Weizmannia coagulans functions as a heat-stress protective probiotic via modulation of the pyridoxal-oxalate interaction

doi:10.21203/rs.3.rs-2759976/v1

Download PDF

Article

Weizmannia coagulans functions as a heat-stress protective probiotic via modulation of the pyridoxal-oxalate interaction

https://doi.org/10.21203/rs.3.rs-2759976/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Global warming has become an urgent issue, and heat stress protective materials should be developed to manage human life and animal husbandry. Here, the heat stress protective effects of the thermostable probiotic Weizmannia coagulans SANK70258 were clarified using a broiler model under thermal conditions in an artificial climate facility. Probiotic exposure promoted growth under the thermal conditions. Based on machine learning (ML) data obtained after omics analysis, structural equations and causal inference for broiler growth under heat-stress estimated the strong relationship with pyridoxal involved in the reduction of hepatic oxalic acid, which is responsible for cold sensitivity. In addition, the relative abundance of the genus Pseudomonas, which may be decrease pyridoxal, were at a low level after exposure of W. coagulans under heat-stress. Thus, this study highlights a novel aspect of W. coagulans as a probiotic candidate that modulates the mechanism of heat tolerance related to the cold sensitivity.

Biological sciences/Microbiology/Industrial microbiology/Environmental microbiology/Food microbiology

Health sciences/Gastroenterology/Gastrointestinal system/Microbiota

Health sciences/Risk factors

Health sciences/Health care/Quality of life

Biological sciences/Physiology/Metabolism/Metabolomics

Heat stress

probiotics

growth

broiler

Weizmannia coagulans

According to the IPCC, the problem of global warming is an extremely serious problem, threatening the entire ecosystem¹ and causing severe damage to human health and economic activity^2,3. To remain healthy under high-temperature conditions and maintain homeostasis, organisms have evolved mechanisms to withstand the increase in external temperature, but the burden on the living body is enormous⁴. Heat stress is a problem in the livestock industry, where many heat waves have caused catastrophic economic losses due to increased mortality. Heat stress can also lead to infections that require the use of antibiotics⁵. The overuse of antibiotics is a global problem, and the emergence of antibiotic-resistant bacteria caused by the overuse and misuse of antibiotics in livestock has become apparent worldwide. The World Health Organization (WHO) has launched an action plan against antimicrobial resistance (AMR)⁶ that includes the eliminating the use of antibiotics for growth promotion in livestock by 2015. Although these issues, that is, combating global warming and reducing the use of antibiotics, appear to constitute distinct fields of research, they both have negative impacts on the food and health chains of humans and animals.

Therefore, research into the effects of heat stress on animals and the relevant countermeasures is essential. For example, studies evaluating behavioural and physiological effects^7–11 are actively underway using relatively manageable livestock. Research on the effects on gene expression, examining genetic as well as epigenetic factors, is progressing, with studies conducted in chickens¹², pigs¹³, cows¹⁴, etc. Adverse effects on immunity and reproductive activities have been also studied^15–17. Research into nutrients and probiotics is underway to overcome these challenges. For example, the administration of vitamin C, with antioxidant activity¹⁸, or dihydromyricetin¹⁹ and the administration of selenium²⁰ or alpha-lipoic acid²¹, which help to produce antioxidants as dietary supplements, can reduce mortality and the deleterious effects of heat stress in livestock. Research on the effect of probiotic administration in domestic animals has also advanced^22,23, and studies on the use of Bacillus subtilis alone²⁴, and in combination with lactic acid bacteria²⁵ have been reported. However, as global warming continues, such studies will not necessarily lead to a breakthrough, and new technologies and perspectives will be needed, including those that can be used in combination with the useful materials that have been developed to date.

