2026 年 63 巻 論文ID: 2026006
Abstract: Insects are commanding increasing attention as a sustainable and alternative nutritional source in animal feed. In this study, the safety and nutritional profile of the dampwood termite, Hodotermopsis sjostedti, were evaluated as a novel ingredient in chicken feed. Colonies were lab-reared, and individual termites were collected by carefully crushing the wood, freeze-drying, and grinding. The resulting termite meal contained 65.6% crude protein, 16.0% fat, and 5.2% ash, on a dry matter basis. No minerals were present at toxic concentrations. Two feeding experiments using Ross 308 female broilers were conducted to evaluate the effects of termite meal inclusion. In Experiment 1, 20 chickens (8-d-old) were assigned to four treatments consisting of a commercial diet plus: 5.0% fish meal, 4.5% fish meal + 0.5% termite meal, 2.5% fish meal + 2.5% termite meal, and 5.0% termite meal. In Experiment 2, 18 chickens (9-d-old) were assigned to three diets: control, control + 2.5% termite meal, and control + 2.5% fish meal. Body weight and feed intake were recorded daily, and samples were collected at 21 d of age. Growth performances, organ weights, and blood plasma profiles were generally comparable across treatments. Chickens receiving 5.0% termite meal had greater alanine aminotransferase values, suggesting a potential hepatic burden. No adverse effects on growth, meat, or organ weight were observed. In conclusion, H. sjostedti meal demonstrated strong potential as a novel dietary component for broiler chickens. However, further long-term studies with larger numbers of birds are necessary to ensure a safe and effective use in poultry diets.
Insects can be grown from resources that cannot be directly used as human food, including valueless agricultural by-products and organic waste from the food industry, producing a valuable protein with a nutritive value[1]. The use of insects as food and feed has thus become a rapidly growing sector in agriculture, seeking alternative, sustainable protein sources[2,3]. Although fish meal is presently the primary component of animal feed[1], environmental issues (e.g., declining fish stocks and biodiversity loss) have prompted the use of insects as an alternative protein source in poultry diets[4].
Termites feed on lignocellulose, which is the most abundant bioresource on Earth. As lignocellulose is only partially utilized by humans, extensive biomass residue remains underused. Because of their abundance in nature and their richness in protein and fat[5], termites are often used for food, feed, and medicinal purposes, especially in Africa, Asia, and Latin America[6,7,8]. Furthermore, it is known that red junglefowls (Gallus gallus), the ancestor of domesticated chickens (Gallus domesticus)[9], fed on a variety of insects, fruits, and seeds, and termites are an essential seasonal food source[10].
The present study assessed the potential of termites as a novel feed ingredient in poultry diets under the project that proposes the establishment of a novel feed-food production system that exploits underutilized biological resources (Fig. 1). The dampwood termite Hodotermopsis sjostedti (Isoptera: Hodotermopsidae), one of the largest termites in size, inhabits subtropical rainforests and typically completes the entire colony lifecycle in decaying wood. Although a single piece of dead wood often serves as the primary habitat, colonies may reach and utilize multiple logs via underground tunnels. Owing to its hidden lifestyle and ecological niche, H. sjostedti has never been used as a food or feed source. The primary objective of this study was to evaluate the safety and nutritional properties of H. sjostedti meal, including its proximate composition, mineral and vitamin contents, and amino and fatty acid profiles. Two short-term feeding experiments were conducted: Experiment 1 (Exp. 1) assessed the effects of substituting fish meal with different proportions of termite meal; Experiment 2 (Exp. 2) compared the supplementation of a commercial diet with either termite or fish meal. Growth performance, feed intake, meat and organ weights, and blood plasma profiles of female broiler chickens were evaluated.
Diagram showing the “WILFood system - Creation of a Novel Food Production System Using Termites as a Poultry Feed Source (above diagonal).” The findings from this study provide essential information (e.g., safety and nutritional value of termites in chicken feed) that will contribute to this system.
The experimental protocol was approved by the Animal Experiment Committee of the Graduate School of Agriculture, Kyoto University (Approval Number: R3-69, April 4, 2021).
TermitesColonies of H. sjostedti were originally obtained from dead logs of brown-rotted Luchu pine wood (Pinus luchuensis) in the natural forests of Amami Oshima Island (28°19′N, 129°22′E) in Japan. The logs of each colony were kept in containers at the Laboratory of Insect Ecology, Graduate School of Agriculture, Kyoto University, at about 25 °C in the dark. Termites (Fig. 2) extracted from six colonies (colony codes: MT578, MT579, MT581, ET471, KM025, and TK058) were freeze-dried (Freeze Dryer FD-1000; EYELA, Bunkyo, Tokyo, Japan) and then ground (Wonder Crusher WC-3L; Osaka Chemical Co., Ltd., Kita, Osaka, Japan). The resulting termite meal was vacuum-sealed (Impulse Vacuum Sealer V301G-10WK, FUJIIMPULSE Co. Ltd., Toyonaka, Osaka, Japan) and kept at −20 °C until use. The freeze-dried termite meal was thoroughly homogenized and analyzed for moisture, crude protein (CP), crude fat (ether extract; EE), crude fiber, ash, vitamin and mineral contents, and amino acid, free amino acid, and fatty acid profiles, with at least two replicates. The analytical methods used (see Table 1) were based on the Feed Analyses Standard of the Food and Agricultural Materials Inspection Center, Food Labeling Standards of the Food Consumer Affairs Agency, and the Japanese Food Standard Ingredients Analysis Manual of the Ministry of Education, Culture, Sports, Science, and Technology, Japan.
