approximately 69.9–72.2% of total arsenic in rice bran occurs in inorganic form (Ruangwises et al., 2011), which is classified as a human carcinogen (International Agency for Research on Cancer, 1998). According to European Union (EU) regulations (2023/915; 2019/1869), the maximum limit for total arsenic in feed materials is 2 mg/kg, while specific limits for inorganic arsenic in food products such as rice, baby food, fruit juices and seafood are typically set at 0.1–0.3 mg/kg. In this study, the total arsenic content of both MF and OW (0.11 mg/kg) was well below the EU limit. Based on the reported proportion of inorganic arsenic in rice bran, the inorganic arsenic content was estimated at 0.069–0.072 mg/kg, which seems to fall within the established limits. However, this estimation was derived from literature-reported inorganic arsenic proportions rather than direct arsenic speciation analysis, which should be recognised as a limitation of the present study. Therefore, future studies should employ direct speciation techniques, such as high-performance liquid chromatography coupled with inductively coupled plasma mass spectrometry (HPLC–ICP–MS), to accurately distinguish and quantify inorganic and organic arsenic species in both feed substrates and BSF biomass.

Notably, large variations in As content were observed in OW feed, reflecting the inconsistent safety/quality of feed formulated from fruit and vegetable wastes from supermarkets. MF contained all heavy metals, exhibiting no notable changes compared to OW. Cd was not detected in any ingredient or feed. Hg was only found in trace amounts in OW. The Pb contents detected in MF and OW were lower than the international maximum levels typically applied to food and feed commodities. Overall, although As was the predominant contaminant in DRB, other heavy metals were present at negligible or undetectable levels across all substrates. Although the total As content of formulated feeds was below the maximum limit for animal feed (EU2023/915), its presence emphasises the importance of monitoring the safety of feed formulations used for BSF production.

Growth performance

BSFL reared on OW, a high-fibre diet and MF, a high-protein diet, showed significant differences in development. BSFL reared on MF showed significantly faster growth during the larval stage than BSFL reared on OW. Specifically, the MF group completed larval development in an average of 15.96 ± 2.01 days, whereas the OW group required 19.94 ± 2.44 days (Figure 1). Furthermore, MF-fed BSFL showed a shorter total developmental period from the larval stage through pupation to adult emergence, completing the entire life cycle in an average of 33.18 ± 2.41 days. However, OW-fed BSFL required 36.61 ± 2.49 days. These results indicate that MF is more effective in promoting larval growth and development, reducing the total developmental period by approximately 3–4 days. The high protein content in MF promotes rapid early weight gain through enhanced amino acid availability for tissue growth and enzyme synthesis, which is consistent with the protein-driven growth responses observed in BSFL (Spranghers et al., 2017). However, the low fat content in MF can restrict long-term biomass accumulation. The sharp decline in larval weight after the peak reflects the onset of prepupal transition, during which energy is redirected toward pupal development and cuticle formation (Gobbi et al., 2013). The higher larval weight observed in the OW group indicates that OW can promote greater nutrient storage before metamorphosis, a pattern consistent with previous findings on substrate-dependent growth in BSFs (Meneguz et al., 2018). On the contrary, the higher lipid and fibre contents in OW enable more sustained energy release, leading to more weight gain and delayed growth. These findings are consistent with those of studies reporting that lipid-rich or heterogeneous substrates extend larval feeding duration and contribute to an increased final body mass (Barragán-Fonseca et al., 2017). Furthermore, the dietary fibre in OW can influence gut microbial composition, contributing to more efficient digestion and nutrient assimilation over an extended period. This phenomenon is consistent with substrate–microbe interactions observed in BSFL (Nyakeri et al., 2017).

View Full Size

Growth performance of BSFL fed two diets. MF, modified feed (a mixture of DRB and DSSCs in a ratio of 9:1); OW, organic waste (expired fruits and vegetables from local supermarkets). Error bars indicate SD.

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

The sex ratio of emerged adults showed notable differences between the MF and OW groups. The MF group showed a male-biased sex ratio of 1.38:1 (male: female), whereas the OW group showed a female-biased sex ratio of 0.61:1. These findings suggest that diet composition may be associated with variations in the sex ratio of adult BSF individuals. However, because other potentially confounding factors, such as larval density and microclimate conditions, were not specifically controlled in the present study, a direct causal relationship cannot be conclusively established. Therefore, diet selection should be aligned with the specific objectives of BSF production. The MF diet may be more suitable when rapid production is prioritised, whereas the OW diet was associated with greater biomass accumulation. In addition, the observed trend toward a higher proportion of female individuals in the OW group may warrant further investigation for its potential relevance to continuous breeding systems.