Here, a broiler model reared under artificially controlled thermal environmental conditions^26,27 was used to estimate the effect of probiotic administration on environmental adaptability to heat. The Weizmannia coagulans (synonym: Bacillus coagulans) SANK70258, which is a highly heat-stable spore-forming lactic acid bacterium with slight environmental variation and has been traditionally used in food or feed since 1949, was used as a probiotic. Previous findings^28–30 obtained under normal conditions suggested an effect on metabolic function in the intestine, so an omics analysis of the intestine and liver was performed. Based on the results, two types of machine learning (ML) methods were applied to extract important feature factors related to heat tolerance from the omics data. These features of the network structure were statistically evaluated by covariance structure analysis and causal inference. Our findings verified the known trends for some effects of W. coagulans administration, and we newly observed that the pyridoxal-oxalic acid (oxalate) metabolic network may play a role in the group of factors related to resistance to cold sensitivity. These estimations may indicate part of the molecular mechanism of heat resistance and will aid in understanding the relationship between environmental stress and gastroenterology.

Scheme and breeding results of this study

An outline of the feeding trial and the flow of analysis for this study is shown in Fig. 1. As shown in Fig. 1a and Fig. S1, broilers were maintained at artificially controlled high temperatures ranging from 25°C to 32°C. As shown in Fig. 1b, all the animals were kept in one cage until the 10th day, and 3 animals were retained in each cage after the 10th day. After 21 days, one bird per cage was bred. For the artificial temperature control, breeding management was performed in a special house for chick (battery brooder) for 10 days. Table 1 shows the results of breeding under these conditions.

As a result, rearing was stopped on day 10, and there was no significant difference among the groups. According to the data collected 10 days after the heat treatment, although there was no significant difference compared to day 21, the body weight tended to increase in the heat-stress test group (H_coagulans) compared to the heat-stress control group (H_control). From days 21 to 49, the body weight increased significantly (p < 0.05) in the heat-stress test group (H_coagulans) compared to the heat-stress control group (H_control). The feed efficiency also tended to improve. A similar trend was observed in the conventional control group (N_control) under non-heat stress conditions. This was confirmed in both under heat treatment from day 10–49 and during all the breeding periods from days 0–49.

As shown in Fig. 1c, the feeding performances were evaluated, and omics analyses of the colon contents and liver tissues were performed on samples obtained upon necropsy on day 49. The comprehensive data were classified by ML. In addition, the mechanism of action was predicted using a structural equation model and causal inference for the statistical causal structure associated with feed intake.

Metabolomics of the intestinal contents and liver

Based on the results of the faecal metabolome analysis and the liver metabolome analysis, substances with p value of 0.05 or less (based on a comparison of the two groups under high-temperature conditions) are shown on the heat map (Fig. 2). A tendency towards strong changes in faeces was observed under three conditions, for example, a greater decrease in substances was observed under heat stress; similar results were observed in the liver. Among the substances with p values of 0.01 or less, in faeces, galactose, lactitol, cytidine, and sebacic acid increased, 3-hydroxypropionic acid, adenine decreased, and in the liver, erythrulose, 2-aminoisobutyric acid, glucose-6-phosphate decreased. Evaluation of the correlation between these metabolites in the enterohepatic group showed that there were large differences between the normal group (N_control) (Fig. S2), the heat-stress control group (H_control) (Fig. S3a) and the heat-stress test group (H_coagulans) (Fig. S3b). In particular, the levels of metabolite that showed a positive correlation with colonic faecal carbohydrate were low in the heat-stress control group (H_control) (Fig. S3a) under the heat treatment and increased in the normal group (N_control) (Fig. S2). The levels further increased in the heat stress test group (H_coagulans) (Fig. S3b) under the heat stress conditions. Pathway analysis was performed further to predict metabolic function (Tables S1 and S2). As shown in Fig. S4, there was a trend towards activation of galactose metabolism, amino sugar and nucleotide sugar metabolism, starch and sucrose metabolism and tryptophan metabolism in the control (N_control) for the normal group and in the heat stress test group (H_coagulans).

Based on these results, it was suggested that the metabolic systems of the heat-stress control group (H_control) under heat-stress conditions were relatively similar to those of the normal group (N_control) under normal conditions and the test group (H_coagulans) under heat-stress conditions. The test group (H_coagulans) was characterized by vitamin B6 metabolism and fructose and mannose metabolism.