Workers (left) and a colony of Hodotermopsis sjostedti (right) termites.
| Termite meal1 | Fish meal2 | ||||
| Item | Value | Lower limit | Analytical method | ||
| Composition | |||||
| Moisture, % | 6.2 | Oven drying | 7.7 | ||
| Crude protein, % | 61.5 | Dumas method | 62.6 | ||
| Ether extract, % | 15.0 | Diethyl ether extraction | 9.5 | ||
| Crude fiber, % | 7.8 | Filtration | 0.0 | ||
| Ash, % | 4.9 | Dry ashing | 17.8 | ||
| Major minerals | |||||
| Ca, g/kg | 4.38 | ICP spectroscopy4 | 44.8 | ||
| P, g/kg | 8.25 | ICP spectroscopy4 | 27.6 | ||
| Mg, g/kg | 1.49 | ICP spectroscopy4 | 2.2 | ||
| K, g/kg | 13.1 | Atomic absorptiometry | 7.6 | ||
| Na, g/kg | 1.58 | Atomic absorptiometry | 10.85 | ||
| Cl, g/kg | 2.18 | Potentiometric titration | 17 | ||
| S, g/kg | 3.6 | Barium sulfate weight method | 7.2 | ||
| Trace elements | |||||
| Mn, mg/kg | 28.3 | Atomic absorptiometry | 18 | ||
| Zn, mg/kg | 190 | Atomic absorptiometry | 99 | ||
| Cu, mg/kg | 108 | Atomic absorptiometry | 11 | ||
| Fe, mg/kg | 321 | ICP spectroscopy4 | 374 | ||
| Se, mg/kg | 0.38 | Fluorescence spectroscopy | 0.4 | ||
| Co, mg/kg | 0.34 | Atomic absorptiometry | 0.09 | ||
| Mo, mg/kg | n.d.3 | 1.0 | ICP spectroscopy4 | 0.1 | |
| I, mg/kg | n.d.3 | 5.0 | Gas chromatography | 3 | |
| F, mg/kg | n.d.3 | 50.0 | Ion chromatography | ||
| As, mg/kg | n.d.3 | 0.1 | Atomic absorptiometry | ||
| Pb, mg/kg | 0.40 | Atomic absorptiometry | |||
| Cd, mg/kg | 1.34 | Atomic absorptiometry | |||
| Hg, mg/kg | 0.17 | Cold vapor atomic absorptiometry | |||
1Termite meal was analyzed at the Japan Food Research Laboratories, Tokyo, Japan.
2Protein content = 62% (Feedinamics, 2025).
3n.d.= non-detectable.
4ICP = inductively coupled plasma.
Two experiments were conducted in an enclosed animal building at the Graduate School of Agriculture, Kyoto University, in August 2021 (Exp. 1) and February 2022 (Exp. 2) with lighting provided for 18 h/d from 04:00 to 22:00 h. Room temperature was controlled by an air conditioning system (PAR-42MA; Mitsubishi Electric Corporation, Tokyo, Japan) set at 26.0 °C and an energy recovery ventilation system (PGL-62DR; Mitsubishi Electric Corporation). Relative humidity was 38.0%. Because Exp. 2 was conducted during the winter, the room was insulated with a thermal aluminum sheet and maintained at 24.4 ± 0.01 °C and 34.6 ± 0.04% relative humidity by a heater (RHJ65L0915; De Longhi Japan Corp., Tokyo, Japan), a circulator (KCF-SDC181T; Iris Ohyama, Inc., Miyagi, Japan), and a humidifier (HD-RXT721E9-WK; Dainichi Co., Ltd., Niigata, Japan). For both experiments, 28 broiler female chicks (Ross 308) at 0 d of age were obtained from Nippon Chunky Co., Ltd., Okayama, Japan, and reared in an incubator (MX BS 500 Avian Brooder; Rcom Company, Wichita, KS, USA) at 29 °C and 55% relative humidity for 2 d, and at 28 °C and 60% humidity thereafter. At 6 d of age, the birds were weighed and selected for experiments. Individual pens (400 W × 500 D × 600 H mm; stainless steel; CLEA Japan, Inc., Tokyo, Japan) were used for adaptation. All birds were fed a commercial diet (Kumiai Broiler Zenki PC; JA Zen-Noh Nishi Nihon Kumiai Feed Co., Ltd., Kobe, Japan) that contained 56% grain (e.g., corn, polished rice, and buckwheat flour), 32% vegetable oil cake (e.g., soybean cake), 8% animal byproduct meal, and 4% feed additives (Table S1) during the adaptation period.
TreatmentsIn Exp. 1, 20 8-d-old birds of similar body weight (initial body weight 238.7 ± 2.96 g) were assigned randomly to four dietary treatments (n = 5 chickens/treatment). Treatments were a commercial diet (Kumiai Broiler Shiage; JA Zen-Noh Nishi Nihon Kumiai Feed Co., Ltd.; Table S1) supplemented with 5.0% fish meal (CON) or 0.5%, 2.5%, and 5.0% termite meal substituting fish meal (Table 2). Substitutions were chosen based on the reported adverse effect of including >10% insect feed[11,12] and a preliminary trial in our laboratory (Tables S2–S4). The chickens were weighed, feed intake was measured, and the bedding was changed at 9:00 h every morning throughout the experiment.
| Treatment | ||||
| Item | CON1 | 0.5% | 2.5% | 5.0% |
| Ingredient, % (as-fed basis) | ||||
| Base diet2 | 95.0 | 95.0 | 95.0 | 95.0 |
| Termite meal3 | 0.0 | 0.5 | 2.5 | 5.0 |
| Fish meal4 | 4.875 | 4.39 | 2.437 | 0.0 |
| Corn meal | 0.125 | 0.11 | 0.063 | 0.0 |
| Chemical composition5 | ||||
| Dry matter, % | 88.6 | 88.6 | 88.7 | 88.9 |
| Crude protein, % of DM | 23.4 | 23.4 | 23.4 | 23.4 |
| Ether extract, % of DM | 5.39 | 5.36 | 5.25 | 5.11 |
| Gross energy6, kcal/kg | 4637 | 4637 | 4635 | 4632 |
1Control.
2Kumiai Broiler Shiage (JA Zen-Noh Nishi Nihon Kumiai Feed Co., Ltd., Kobe, Japan).
3Dry matter (DM) = 93.9%, crude protein (CP) = 62.9% DM, and ether extract (EE) = 8.2% DM.
4DM = 88.4%, CP = 68.3% DM, and EE = 13.9% DM.
5Calculated values.
6Gross energy = 4341 + 11×CP% + 54×EE% - 24×CA% (Song et al., 2025).