Nutritional composition of BSFs

Each of the larval, prepupal and pupal stages of BSFs showed differences in nutritional composition between the MF and OW groups. BSFs are a good source of protein and lipids (Table 3). Crude protein content increased with the progress of each developmental stage, peaking during the pupal stage ($p\mathrm{<}0.05$). These trends are consistent with those reported by Liu et al. (2017). The MF group showed a higher protein content at all developmental stages ($p\mathrm{<}$ 0.05). Similarly, studies have reported that BSFL reared on nutrient-rich substrates accumulate more protein than those reared on organic residues (Barragán-Fonseca et al., 2017). Compared with MF, OW, which had a higher fat content (Table 1), consistently resulted in higher fat levels in BSF individuals at all developmental stages ($p\mathrm{<}0.05$). This trend supports previous findings indicating that OW substrates enhance lipid deposition in BSFs, highlighting their potential in biofuel and feed applications (Wang and Shelomi, 2017). In this study, feed type is the main factor governing fat content of BSF with OW resulting in higher content ($p\mathrm{<}0.05$). The total lipid content detected in OW-fed BSF prepupae is comparable (approx. 37%, dry basis) to that reported by Spranghers et al. (2017). Furthermore, OW-fed BSFs showed lower protein and higher fat contents (Fuso et al., 2021).

View Full Size

Chemical composition of BSFs reared on MF and OW at three developmental stages

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

Seafood-derived substrates such as fish trimmings and shrimp carcasses have been reported to yield BSF larvae with protein contents of approx. 30-45% (dry basis), fat contents up to approx. 35% (dry basis) and material reduction efficiencies exceeding 50% (Hu et al., 2024; Lopes et al., 2025). In comparison, the use of surimi crabstick byproducts (DSSCs) in this study provides a more homogeneous and protein-rich substrate, potentially enabling improved digestibility and a more balanced protein–lipid profile while maintaining comparable bioconversion performance.

Ash content increased with developmental stage, irrespective of the feed type used ($p\mathrm{<}0.05$), while carbohydrate contents were comparable in all samples ($p\mathrm{\ge }$ 0.05). Diener et al. (2009) reported that mineral content was less influenced by feed type than protein or lipid content. Our results indicate that the nutritional value of BSFs at all developmental stages can be effectively modulated by using agro-industrial byproducts as rearing substrates.

Most minerals were not affected by feed type or developmental stage ($p\mathrm{\ge }$ 0.05). However, calcium (Ca) content appeared to increase with advancing developmental stage ($p\mathrm{<}0.05$, Table 4). Magnesium (Mg) concentrations in BSFs have been reported to range from approximately 1.0 to 3.5 g/kg dry matter, with variability largely attributed to differences in rearing substrates and processing methods (Lu et al., 2022). However, in the present study, feed type did not exert a significant effect on Mg content ($p\mathrm{\ge }$ 0.05), indicating that Mg levels remained relatively stable across the dietary treatments evaluated. Arsenic (As) was the only predominant heavy metal detected, with significantly higher levels observed in MF-fed BSF at all stages ($p\mathrm{<}0.05$). Although the concentrations remained within the safety limits established by the EU for food applications, this finding raises concerns regarding the use of rice bran as a co-substrate in BSF rearing. In this context, the use of OW appears to be a more favorable alternative in terms of minimising arsenic accumulation. Cadmium (Cd) and lead (Pb) were detected in all samples at low levels, whereas mercury (Hg) was mostly undetectable. Heavy metal analysis confirmed that heavy metal concentrations in BSFs in both the MF and OW groups remained within international safety thresholds. Similarly, previous studies have shown that heavy metal accumulation in BSFs largely depends on the source and composition of rearing substrates (Purschke et al., 2017). Therefore, although DRB is a nutritionally valuable rearing substrate, its use in large-scale insect farming requires consistent monitoring to prevent potential arsenic accumulation in food-grade insects.