Extraction of key metabolites and development of network models

Based on these metabolomic data, essential factors were extracted using machine learning (ML). The results are shown in Fig. 3. As a result of the random forest analysis, the MeanDecreaseGini value, which was an index of the characteristic quantity, was classified in colonic excreta (Fig. 3a) and liver (Fig. 3b). In faeces from the large intestine, carbohydrates were ranked higher. Organic acids such as glucose-6-phosphatase, which is involved in the glycolytic pathway, were ranked highest in the liver. In the XGBoost analysis, the importance value, which became an index of the characteristic quantity, was high only for pyridoxine, mannitol, and oxalic acid in the faeces and liver (Fig. 3c).

A comparison of the concentrations of metabolites extracted only from random forests in the three groups (Fig. S5) showed that the behaviour of galactose and lactitol as carbohydrates, acetylglycine as an organic acid, and 3-hydroxypropionic acid, 3-hydroxyisobutyric acid and adenine as a constituent of nucleic acids in faeces was similar between normal conditions (N_control) and the heat-stress test group (H_coagulans). Although there were significant differences, galactose and lactitol tended to increase in both groups, and acetylglycine, 3-hydroxypropionic acid, 3-hydroxyisobutyric acid, and adenine tended to decrease. Sebacic acid only increased significantly in the heat-stress test group (H_coagulans) among the carbohydrates, maltitol, and among organic acids. The metabolite levels in the liver (Fig. S6) were similar between the normal group (N_control) and the heat-stress test group (H_coagulans), but not necessarily similar for 2-aminobutyric acid and erythrulose, whose levels decreased significantly in the test group (H_coagulans). The change in glucose-6-phosphate levels was similar between normal conditions (N_control) and the heat-stress test group (H_coagulans).

Characteristic factors that were identified in both the random forest and XGBoost analyses were that faecal mannitol (p = 0.0007) as a carbohydrate, faecal pyridoxal as a vitamin, and liver oxalic acid (p = 0.001) as an organic acid (Fig. 4a). Comparison of concentrations among the three groups showed that mannitol and pyridoxal were significantly increased in the feces of only the heat-stress test group (H_coagulans).

Based on these classifications, structural equation modeling and causal inference were used to assess the degree of association between groups of factors, rather than to establish a clear causal relationship between these groups. In addition, when selecting the statistical method, we assessed whether the distribution was a Gaussian distribution and whether equal variance was obtained (Fig. S7). Calculations for the three factors showed a mixture of data with a Gaussian or non-Gaussian distribution. Therefore, an evaluation was performed on a statistical model assuming a Gaussian or non-Gaussian distribution.

By calculation of the structural equation showing the optimum value for the maximum likelihood method assuming a Gaussian distribution (Table S3 and Fig. 4b), it was demonstrated that W. coagulans administration had a positive relationship with faecal mannitol and growth and W. coagulans administration had a negative relationship with liver oxalic acid. Faecal pyridoxal was in a positive relationship with W.coagulans.

A non-Gaussian distribution was then assumed and independent causal analysis was applied to evaluate the relationship with environmental factors. The result (Fig. 4c) showed that the metabolites with a positive relationship with heat stress and growth were pyridoxal and mannitol, and the metabolite with a negative relationship with heat stress and growth was oxalic acid.

Effects on the fecal bacterial population

The bacterial population in faeces of the colon was then analyzed. There was no statistically significant difference in the diversity of the flora (Fig. S8a and S8b) or at the phylum level (Fig. S8c). As a result, there was no statistically significant difference in the diversity of the flora (Fig. S8a and S8b) or at the phylum level (Fig. S8c). At the genus level, a decreasing trend in Lactobacillus was observed under heat treatments (Fig. 5a).

A decreasing trend for the genus Pseudomonas and an increasing trend for the genera Flavonifractor and Romboutsia (Fig. 5b and S9a) were observed especially in the heat-stress test group (H_coagulans), although a few samples showed significant changes with p values less than or equal to 0.2. A negative correlation was observed between the genera Lactobacillus and Flavonifractor when evaluating the correlations of four factors (Fig. S9b). In addition, because the data showed a mixture of Gaussian and non-Gaussian distributions (Fig. S10), the evaluation was performed in a statistical model assuming a Gaussian or non-Gaussian distribution.