In Exp. 2, 18 9-d-old birds (initial body weight 312.0 ± 3.31 g) were assigned randomly to three dietary treatments (n = 6 chickens/treatment). Treatments included a commercial diet (CTL; Kumiai Broiler Shiage JA Zen-Noh Nishi Nihon Kumiai Feed Co., Ltd.), CTL + 2.5% termite meal (TM), and CTL + 2.5% fish meal (FM). Ingredients and chemical composition of the treatment diets in Exp. 2 are shown in Table 3. Body weight and feed intake were measured at 9:00 h, followed by bedding changes. The feed and residues were sampled daily for dry matter (DM) and CP (DUMATHERM N Pro; Gerhardt Japan Co., Ltd., Tokyo, Japan) analysis.
| Treatment1 | |||
| Item | CTL | TM | FM |
| Ingredient, % (as-fed basis) | |||
| Base diet2 | 100.0 | 93.0 | 93.5 |
| Termite meal3 | 0.0 | 2.5 | 0.0 |
| Fish meal4 | 0.0 | 0 | 2.5 |
| Corn meal | 0.0 | 4.5 | 4.0 |
| Chemical composition5 | |||
| Dry matter, % | 88.4 | 89.3 | 89.1 |
| Crude protein, % of DM | 27.9 | 27.9 | 27.9 |
| Ether extract, % of DM | 4.95 | 6.25 | 6.40 |
| Gross energy6, kcal/kg | 4598 | 4624 | 4624 |
1CTL = control, TM = termite meal treatment, FM = fish meal treatment.
2Kumiai Broiler Shiage (JA Zen-Noh Nishi Nihon Kumiai Feed Co., Ltd., Kobe, Japan).
3Dry matter (DM) = 93.2%, crude protein (CP) = 61.4% DM, and ether extract (EE) = 8.2% DM.
4DM = 85.8%, CP = 58.1% DM, EE = 14.3% DM, and < 23% crude ash.
5Calculated values.
6Gross energy = 4341 + 11 × CP% + 54 × EE% - 24 × CA% (Song et al., 2025).
Before each experiment, DM, CP, and EE (Soxtherm; Gerhardt Japan Co., Ltd.) were analyzed for each feed ingredient to calculate the treatment diet formulation. Gross energy was calculated using the equation proposed by Song et al.[13]. Experimental diets were formulated to contain equal amounts of nitrogen (isoprotein); if needed, corn powder (Hominy Feed; Hiyobaku Co., Ltd., Kobe, Japan) was used for adjustment. The calculated chemical composition of feed for each diet is listed in Tables 2 and 3.
SamplingAt 21 d of age, the chickens were fasted for 6 h before slaughter. Before slaughter, each bird was weighed and placed in a small bucket with cotton pads impregnated with isoflurane (Pfizer Japan Inc., Shibuya, Tokyo, Japan) attached to the lid. The order of fasting and slaughtering of the chickens was fixed to ensure the same fasting hours. After confirming the loss of consciousness, birds were decapitated with a guillotine, and 5 cc of blood was sampled from the cut into a blood collection tube with a sodium heparin film (VENOJECT™ II VP-H050K; Terumo Corporation, Tokyo, Japan). Then, the thighs were disarticulated at the femoral head, the breast meat and inner breast fillets were separated from the bones, the organs were sampled, and the meat and organ weights were measured. Organ abnormalities were visually examined. Blood samples were centrifuged at 4 °C, 1,700 × g for 10 min, and blood plasma parameters were measured by Obihiro Clinical Laboratory Inc., Obihiro, Hokkaido, Japan.
Statistical analysisData obtained from the feeding experiments were analyzed using a general linear model in SAS[14] with a fixed effect of the treatment. Initial body weight was a covariate for final body weight, average daily gain, and feed intake. Data obtained in Exp. 1 were subjected to orthogonal contrasts for the inclusion of termite meal (i.e., control vs. mean of 0.5%, 2.5%, and 5.0% inclusion), and linear and quadratic effects for the level of termite meal in 0.5%, 2.5%, and 5.0% treatments. Statistical significance was set at P < 0.05. Pairwise least squares mean differences and their significance were tested with the PDIFF option in SAS[14].
The chemical composition and mineral content of H. sjostedti termite meal (as-fed basis) are listed in Table 1, along with the reported values for fish meal[15] as a comparison. Dry matter of freeze-dried and ground termite meal contained 65.6% CP, 16.0% EE, 8.3% crude fiber, and 5.2% ash. The termite meal contained 0.04, 3.99, 25.0, and 0.7 mg/100 g of vitamins A, B, C, and E, respectively, and 116 and 300 µg/100 g of vitamin K and folic acid, respectively (Table 4). The amino acid and free amino acid concentrations (as-fed basis) are listed in Table 5, along with the CP and amino acid content of fish meal[15]. Fatty acid profiles of freeze-dried termite meal and those reported for fish meal[15] are presented in Table 6. A total of 45 fatty acids were measured, including 17 saturated fatty acids (SFA), 10 monounsaturated fatty acids (MUFA), and 18 polyunsaturated fatty acids (PUFA). Of these, only nine were detected at a sensitivity of 0.1%, MUFA accounted for 54.3% of the total, followed by PUFA and SFA.
| Termite meal1 | Fish meal2 | ||||
| Item | Value | Lower limit | Analytical method | ||
| Vitamin A | 0 | ||||
| Retinol, mg/100 g | n.d.3 | 0.01 | HPLC4 | ||
| α-carotene, mg/100 g | n.d. | 0.02 | HPLC | ||
| β-carotene, mg/100 g | 0.04 | HPLC | |||
| β-cryptoxanthin, mg/100 g | n.d. | 0.02 | HPLC | ||
| Vitamin B | |||||
| B1 thiamin, mg/100 g | 1.05 | HPLC | 1.7 | ||
| B2 riboflavin, mg/100 g | 2.70 | HPLC | 9.1 | ||
| B12 cobalamin, µg/100 g | 240 | 9.028 | |||
| Vitamin C | |||||
| Total ascorbic acid, mg/ 100 g | 25.0 | HPLC | 0 | ||
| Vitamin D, IU/100 g | n.d. | 30 | HPLC | 200 | |
| Vitamin E | 0.85 | ||||
| α-tocopherol, mg/100 g | 0.7 | HPLC | |||
| β-tocopherol, mg/100 g | n.d. | 0.1 | HPLC | ||
| γ-tocopherol, mg/100 g | n.d. | 0.1 | HPLC | ||
| δ-tocopherol, mg/100 g | n.d. | 0.1 | HPLC | ||
| Vitamin K | |||||
| Phylloquinone, µg/100 g | n.d. | 1.0 | HPLC | ||
| Menaquinone-4, µg/100 g | n.d. | 1.0 | HPLC | ||
| Menaquinone-7, µg/100 g | 116 | HPLC | |||
| Folic acid, µg/100 g | 300 | Microbial quantification | 40 | ||
1Termite meal was analyzed at the Japan Food Research Laboratories, Tokyo, Japan.
2Protein content = 62% (Feedinamics, 2025).
3n.d.= non-detectable.