View Full Size

Mineral and heavy concentrations in BSFs reared on MF and OW at three developmental stages (mg/kg wet weight basis)

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

Amino acid composition of BSFs

The amino acid composition of BSFs significantly varied with the developmental stage and feed type (Table 5). The total amino acid (TAA) content of MF-reared BSFs was higher than that of their OW-reared counterparts ($p\mathrm{<}0.05$), indicating that MF markedly enhances overall amino acid accumulation. In contrast, in OW-fed BSFs, TAA content increased progressively with developmental stage, reaching its maximum at the pupal stage ($p\mathrm{<}0.05$). This trend is consistent with the physiological requirement for protein reserves before metamorphosis (Finke, 2013). The feed type influenced the amino acid profile of BSFs with the MF group consistently showing higher levels than the OW group, especially histidine, isoleucine, leucine, threonine, arginine, aspartic/asparagine, glycine, serine and glycine. OW consistently resulted in a lower TAA content, with the lowest content observed in OW-fed larvae. These findings indicate that protein-rich diets can enhance protein accumulation in BSFs. Similar diet-related effects on insect nutrient profiles have been reported by other studies (Barragán-Fonseca et al. 2017).

View Full Size

TAA content in BSFs reared on MF and OW at three developmental stages (g/100 g sample; wet basis)

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

Among essential amino acids (EAAs), leucine, lysine, phenylalanine and valine showed the highest concentrations across all developmental stages, validating the nutritional value of BSFs as a protein source (Wang and Shelomi, 2017). These EAA levels satisfy the Food and Agriculture Organisation (FAO) requirements for human nutrition (Fuso et al, 2021). Among nonessential amino acids (NEAAs), alanine (Ala), asparagine (Asn)/aspartic acid (Asp) and glutamine (Gln)/glutamic acid (Glu) dominated the profile, with their concentrations exceeding 1.0 g/100 g in most samples. These findings are consistent with previous studies reporting that NEAAs, especially Glu and Asp, are abundant in insects and contribute to palatability and nutritional balance (Rumpold and Schlüter, 2013).

Histidine, isoleucine, leucine and threonine exhibited significantly higher concentrations in MF-fed BSFs than in OW-fed BSFs during the larval stage. This difference can be attributed to the higher protein quality and amino acid availability in MF, which is formulated using nutrient-balanced substrates such as DRB and surimi byproducts. Diets rich in high-quality proteins can enhance EAA deposition during the larval stage, when protein synthesis for rapid growth is most active (Barragán-Fonseca et al., 2017). During the prepupal and pupal stages, the concentrations of these amino acids were not significantly different between MF- and OW-fed BSF individuals. This convergence can be explained by the metabolic reallocation of nutrients during metamorphosis, wherein stored proteins and amino acids are mobilised for cuticle formation, pupal development and energy metabolism (Tschirner and Simon, 2015). During these later stages, the influence of diet composition becomes less pronounced, as endogenous metabolic processes dominate nutrient utilisation (Spranghers et al., 2017). In the OW group, the levels of EAAs, including histidine, isoleucine, leucine and threonine and NEAAs, including arginine and tyrosine, were low during the larval stage but increased during the pupal stage (Table 5; $p\mathrm{<}$ 0.05). These trends suggest a compensatory mechanism wherein larvae feeding on substrates with a low protein content accumulate fewer amino acids, whereas nutrient recycling during metamorphosis enables partial restoration of amino acid pools during the pupal stage (Gold et al., 2018). Conversely, MF-fed insects maintained relatively stable amino acid levels across all developmental stages, indicating the consistent availability and efficient assimilation of dietary protein from the nutrient-rich feed.

Branched-chain amino acids (BCAAs) such as leucine, isoleucine and valine showed consistent increases, especially in OW-fed BSFs, which is consistent with their established role in muscle protein synthesis and tissue growth (Wu, 2009). BCAAs are particularly important for enhancing muscle protein synthesis, facilitating muscle growth and repair and reducing muscle fatigue after exercise for athletic performance or recovery. Whey protein is a great source of BCAAs, with approximately 26% of the TAAs being BCAAs (Bos et al., 2000). Herein, BCAAs accounted for approximately 20% of the TAAs in BSFs. Furthermore, the EAA/NEAA ratio ranged from 77.71% to 93.16%, with the highest ratio observed in MF-fed BSF prepupae. These findings indicate that MF contains high-quality protein compared to conventional animal protein sources such as meat, fish and milk (FAO, 2013; Fuso et al., 2021). Irrespective of the feed type, BSF prepupae and pupae exhibited higher concentrations of all amino acids than BSFL (Table 5). Prepupae reared on MF consistently showed the richest amino acid profile, indicating their suitability as an optimal harvest point for protein yield. The enhanced amino acid accumulation under MF feeding suggests that high-quality protein substrates, such as DRB and DSSCs, can improve the efficiency of protein biosynthesis during larval growth and metamorphosis.