First, the structural equation was calculated for the maximum likelihood method assuming the Gaussian distribution (Table S4). In the model that showed the optimum value, the administration of W.coagulans showed a positive relationship with the genera Romboutsia and Flavonifractor. The administration of W.coagulans showed a negative relationship with the genus Pseudomonas, indicating a negative relationship with growth (Fig. 5c).

A non-Gaussian distribution was then assumed and the relationship with environmental factors was assessed computationally by independent analysis (Fig. 5d). As a result, the genera Lactobacillus and Flavonifractor did not show a strong relationship with growth. A positive growth value (value = 6.53) was observed for the genus Romboustia, whereas a more negative growth value (value = -24.99) was observed for the genus Pseudomonas.

In this study, among the three groups, the breeding result for the heat-stress test group (H_coagulans) was significantly better than that for the heat-stress control group (H_control), and the finding was equivalent to that for the normal group (N_control) under normal temperature conditions. Prediction of the mechanism of action by ML and statistical structural model evaluation confirmed large differences among the three groups (Fig. 6). A bird's eye view of the data showed that at the metabolite level (Fig. 6a), mannitol in colonic faeces was common except the heat-stress control group (H_control), although there were some differences in the levels of mannitol in colonic faeces. Pyridoxal was familiar to the heat-stress conditions (H_control and H_coagulans), and oxalic acid in the liver tended to have a minimal effect only in the heat-stress test group (H_coagulans). Although there was a difference in the abundance of the genus Romboustia in the bacterial flora level (Fig. 6b), it was observed in the heat-stress test group (H_coagulans) and the control under heat stress conditions (H_control). The effect of the genus Pseudomonas tended to extremely small in the heat-stress test group (H_coagulans) compared to the heat stress control group (H_control).

Administration of W. coagulans may control heat tolerance, although heat stress negatively affects the growth inhibition and quality of life of animals as described above (Fig. 7a). Interestingly, it was recently suggested that oxalic acid increases sensitivity to reactive oxygen species (ROS) and temperature by inhibiting proline hydroxylation by transient receptor potential ankyrin1 (TRPA1)³¹, which is involved in several sensory processes across animal species³², including the sensing of taste and temperature³¹ (Fig. 7b). It is known that the accumulation of oxalic acid harms living organisms. Oxalic acid, also known as the cause of urinary stones, is formed from glyoxalic acid. Generally, high levels of oxalic acid occur in primary hyperoxaluria type 1 (PH1), which is characterized by progressive renal failure due to the accumulation of insoluble calcium oxalate. The cause of PH1 is a deficiency of glyoxylate aminotransferase (AGT), a pyridoxal 5’-phosphate (PLP)-dependent enzyme³³. As a way of controlling this production of oxalic acid, the use of pyridoxal or pyridoxine results in the formation of glycine without an increase in oxalic acid (https://www.sciencedirect.com/topics/nursing-and-health-professions/glyoxylic-acid)³⁴. In this study, faecal pyridoxal was selected from the ML and statistical structural equation model (Fig. 4). Although pyridoxine was excluded from the ML analysis due to its low concentration, its level was significantly higher in the heat-stress test group (H_coagulans) (Fig. S11).

It is therefore reasonable to assume that the cause of the low oxalic acid levels in the heat-stress test group (H_coagulans) was the high levels of pyridoxal and pyridoxine detected in the faeces. Furthermore, a low abundance of the genus Pseudomonas in the intestine is mentioned as a factor when considering the enriched factors. Although the genus Pseudomonas is known to possess pyridoxal kinase, ATP and pyridoxal react to form ADP, H₂, and pyridoxal 5'-phosphate (https://www.uniprot.org/uniprotkb/J2RGI2/entry)³⁵.