4HPLC = high-performance liquid chromatography.
| Termite meal1 | Fish meal2 | ||
| Crude protein | 61.5 | 62.6 | |
| Amino acid, % | Free amino acid, % | Amino acid, % | |
| Arginine | 2.76 | 0.33 | 3.74 |
| Histidine | 1.35 | 0.25 | 1.62 |
| Isoleucine | 2.11 | 0.05 | 2.56 |
| Leucine | 3.52 | 0.08 | 4.48 |
| Lysine | 3.03 | 0.12 | 4.68 |
| Methionine | 0.72 | 0.02 | 1.66 |
| Phenylalanine | 1.99 | 0.03 | 2.44 |
| Threonine | 2.20 | 0.10 | 2.59 |
| Valine | 2.88 | 0.17 | 3.11 |
| Alanine | 3.70 | 0.07 | 3.98 |
| Aspartic acid4 | 4.18 | 0.03 | 5.70 |
| Cystine | 0.50 | n.d.3 | 0.55 |
| Glutamic acid5 | 6.10 | 0.25 | 7.80 |
| Glycine | 3.25 | 0.10 | 4.35 |
| Proline | 2.75 | 0.16 | 2.90 |
| Serine | 2.30 | 0.04 | 2.43 |
| Tryptophan | 0.57 | 0.06 | 0.61 |
| Tyrosine | 3.52 | 0.23 | 1.88 |
1Moisture content = 6.2%.
2Protein content = 62% (Feedinamics, 2025).
3n.d. = non-detectable at 0.01% sensitivity.
4Aspartic acid contains asparagine.
5Glutamic acid contains glutamine.
| Termite meal1 | Fish meal2 | |||
| Concentration, % | Ratio, % | Concentration, % | ||
| C6:0, caproic acid | n.d.3 | |||
| C7:0, enanthic acid | n.d. | |||
| C8:0, caprylic acid | n.d. | |||
| C10:0, capric acid | n.d. | |||
| C10:1, caproleic acid | n.d. | |||
| C12:0, lauric acid | 0.1 | 0.8 | 0.009 | |
| C13:0, tridecylic acid | n.d. | |||
| C14:0, myristic acid | 0.2 | 1.6 | 0.42 | |
| C14:1, myristoleic acid | n.d. | |||
| anteiso-C15:0 | n.d. | |||
| C15:0, pentadecylic acid | n.d. | |||
| C15:1, cis-10-pentadecenoic acid | n.d. | |||
| iso-C16:0, isopalmitic acid | n.d. | |||
| C16:0, palmitic acid | 1.4 | 10.9 | 1.24 | |
| C16:1, palmitoleic acid | 0.2 | 1.6 | 0.50 | |
| C16:2, hexadecadienoic acid | n.d. | |||
| C16:3, omega-3 fatty acid | n.d. | |||
| anteiso-C17:0, anteiso heptadecanoic acid | n.d. | |||
| C17:0, heptadecanoic acid | n.d. | |||
| C17:1, heptadecenoic | n.d. | |||
| C18:0, stearic acid | 1.1 | 8.5 | 0.25 | |
| C18:1n-9, oleic acid | 6.6 | 51.2 | 0.85 | |
| C18:1n-7, cis-vaccenic acid | 0.2 | 1.6 | ||
| C18:2n-6, linoleic acid | 3.0 | 23.3 | 0.14 | |
| C18:3n-6, γ-linolenic acid | n.d. | |||
| C18:3n-3, α-linolenic acid | n.d. | 0.13 | ||
| C18:4n-3, stearidonic acid | n.d. | 0.10 | ||
| C20:0, arachidic acid | n.d. | 0.02 | ||
| C20:1n-9, eicosenoic acid | n.d. | 0.46 | ||
| C20:2n-6, eicosadienoic acid | n.d. | |||
| C20:3n-6, dihomo-γ-linolenic acid | n.d. | |||
| C20:3n-3, eicosatrienoic acid | n.d. | |||
| C20:4n-6, arachidonic acid | 0.1 | 0.8 | 0.16 | |
| C20:4n-3, eicosatetraenoic acid | n.d. | |||
| C20:5n-3, eicosapentaenoic acid | n.d. | 0.62 | ||
| C21:5n-3, heneicosapentaenoic acid | n.d. | |||
| C22:0, behenic acid | n.d. | 0.02 | ||
| C22:1n-9, erucic acid | n.d. | 0.54 | ||
| C22:2, docosadienoic acid | n.d. | |||
| C22:4n-6, docosatetraenoic acid | n.d. | |||
| C22:5n-6, docosapentaenoic acid | n.d. | 0.18 | ||
| C22:5n-3, docosapentaenoic acid | n.d. | |||
| C22:6n-3, docosahexaenoic acid | n.d. | 0.46 | ||
| C24:0, lignoceric acid | n.d. | 0.00 | ||
| C24:1n-9, nervonic acid | n.d. | |||
| Σ Saturated fatty acids | 2.8 | 21.7 | 1.959 | |
| Σ Monounsaturated fatty acids | 7.0 | 54.4 | 2.35 | |
| Σ Polyunsaturated fatty acids | 3.1 | 24.1 | 1.79 | |
1Crude fat content = 15.0%.
2Protein content = 62% (Feedinamics, 2025).
3n.d. = non-detectable at 0.1% sensitivity.
The results of Exp. 1, whereby fish meal was replaced by termite meal and added to a commercial diet fed to female broiler chickens, are summarized in Tables 7 and 8. The diet treatments had no significant effect, except for lower liver weight in chickens given 2.5% termite meal compared with those given 0.5% and 5.0% (quadratic contrast, P = 0.015). Average daily gain tended to decrease with increasing termite meal substitution (linear contrast, P = 0.097). Cecum weight was greater in the CON group (inclusion contrast, P = 0.038) than in the termite meal inclusion groups, whereas the heart weight showed a quadratic contrast effect (P = 0.032). Plasma alanine aminotransferase (ALT) concentration showed a significant linear contrast (P = 0.048), with higher values for 5.0% meal substitution. Plasma lactate dehydrogenase (LD) tended to increase proportionately with termite meal substitution level (linear contrast, P = 0.080). Plasma β-hydroxybutyric acid tended to be greater (P = 0.058) at 5.0% meal substitution than with other treatments.