Fatty acid composition of BSFs

The fatty acid composition of BSFs significantly varied based on the developmental stage and feed type (Table 6). Across all developmental stages, saturated fatty acids (SFAs) consistently dominated over unsaturated fatty acids (UFAs), irrespective of the feed type. Lauric acid (C12:0) was the most abundant SFA, followed by palmitic acid (C16:0) and myristic acid (C14:0), which is consistent with previous findings indicating that BSFs preferentially synthesise medium-chain fatty acids (MCFAs) from carbohydrate-rich substrates (Spranghers et al., 2017). C12:0 content was the highest in OW-fed BSF pupae (8.27 ± 1.28 g/100 g), which aligns with previous studies reporting C12:0 as the main lipid component in BSFs (Surendra et al., 2016). The OW group usually showed higher SFA levels than the MF group, with OW-fed larvae showing significantly higher C12:0 and C16:0 contents ($p\mathrm{<}$ 0.05). These findings suggest that nutrient composition and substrate digestibility in OW, which contains more complex carbohydrates and fibres, promote de novo lipogenesis pathways, favouring the synthesis of medium-chain SFAs (Barragán-Fonseca et al., 2017).

View Full Size

Fatty acid content in BSFs reared on MF and OW at three developmental stages (g/100 g sample; wet basis)

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

Among UFAs, oleic acid (C18:1n9c) and linoleic acid (C18:2n6) were predominant. These UFAs substantially contribute to the lipid profile of BSFs irrespective of the feed type (Meneguz et al., 2018; Spranghers et al., 2017). C18:1n9c content peaked in OW-fed larvae (2.38 ± 0.28 g/100 g), whereas C18:2n6 content peaked in OW-fed pupae (1.17 ± 0.09 g/100 g). Fatty acid accumulation in BSFs was found to be directly dependent on the amount of fat in the rearing substrate, indicating some possibilities for modulating the fatty acid profile (Li et al., 2022). OW exhibited the highest fat content (Table 1), resulting in increased levels of both SFAs, including C12:0 (a medium-chain SFA) and C16:0 and UFAs, including C18:1n9c, C18:2n6, linolenic acid, which are essential fatty acids, across all development stages of BSFs in comparison to MF. However, no significant difference was observed in docosahexaenoic acid (DHA, C22:6n3) content between the MF and OW groups (Table 6). Similarly, Liu et al. (2017) investigated the fatty acid content of BSFs at various developmental stages and primarily identified C12:0, C14:0, C18:1n9c and C18:2n6. In addition, a study by Surendra et al. (2016) confirmed the presence of medium-chain SFAs in BSFs. C12:0 accounts for approximately 50% of the total SFAs in coconut oil and exhibits antibacterial and antiviral activities. C16:0 is a prevalent SFA in nature, with palm oil (containing up to 44% C16:0) being its significant source. It is a long-chain SFA whose excessive intake is associated with an increased risk of cardiovascular disease and insulin resistance. OW-fed BSFs contained higher fat and fatty acid contents than MF-fed BSFs (Tables 1 and 6). These findings suggest that MF is an appropriate low-fat diet formulation for BSFs.

The UFA/SFA ratio was usually low, ranging from 0.22 to 0.35, with OW-fed BSFs maintaining slightly higher ratios than MF-fed BSFs. The ratio (Table 6) decreased as the growth stages of BSFs progressed, indicating that UFA content decreases with increasing BSF age. Similarly, Liu et al. (2017) reported that UFA levels significantly decreased after the prepupal stage. Overall, BSFs are a valuable source of healthy fatty acids, especially C12:0, compared to other insects, as these fatty acids exhibit antimicrobial and anticancer properties (Suryati et al., 2023).

OW-fed insects showed increased accumulation of lipids, especially MCFAs such as C12:0, whereas MF-fed insects maintained a relatively balanced fatty acid profile with moderate lipid accumulation and potentially improved nutritional quality. This composition can be advantageous for developing BSF-based ingredients for human food, wherein lower saturated fat levels are desirable. These results highlight the influence of feed composition on lipid metabolism and fatty acid deposition during BSF development.