| Treatment1 | P-value | Contrast P-value2 | |||||||||
| Item | CON | 0.5% | 2.5% | 5.0% | SEM | Treatment | Inclusion | Linear | Quadratic | ||
| Final body weight3, g | 1085 | 1111 | 1102 | 1090 | 18.2 | 0.737 | 0.455 | 0.426 | 0.950 | ||
| Average daily gain3, g | 65.4 | 68.8 | 64.9 | 65.6 | 1.26 | 0.164 | 0.504 | 0.097 | 0.153 | ||
| Total intake3, g | 928 | 934 | 947 | 938 | 16.5 | 0.878 | 0.565 | 0.851 | 0.591 | ||
| Gross energy intake4, kcal | 4304 | 4331 | 4389 | 4345 | |||||||
| Feed conversion ratio | 0.84 | 0.88 | 0.83 | 0.84 | 0.016 | 0.118 | 0.696 | 0.086 | 0.086 | ||
| Meat weight | |||||||||||
| Thigh, % of FBW5 | 19.5 | 19.1 | 18.8 | 19.1 | 0.37 | 0.637 | 0.294 | 0.943 | 0.470 | ||
| Breast, % of FBW | 17.6 | 17.9 | 18.7 | 18.7 | 0.48 | 0.265 | 0.136 | 0.255 | 0.500 | ||
| Inner fillet, % of FBW | 3.43 | 3.37 | 3.51 | 3.57 | 0.12 | 0.676 | 0.703 | 0.263 | 0.834 | ||
| Organ weight | |||||||||||
| Liver, % of FBW | 2.31 | 2.38 | 2.18 | 2.40 | 0.064 | 0.097 | 0.934 | 0.875 | 0.015 | ||
| Heart6, % of FBW | 0.46 | 0.49 | 0.43 | 0.47 | 0.017 | 0.132 | 0.635 | 0.385 | 0.032 | ||
| Spleen, % of FBW | 0.09 | 0.06 | 0.08 | 0.08 | 0.011 | 0.393 | 0.260 | 0.253 | 0.509 | ||
| Gizzard, % of FBW | 1.12 | 1.12 | 1.20 | 1.02 | 0.065 | 0.331 | 0.902 | 0.344 | 0.120 | ||
| Proventriculus, % of FBW | 0.42 | 0.45 | 0.42 | 0.38 | 0.033 | 0.584 | 0.833 | 0.184 | 0.917 | ||
| Cecum with fill, % of FBW | 0.47 | 0.47 | 0.38 | 0.43 | 0.063 | 0.690 | 0.508 | 0.674 | 0.372 | ||
| Cecum, % of FBW | 0.37 | 0.33 | 0.31 | 0.31 | 0.022 | 0.172 | 0.038 | 0.500 | 0.902 | ||
| Intestine, % of FBW | 3.22 | 3.22 | 3.26 | 3.05 | 0.148 | 0.762 | 0.826 | 0.446 | 0.492 | ||
1See Table 2 for the composition of treatment diets.
2Inclusion = inclusion of termite meal (control vs. mean of the 0.5%, 2.5%, and 5.0% treatments); Linear = linear effect of termite meal supplement; Quadratic = quadratic effect of the level of termite meal supplement.
3Covariate (initial body weight) P-values were < 0.001, 0.457, and 0.014 for final weight, average daily gain, and intake, respectively.
4Calculated value.
5FBW = fasting body weight.
6Covariate (final body weight) P = 0.006.
| Treatment1 | P-value | Contrast P-value2 | |||||||||
| Item3 | CON | 0.5% | 2.5% | 5.0% | SEM | Treatment | Inclusion | Linear | Quadratic | ||
| Total protein, g/dL | 2.90 | 2.92 | 2.86 | 2.70 | 0.110 | 0.500 | 0.570 | 0.175 | 0.714 | ||
| Albumin, g/dL | 1.50 | 1.50 | 1.48 | 1.42 | 0.044 | 0.546 | 0.522 | 0.218 | 0.716 | ||
| AST, U/L | 245 | 342 | 287 | 452 | 65.1 | 0.171 | 0.145 | 0.248 | 0.188 | ||
| ALT, U/L | 2.80 | 3.00 | 3.00 | 3.80 | 0.265 | 0.073 | 0.146 | 0.048 | 0.235 | ||
| LD, U/L | 856 | 1391 | 1302 | 3404 | 762.4 | 0.126 | 0.200 | 0.080 | 0.258 | ||
| ALP, U/L | 4424 | 5757 | 5206 | 5150 | 519.1 | 0.374 | 0.134 | 0.421 | 0.702 | ||
| γ-GT, U/L | 14.8 | 13.4 | 13.8 | 12.8 | 1.19 | 0.689 | 0.302 | 0.727 | 0.638 | ||
| Total bilirubin, mg/dL | 1.12 | 1.04 | 1.12 | 1.00 | 0.060 | 0.418 | 0.350 | 0.644 | 0.192 | ||
| Glucose, mg/dL | 217 | 229 | 214 | 214 | 6.81 | 0.358 | 0.828 | 0.129 | 0.369 | ||
| Na, mEq/L | 148 | 148 | 148 | 149 | 0.69 | 0.383 | 0.805 | 0.170 | 0.302 | ||
| K, mEq/L | 6.44 | 6.68 | 5.82 | 6.64 | 0.436 | 0.497 | 0.907 | 0.949 | 0.135 | ||
| Ca, mg/dL | 10.3 | 10.9 | 10.1 | 10.5 | 0.402 | 0.509 | 0.632 | 0.492 | 0.213 | ||
| Mg, mg/dL | 2.08 | 2.00 | 2.00 | 2.06 | 0.063 | 0.738 | 0.423 | 0.512 | 0.704 | ||
| P, mg/dL | 7.68 | 7.74 | 7.56 | 8.26 | 0.246 | 0.234 | 0.550 | 0.155 | 0.164 | ||
| Total cholesterol, mg/dL | 139 | 144 | 139 | 140 | 6.73 | 0.934 | 0.800 | 0.665 | 0.694 | ||
| Triglyceride, mg/dL | 16.4 | 15.6 | 14.2 | 13.2 | 1.28 | 0.330 | 0.183 | 0.205 | 0.900 | ||
| NEFA, mEq/L | 0.406 | 0.384 | 0.488 | 0.414 | 0.0448 | 0.411 | 0.667 | 0.642 | 0.124 | ||
| Phospholipid, mg/dL | 227 | 230 | 229 | 229 | 9.84 | 0.995 | 0.817 | 0.910 | 0.974 | ||
| BUN, mg/dL | 0.52 | 0.74 | 0.60 | 0.68 | 0.118 | 0.588 | 0.276 | 0.723 | 0.457 | ||
| Creatinine, mg/dL | 0.07 | 0.07 | 0.05 | 0.07 | 0.007 | 0.301 | 0.935 | 0.842 | 0.065 | ||
| β-Hydroxybutyric acid, µmol/L | 656 | 804 | 685 | 1104 | 116.9 | 0.058 | 0.142 | 0.088 | 0.079 | ||
1See Table 2 for the composition of treatment diets.