Transcriptomic analysis

In this study, five independent biological replications were used for evaluating BSF growth performance and proximate composition to provide sufficient statistical power for phenotypic and nutritional traits. For transcriptomic analysis, three representative biological replications per treatment were selected from independently reared batches. This replication level is commonly used in RNA-seq experiments and is considered adequate for identifying differentially expressed genes when supported by quality control, clear clustering among biological replicates and validation by qRT-PCR (Fu et al., 2021; Nag et al., 2021). Approximately 20 million reads were generated from six libraries, with clean reads accounting for > 96% of the total reads (Table 7). According to standard RNA-Seq screening criteria, DEGs were identified based on an FDR < 0.05 and | log₂(fold change) | ≥ 1. This resulted in the identification of 711, 3757 and 5178 DEGs from the larval, prepupal and pupal stages, respectively (Figure 2A). Among them, 644, 2480 and 3675 DEGs were upregulated, while 67, 1277 and 1503 DEGs were downregulated during the larval, prepupal and pupal stages, respectively (Figure 2A–C). DEGs associated with molecular functions, cellular components, lipid metabolic processes, catalytic activity and biological processes were analysed based on GO classifications. Increased activities were observed for metabolic and cellular processes within the biological process category, for cellular anatomical entities and lipid metabolic processes within the cellular component category and for catalytic activity and binding within the molecular function category (Figure 2B–D). Furthermore, KEGG pathway analysis revealed potential growth-promoting pathways in larvae, prepupae and pupae. These pathways included oxidative phosphorylation, citrate cycle (TCA cycle) and dilated cardiomyopathy (DCM) in larvae; protein digestion and absorption, insect hormone biosynthesis and amino sugar and nucleotide sugar metabolism in prepupae; and Toll and IMD signaling pathway, cardiac muscle contraction and phototransduction (fly) in pupae (Figure 2E–G).

View Full Size

RNA sequencing (RNA-Seq) data of BSFs at different developmental stages

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Transcriptomic analysis of BSFs fed different diets (MF and OW). (A) Differentially expressed genes (DEGs) identified between the MF and OW groups. (B) Venn diagram illustrating shared and unique DEGs among the larval, prepupal and pupal stages. DEGs were defined using a false discovery rate (FDR) < 0.05 and an absolute fold-change threshold of $\mid {log}_2(\mathrm{fold}\mathrm{change})\mid \mathrm{\ge }1$. (C–E) Gene Ontology (GO) classification of DEGs into biological processes (red), cellular components (blue) and molecular functions (green). (F–H) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of DEGs in MF- and OW-fed BSFs at the (F) larval, (G) prepupal and (H) pupal stages.

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

Overall, the transcriptomic data provided a comprehensive overview of gene expression patterns in Hermetia illucens L. at different developmental stages (Groen et al., 2020). Previous studies have investigated the molecular mechanisms underlying fat accumulation in BSFs, constructing detailed models based on transcriptomic analysis (Zhu et al., 2019). Our findings revealed distinct gene expression profiles across the larval, prepupal and pupal stages of BSFs reared on MF. The DEGs were significantly enriched in metabolic pathways related to protein digestion and absorption, oxidative phosphorylation and ribosomal biogenesis, indicating enhanced energy metabolism and protein synthesis (Figure 2B–D). BSFs can efficiently degrade OW and convert it into insect biomass containing high fat and protein contents. Studies have identified several putative genes involved in key fat metabolic pathways, including pyruvate formation, acetyl-CoA biosynthesis, acetyl-CoA transcription, fatty acid biosynthesis and triacylglycerol biosynthesis (Zhu et al., 2019). Herein, KEGG analysis highlighted key pathways such as the TCA cycle, insect hormone biosynthesis and phototransduction, suggesting that MF can influence both developmental and physiological processes in BSFs (Figure 2E–G). Studies on insect models such as Drosophila melanogaster have shown that transcriptional changes play a crucial role in metabolic regulation, particularly in response to dietary variations (Farhadian et al., 2012). A study on Tribolium castaneum revealed the involvement of genes related to detoxification, energy metabolism, cytoskeletal organisation, vesicle transport and immune responses after exposure to diflubenzuron (Merzendorfer et al., 2012). These findings emphasise the complexity of gene regulation in insects and highlight the need for further research on BSFs. Our findings suggest that transcriptomic changes can affect growth performance and protein synthesis in BSFs. Overall, this study provides novel insights into the molecular mechanisms underlying BSF development and nutrient assimilation. However, further investigation is warranted to elucidate the transcriptomic response of BSFs to different diets, facilitating their optimisation as a sustainable alternative protein source. Based on the transcriptomic results, DEGs of each BSF developmental stage were randomly selected and used to validate their gene expressions by qRT-PCR analysis.