2Inclusion = inclusion of termite meal (control vs. mean of the 0.5%, 2.5%, and 5.0% treatments); Linear = linear effect of termite meal supplement; Quadratic = quadratic effect of level of termite meal supplement.
3AST = aspartate aminotransferase; ALT = alanine aminotransferase; LD = lactate dehydrogenase; ALP = alkaline phosphatase; γ-GT = γ-glutamyl transferase; NEFA = non-esterified fatty acid; BUN = blood urea nitrogen.
The overall effect of termite and fish meal supplementation (Exp. 2) is listed in Tables 9 and 10. Growth performance and meat weights were similar among treatments. Despite an isoprotein diet, CP of the feed residues was lower in the TM group than in the CTL and FM groups, suggesting selective feeding. Spleen weight tended to be greater (P = 0.082) in chickens treated with TM, whereas cecum fill weight was greater (P = 0.040) in FM chickens. Plasma total protein (P = 0.008), Na (P = 0.004), Ca (P = 0.025), and P (P = 0.028) concentrations were greater in TM and FM chickens than in CTL chickens.
| Treatment1 | ||||||
| Item | CTL | TM | FM | SEM | P-value | |
| Final body weight2, g | 1191 | 1178 | 1184 | 21.5 | 0.909 | |
| Average daily gain2, g | 76.1 | 74.3 | 74.5 | 1.73 | 0.736 | |
| Total intake2, g | 990 | 980 | 983 | 21.1 | 0.945 | |
| Gross energy intake3, kcal | 4552 | 4531 | 4545 | |||
| Average daily gain/daily feed intake | 0.92 | 0.91 | 0.91 | 0.013 | 0.699 | |
| Feed crude protein, % DM4 | 27.6 | 27.5 | 27.3 | 0.500 | 0.939 | |
| Residue crude protein, % DM | 26.5a | 25.9b | 26.7a | 0.187 | 0.009 | |
| Meat weight | ||||||
| Thigh, % of FBW5 | 19.3 | 19.6 | 19.0 | 0.281 | 0.355 | |
| Breast, % of FBW | 19.3 | 18.4 | 18.5 | 0.482 | 0.404 | |
| Inner fillet6, % of FBW | 3.50 | 3.67 | 3.44 | 0.091 | 0.237 | |
| Organ weight | ||||||
| Liver, % of FBW | 2.16 | 2.09 | 2.09 | 0.064 | 0.661 | |
| Heart, % of FBW | 0.618 | 0.574 | 0.614 | 0.0283 | 0.499 | |
| Spleen7, % of FBW | 0.078 | 0.093 | 0.072 | 0.0063 | 0.082 | |
| Gizzard, % of FBW | 1.24 | 1.27 | 1.30 | 0.071 | 0.869 | |
| Proventriculus, % of FBW | 0.42 | 0.42 | 0.50 | 0.034 | 0.204 | |
| Pancreas, % of FBW | 0.224 | 0.210 | 0.224 | 0.0130 | 0.674 | |
| Cecum with fill, % of FBW | 0.38b | 0.41b | 0.53a | 0.040 | 0.040 | |
| Cecum, % of FBW | 0.32 | 0.31 | 0.33 | 0.017 | 0.718 | |
| Intestine, % of FBW | 2.66 | 2.83 | 2.73 | 0.102 | 0.498 | |
1CTL = control, TM = termite meal, FM = fish meal; see Table 3 for the composition of treatment diets.
2Covariate (initial body weight) P = 0.002 for final weight, P = 0.009 for average daily gain, and P = 0.002 for total intake.
3Calculated value.
4DM = dry matter.
5FBW = fasting body weight.
6Covariate (final body weight) P = 0.052.
7Covariate (final body weight) P = 0.034.
a-bTreatment means without a common superscript letter differ by P < 0.05.
| Treatment1 | |||||
| Item2 | CTL | TM | FM | SEM | P-value |
| Total protein, g/dL | 2.57b | 2.90a | 2.82a | 0.067 | 0.008 |
| Albumin, g/dL | 1.43 | 1.53 | 1.53 | 0.037 | 0.116 |
| AST, U/L | 299 | 373 | 380 | 38.0 | 0.282 |
| ALT, U/L | 1.67 | 1.67 | 1.83 | 0.197 | 0.791 |
| LD, U/L | 1260 | 1421 | 1409 | 251.9 | 0.882 |
| ALP, U/L | 1582 | 1883 | 1610 | 160.4 | 0.368 |
| γ-GT, U/L | 14.2 | 17.2 | 14.5 | 1.60 | 0.124 |
| Total bilirubin, mg/dL | 0.87 | 0.98 | 0.95 | 0.051 | 0.283 |
| Glucose, mg/dL | 219 | 231 | 229 | 5.55 | 0.291 |
| Na, mEq/L | 140.8b | 150.8a | 149.5a | 1.91 | 0.004 |
| K, mEq/L | 6.27 | 6.45 | 6.37 | 0.342 | 0.931 |
| Ca, mg/dL | 9.72b | 10.38a | 10.28a | 0.164 | 0.025 |
| Mg, mg/dL | 2.02 | 2.10 | 2.00 | 0.081 | 0.653 |
| P, mg/dL | 7.10b | 7.92a | 7.75a | 0.201 | 0.028 |
| Total cholesterol, mg/dL | 123.7 | 140.0 | 137.3 | 5.46 | 0.115 |
| Triglyceride, mg/dL | 16.2 | 14.3 | 18.0 | 1.30 | 0.172 |
| NEFA, mEq/L | 0.652 | 0.505 | 0.555 | 0.0613 | 0.259 |
| Phospholipid, mg/dL | 234.3 | 238.8 | 248.0 | 9.32 | 0.584 |
| BUN, mg/dL | 0.43 | 0.50 | 0.47 | 0.040 | 0.521 |
| Creatinine, mg/dL | 0.068 | 0.068 | 0.073 | 0.0032 | 0.454 |
| β-Hydroxybutyric acid, µmol/L | 757.8 | 805.0 | 775.2 | 73.94 | 0.902 |
1CTL = control, TM = termite meal, FM = fish meal; see Table 3 for the composition of treatment diets.
2AST = aspartate aminotransferase; ALT = alanine aminotransferase; LD = lactate dehydrogenase; ALP = alkaline phosphatase; γ-GT = γ-glutamyl transferase; NEFA = non-esterified fatty acid; BUN = blood urea nitrogen.
a-bTreatment means without a common superscript letter differ by P < 0.05.
This study comprises the initial part of a project by the Moonshot Research and Development Program, funded by the Cabinet Office of the Government of Japan, under the concept ‘Creation of a Novel Food Production System Using Termites as a Poultry Feed Source.’ This innovative and sustainable system, which brings together wood, insects, and livestock to produce food, was named the WILFood system. The WILFood system adheres to the principles of the natural food chain; plants offer sustenance for insects, which in turn become prey for birds. The birds then conclude this chain by becoming part of the diet of the predators, and continuing the cycle. In essence, the WILFood system is a food production approach that is closely aligned with the natural food chain. Substantial challenges in this project included the mass production of H. sjostedti using unutilized plant biomass, a task human beings had never attempted. Examples of unutilized plant biomass include forest residues, wood waste from natural disasters, and food industry byproducts, such as bagasse, mushroom cultivation residues, and bamboo waste. These resources have shown promising results, as used by H. sjostedti (unpublished data). Another aspect of the project focuses on identifying functional substances from H. sjostedti. These termites exhibit extraordinary longevity[16], suggesting they may produce compounds of interest for human health. Thus, this project aims to achieve biological assimilation or cellulose transformation using termites, rather than simply seeking an alternative protein source for chicken feed.
TermitesTermites digest lignocellulose, the principal cell-wall component of woody plants, by a diverse community of bacterial, archaeal, and eukaryotic gut symbionts[17]. Their ability to digest lignocellulose, combined with their eusocial systems, has led to tremendous evolutionary success. The estimated global biomass of termites is 440 Mt (wet biomass), accounting for 40–60% of soil macrofaunal biomass[18,19]. Being rich in protein and fat[5], termites are often used for food, feed, and medicine, in Africa, Asia, and Latin America[6,7,8]. The majority (87%) of termites of interest to humans belong to the Termitidae family[6], which constructs huge soil mounds[20]. Other termite families utilized by humans include Hodotermitidae, Rhinotermitidae, and Kalotermitidae[6], which are often recognized as harmful pests[21]. They are traditionally collected by chipping mounds or underground nests[22], trapping during nuptial flights at night[23,24], or rearing on crop residues in clay pots for several weeks[25]. In contrast, the dampwood termite, H. sjostedti of the Hodotermopsidae family, is distributed from southern Japan to northern Vietnam[26,27]. Being the most basic and primitive termites[28] that completes its entire colony life in a single piece of dead wood in temperate rainforests[20], this species is thus most unlikely to be a pest of wooden structures. It is also one of the largest termite species, with a length of 0.5–1.2 cm for workers, 1.5–2.0 cm for soldiers (Fig. 2), and 2.0–2.2 cm for queens, and thus will have a high yield when reared for livestock feed.
Nutritional values and safetyCrude protein was slightly lower in termite meal (65.6% of DM) than in the reported value of the fish meal (67.8% of DM; Fish meal protein 62%) by Feedinamics[15]. Masagounder et al.[29] reported variations in CP in fish meal by year, species, and season of production, ranging from 58.6% to 69.8%. The CP composition of other termite species has also been reported to range from 8.8% to 66.7% of DM, depending on species and caste[7]. The termites used in this study feed on decayed Luchu pine wood; hence, their CP may also vary according to habitat and caste composition.
Crude fat content was greater in termite meal (16.0% of DM) than in fish meal (10.3% of DM)[15], but less than half of that in winged termite species (33–46%)[30]. A moderate crude fat content suggests unnecessary defatting when used in a chicken diet. Fat is an essential energy source and is usually added to broiler chicken feed. Saturated fatty acid utilization is affected by the presence of unsaturated fatty acids[31]. Nonetheless, neither the National Research Council[31] nor the National Agriculture and Food Research Organization (NARO)[32] state the amount of fat required for chickens; the former has claimed that linoleic acid (C18:2, n-6) and α-linolenic acid (C18:3, n-3) are the only metabolically essential fatty acids for fowl and should reach 0.83–1.25% of dietary concentrations in chickens[31].
Ash was lower in termite meal (5.2% of DM) than in fish meal (19.3% of DM)[15] owing to a lack of an internal calcified skeleton[7]. Similarly low were most of the main minerals (Table 1), except K, which was more abundant in termite meal (14.0 g/kg of DM) than in fish meal (8.2 g/kg of DM)[15]. This anomaly can be explained by termites feeding on pine trees rich in K (390 mg/kg)[33]. Termite meal had a high Cu content; whereas the amounts of Pb, Cd, and Hg were marginal, and As was not detected. According to NARO[34], the toxic level of Cu in growing chickens is 324 ppm, which is threefold higher than the value found in this study. Therefore, termite meal can be considered a safe alternative to fish meal for partial replacement.
The vitamin content of termite meal was generally low or below detection in this study, as reported also by Williams et al.[7]. Interestingly, however, vitamin B12 was much more abundant in termite meal (240 µg/kg of DM) than in fish meal (9.028 µg/kg of DM)[15]. Elevated vitamin B12 content has been reported previously in winged termite powder from Africa[30].
Dietary protein requirements are determined by the amino acid content[31]. In chicken diets, a balanced amino acid profile is essential to provide the amino acids required for the synthesis of structural and protective tissues and to support optimal metabolic function for maximal productivity[31]. Most amino acids were less abundant in termite meal than in fish meal[15]; however, that should not pose a risk if up to 5.0% of fish meal were replaced with termite meal. For example, methionine in termite meal accounted for less than half of that in fish meal, but when combined with cystine, which can be synthesized from methionine, they reached 55% of the value in fish meal[15]. Conversely, termite meal contained more tyrosine than fish meal[15]. A balanced amino acid composition is essential, although it can sometimes be difficult to achieve owing to specific dynamics between amino acids[31]. Monitoring CP and amino acid profiles is required when termite meal is used in chicken diets for more extended feeding periods. Amino acid corrections, such as DL-methionine supplementation, should be considered in future studies.