Gene expression validation through quantitative reverse transcription PCR

To determine the effects of MF on the transcriptomic profile of BSFs, gene expression was compared among the larval, prepupal and pupal stages. The expression levels of 16 randomly selected genes related to BSF growth were measured using quantitative reverse transcription PCR (qRT-PCR). These genes included Culex pipiens pallens ATP synthase subunit alpha (ATP5A1), lysozyme 1 (LYZ), heat shock protein Grp78/BiP/HspA5 (HSPA5-BIP), lipase (LIPA), lipase 3 (LIPA3), calcium/calmodulin-dependent protein kinase type II alpha (CAMK), lysozyme 2 (LYZ2), glutathione S-transferase 1 (GST), superoxide dismutase (Cu-Zn) (SOD1), peroxidase (PXDN), spaetzle (SPZ), the transcription factor Jun (JUN), troponin I transcript variant X13 (TNNI3), tropomyosin-1 (TPM1), Hermetia illucens transient receptor potential-like (TRPL) and eyes shut (EYS). qRT-PCR was used to validate the RNA sequencing results (Figure 3). The significantly upregulated genes in larvae included ATP5A1, LYZ, LYZ2, HSPA5-BIP, GST, TNNI3 and TPM1, with their expression levels showing an increase of 1.9–8 fold. During the prepupal stage, ATP5A1, LYZ, LYZ2, HSPA5-BIP, GST, SOD1, TNNI3, TPM1 and EYS were upregulated, showing an increase of 3–40 fold ($p\mathrm{>}0.05$). Similarly, ATP5A1, LYZ, LYZ2, HSPA5-BIP, SOD1, TNNI3, TPM1, TRPL and EYS were upregulated during the pupal stage, with their expression levels showing an increase of 1.5–16 fold ($p\mathrm{>}0.05$) (Figure 3). LIPA, LIPA3, CAMK, PXDN, SPZ and JUN were downregulated in MF-fed BSFs, with their expression levels showing a decrease of 0.5–10 fold, 0.2–10 fold and 0.2–5 fold in larvae, prepupae and pupae, respectively ($p\mathrm{>}0.05$; Figure 3). Although several genes did not show statistically significant differences by qRT-PCR, the overall expression trends were generally consistent with the RNA-seq results. The lack of statistical significance for some targets may reflect biological variability among samples and the limited number of biological replicates (n = 3), which could reduce the statistical power for detecting moderate expression differences. Comparison of differentially expressed genes (DEGs) identified by RNA-seq and validated by qRT-PCR revealed a significant positive correlation ($R^{\mathrm{2}}=$ 0.7567, $p\mathrm{<}0.0001$). This strong agreement indicates that the RNA-seq data were highly reliable and accurately reflected the transcriptional trends confirmed by qRT-PCR analysis. Similar levels of consistency between RNA-seq and qRT-PCR have been widely reported in previous studies, with $R^{\mathrm{2}}$ values generally ranging from 0.74 to 0.99 (Ai et al., 2022; Everaert et al., 2017). Therefore, the non-significant qRT-PCR results for some genes should be interpreted cautiously and do not necessarily invalidate the overall transcriptomic patterns observed in the RNA-seq dataset. Collectively, these findings support the robustness of the transcriptomic dataset and suggest that the validated gene expression patterns are associated with the enhanced growth performance observed in MF-fed BSFs.