Termite meal contained twice as many MUFA as PUFA and SFA. Oleic acid (C18:1, n-9) was the main fatty acid in termite meal, followed by linoleic acid, which is an essential fatty acid in chickens[31]. Fish meal contained similar amounts of SFA, MUFA, and PUFA as termite meal, with 0.85% oleic acid but only 0.14% linoleic acid[15]. Fish meal also contains very long-chain fatty acids (0.02% C22:0, 0.54% C22:1, 0.18% C22:5, and 0.46% C22:6)[15], which were below the detection level in termite meal. Cheseto et al.[30] detected 33 fatty acids in winged termite oils using gas chromatography coupled with mass spectrometry. Overall, the termite meal used in this study contained more fatty acids than fish meal due to its higher EE percentage.
Feeding trialsThe primary purpose of the feeding trial in this study was to evaluate the adverse effects (e.g., reduced weight gain, organ abnormalities, or health issues) and growth performance of young chickens, in which fish meal was partly replaced by termite meal. A preliminary study (see Table S2) showed that feed intake was lower at 10% than at 6.7% termite meal inclusion, which would be a major concern if the diet were applied for an extended period. Therefore, Exp. 1 was conducted at lower substitution levels. Exp. 2 was designed to compare termite and fish meals at a 2.5% supplementation level. As the dampwood termite H. sjostedti is a novel feed source for poultry diets, there is no information on its metabolizable energy. Accordingly, gross energy rather than metabolizable energy was used in this study. There were no apparent adverse effects on the growth performance, feed intake, feed efficiency, or meat production in either Exp. 1 or Exp. 2. Egbewande et al.[35] demonstrated that replacing fish meal with termite meal at 5% of the starter diet for 4 weeks and then at 2.5% of a finisher diet for another 4 weeks to raise broilers, which resulted in lower feed intake and smaller final live weight. In this study, birds were treated for 13 d (Exp. 1) or 12 d (Exp. 2); however, a more extended experimental period would determine the long-term effects of substituting fish meal for termite meal. Interestingly, lower CP content in the TM residues of Exp. 2 suggests that female broiler chickens selected termite meal over other ingredients. Organ weights relative to fasting body weight were also similar among treatments in both experiments, with a few exceptions. In Exp. 1, chicken liver weight was greater for 0.5% and 5.0% than for 2.5% termite meal, which coincided with higher ALT at 5.0% but not at 0.5% treatment. Increased ALT levels may indicate hepatic dysfunction. Although the value was within the normal range for broiler chickens (3.35–78.0 U/L)[36], hepatic burden or immune stress should be monitored in future studies. The plasma concentration of β-hydroxybutyric acid increased between 0.5% and 5.0% termite meal. β-hydroxybutyric acid is a source of ketone bodies in the liver, which provide energy to extrahepatic tissues[37]. In Exp. 2, the TM group presented a slightly higher spleen weight than the CTL and FM groups. An immune reaction or liver burden can cause an enlarged spleen; however, this did not occur in Exp. 1 at the 2.5% and 5.0% treatment levels. Cecal weight decreased as the termite meal substitution level increased in Exp. 1. In Exp. 2, the cecal fill percentage was greater in the FM group than in the CTL and TM groups. The reason for these conflicting results is unclear, but the differences were negligible (0.01–0.1% level). The above results allude to possible liver burden in chickens fed 5.0% termite meal.
Preliminary feeding trials in this study were limited by a small sample size and short duration. A long-term and large-scale trial using a nutritionally controlled diet for both sexes is necessary to determine the effects of substituting fish meal with termite meal.
ConclusionsThe dampwood termite H. sjostedti meal was evaluated for its safety, nutrient values, and adverse effects in growing chickens to explore the potential of termite (H. sjostedti) meal. Nutritional values differed from those of a comparable meal obtained from the African winged termite. H. sjostedti meal had a high protein content, comparable to that of fish meal. Amino acid concentrations were comparable to the reported values of fish meal, except for methionine (lower) and tyrosine (higher). The termite meal contained more MUFA, followed by PUFA and SFA. Minerals and vitamins were generally low, except for vitamin B12, K, and Cu, which were much more abundant. All these disparities would not constitute a problem if termite meal accounted for up to 5% of the supplementation in a chicken diet. None of the minerals tested was present at toxic levels. Female broiler chickens displayed similar growth, intake, and meat or liver weight with either termite or fish meal. Blood plasma profiles revealed only minor discrepancies when 0–5.0% fish meal was replaced by termite meal. Plasma ALT and β-hydroxybutyric acid were more abundant at 5.0% substitution with termite meal, which suggested potential hepatic burden. Nevertheless, no adverse effects were observed in any of the traits tested. In conclusion, H. sjostedti termite meal has strong potential as a source of chicken feed. Monitoring the nutritional value of termite meal is required owing to fluctuations in substrate and caste composition. Longer feeding trials with larger numbers of birds will confirm the long-term effects of termite meal supplementation.
This work was supported by the Cabinet Office, Government of Japan; Moonshot Research and Development Program for Agriculture, Forestry, and Fisheries (funding agency: Bio-oriented Technology Research Advancement Institution); and Project JPJ009237. We thank Dr. Mamoru Takata, Assistant Professor of the Laboratory of Insect Ecology, Graduate School of Agriculture, Kyoto University, for his assistance.
Conceptualization, Kenji Matsuura, Hiroyuki Hirooka, Eisuke Tasaki, and Yuki Mitaka; methodology, Hajime Kumagai, Yoko Tsukahara, and Shozo Tomonaga; validation, Yoko Tsukahara, Hajime Kumagai, Hiroyuki Hirooka, and Kenji Matsuura; formal analysis, Yoko Tsukahara, Hajime Kumagai, and Nami Tomonaga; investigation, Yoko Tsukahara, Nami Tomonaga, Hajime Kumagai, and Shozo Tomonaga; resources, Kenji Matsuura, Hiroyuki Hirooka, Hajime Kumagai, Yuki Mitaka, and Eisuke Tasaki; data curation, Yoko Tsukahara, Nami Tomonaga, and Hajime Kumagai; writing – original draft preparation, Yoko Tsukahara; writing – review & editing, Yoko Tsukahara, Hajime Kumagai, Hiroyuki Hirooka, Kenji Matsuura, Shozo Tomonaga, Eisuke Tasaki, and Yuki Mitaka; visualization, Yoko Tsukahara and Kenji Matsuura; supervision, Kenji Matsuura, Hiroyuki Hirooka, and Hajime Kumagai; project administration, Kenji Matsuura, Yoko Tsukahara, and Yuki Mitaka; funding acquisition, Kenji Matsuura.
The authors declare no conflict of interest.