View Full Size

Relative gene expression profiles of BSFs reared on MF and OW at different developmental stages (larva, prepupa and pupa). Gene expression levels were quantified by qRT-PCR using 18S rRNA as the internal reference gene. Relative expression was calculated using the $2^{\mathrm{-}\Delta \Delta C_t}$ method and expressed as mean ± SD (n =9, derived from 3 biological replicates × 3 technical replicates). Bar graphs represent qRT-PCR results, whereas red triangles (▲) indicate genes identified as differentially expressed between MF- and OW-fed BSFs based on RNA-Seq (next-generation sequencing, NGS) analysis. Asterisks (*) indicate significant differences between MF- and OW-fed BSFs within the same developmental stage ($p\mathrm{<}0.05$) based on qRT-PCR. qRT-PCR and RNA-Seq results showed a significant positive correlation, with $R^2=0.7567$ and $p\mathrm{<}0.0001$.

Citation: Journal of Insects as Food and Feed 2026; 10.1163/23524588-bja10424

• Download Figure
• Download figure as PowerPoint slide

Herein, we evaluated the expression profiles of 16 genes associated with various metabolic pathways in Hermetia illucens, such as protein digestion and absorption, amino sugar and nucleotide sugar metabolism, insect hormone biosynthesis and the Toll and IMD signaling pathways (Figure 3). Results revealed significant differences in gene expression levels across the three developmental stages in the MF group. Transcriptomic analysis further highlighted distinct gene expression patterns at each stage, emphasising the complex molecular mechanisms driving BSF growth and physiological transitions (Zhu et al., 2019). During the larval stage, genes related to metabolic processes, protein synthesis and immune responses, such as ATP5A1, LYZ, LYZ2, HSPA5-BIP, GST, TNNI3 and TPM1, were upregulated (Figure 3). Similarly, a study showed that the LYZ gene in Bombyx mori was constitutively expressed in the fat body and exhibited antiviral properties, suggesting the presence of an innate immune mechanism in invertebrates (Li et al., 2018). ATP5A1, a subunit of mitochondrial ATP synthase, plays a crucial role in ATP production (Noble, 2024). HSPA5 and GST-2 have been shown to mitigate TDP-43-induced toxicity in Drosophila, indicating their role in alleviating oxidative stress, particularly in neuronal tissues (François-Moutal et al., 2012). Notably, we found that the concentrations of valine, leucine and isoleucine were higher during the larval stage than during the other two stages (Table 5), suggesting active nutrient absorption and assimilation in preparation for metamorphosis.

During the prepupal stage, genes associated with developmental processes and stress responses, such as SOD1 and EYS, were upregulated (Figure 3). SOD1 plays a key role in superoxide scavenging and lifespan modulation under different dietary conditions in Drosophila (Sun et al., 2012). EYS is essential for eye development in both embryos and adults, suggesting broader functional significance beyond visual processing (Hanson, 2001). These findings indicate that the transition from larva to prepupa involves significant physiological shifts, including altered energy utilisation and activation of genes crucial for pupation.

During the pupal stage, genes involved in developmental regulation and morphogenesis, including TRPL and EYS, were upregulated. TRPL expression was higher in prepupae than in larvae, suggesting its role in gustatory receptor neurons, particularly in response to long-term camphor feeding (Chyb et al., 1999). Furthermore, TRPL contains two potential calmodulin-binding sites, particularly in the brain isoform of the voltage-sensitive Ca^{2 +} channel, suggesting that both genes encode light-sensitive ion channels (Phillips et al., 1992).

Conversely, LIPA, LIPA3, CAMK, PXDN, SPZ and JUN were consistently downregulated across all samples. Notably, SPZ and JUN are involved in the Toll and IMD signalling pathways, which are the two primary innate immune response pathways in Drosophila (Han et al., 2013). Recent studies have suggested that downregulation of Toll-6 expression influences tumour cell migration and invasion in Drosophila metastasis models (Mishra-Gorur et al., 2019). In addition, JUN downregulation appears to play a crucial role in refining photoreceptor differentiation and ommatidia assembly during eye development (Kockel et al., 1997). LIPA3 expression exhibits conditional regulation, with downregulation in the absence of fasting or aging (Hänschke et al., 2022). PXDN can contribute to eye development by regulating cell proliferation and differentiation, with our findings showing a marked reduction in PXDN expression during the larval and prepupal stages (Yang et al., 2016). These findings highlight the dynamic and stage-specific nature of gene expression during BSF development. MF, which is rich in protein, appears to influence metabolic and growth-related pathways at each developmental stage of BSFs, potentially improving overall growth performance and nutrient assimilation efficiency. Further studies are warranted to elucidate the mechanistic basis of the observed gene expression changes, which can help optimise feed formulations for BSFs and promote their practical application as a sustainable protein source.