In this context, the behaviour of the metabolites involved in the metabolism of pyridoxine – oxalic acid was confirmed, including metabolites present at low concentrations (< 0.1 µ g/mg) that were not necessarily significantly different (Fig. S11). ATP and pyridoxal are converted to ADP, hydrogen, and PLP by the enzymatic reaction of AGT. In this metabolomic analysis, the PLP concentrations was below the lower detection limit. Adenosine triphosphate (ATP) and adenosine diphosphate (ADP) could not be measured, faecal adenine levels tended to be reduced (p = 0.0588) in the heat-stress test group (H_coagulans). Since glyoxylic acid is converted to glycine by pyridoxal as above described³⁴, the concentrations of glyoxylic acid and glycine in the faeces and the liver was confirmed. As a result, the level of the liver glyoxylic acid tended to decrease, at least in the liver, in the heat-stress test group (H_coagulans), although that of glycine did not change significantly. Although a study on genetically modified animals was not performed, it can be important phenomena that the increase in faecal vitamin B₆ and the decrease in liver oxalic acid in the broiler chicken model here were significant, which could be important. Thus, it was notable that the contents of oxalic acid for a reaction to alleviate excessive oxidative stress was decreased in the heat-stress test group (H_coagulans).

Based on the combination of the above selected omics data, a hypothesis derived from this study was visualized in Fig. 7. Although heat stress can inhibit growth, immunity, reproductive activities, and quality of life for the organism^15–17, the combined use of W. coagulans made it resistant to heat stress (Fig. 7a). The mechanism predicted by this study is illustrated in the conceptual diagram in Figs. 7b and 7c. In other words, the increase in growth and mannitol after the administration of W. coagulans is assumed to be an indicator of effective carbohydrate utilization, as previously reported²⁹. Second, it is expected that the reduction phenomenon of the genus Pseudomonas attenuates the enzymatic reaction of their PLPs as described above, resulting in the prevention of the reduction of pyridoxal/pyridoxine in the intestine (https://www.uniprot.org/uniprotkb/J2RGI2/entry). In fact, when the genus Peudomonas was reduced by administration of W. coagulans, the concentration of pyridoxal/pyridoxine in the gut was higher (Figs. 4a and S11). Given these reactions, it follows that the growth promotion by W. coagulans in this case is expected to regulate the temperature sensitivity of TRPA1³¹, since oxalic acid negatively regulates the TRPA1 response³¹, as described above. To support this hypothesis, we also examined the amount of TRPA1 present at the protein level to see whether it is the concentration of oxalic acid that regulates prolyl hydroxylases (PHDs) or the actual amount of TRPA1 present itself. The results did not show statistically significant differences under the three conditions (FigS12). Next, the protein levels of HIF1α, which is known to affect TRPA-1 expression and act in a co-beneficial way during heat stress in various animal species^36–40, were examined. The results confirmed a trend toward higher protein levels of HIF1α during heat stress. In view of the above, the cascade in Fig. 7c is envisioned in the present data: the possibility that PHDs are known enzymes that regulate the reactivity of TRPA-1 and that their regulation is controlled by oxalic acid. Since it is difficult to assess the response of PHDs in tissues, the protein abundance of HIF1α as an indirect marker molecule was examined to assess the response (Fig. S12). The protein level of HIF1α involved in the induction of TRPA1 expression was observed, although the different tendencies between normal and heat stress conditions were observed (p = 0.1211). The presence of TRPA-1 is common from insects to humans³², and this perspective may also take into account perspectives beyond the model animal. From this reactionary perspective, antioxidants or peroxides are important metabolites that reactive oxygen species (ROS) themselves take into account. Notably, ascorbic acid concentrations in faeces and liver tended to be higher in the heat-stress test group (H_coagulans) (p < 0.1) (Fig. S13a). However, the levels of malondialdehyde, a stable indicator of lipid peroxides⁴¹, were not necessarily affected in the liver and white adipocyte tissue (Fig. S13b). In addition, there is a decreasing trend in methylmalonic acid^42–44 and paraxanthine⁴⁵, which may have an inhibitory effect on liver function (Fig. S13b). However, the significance of the relationship with these changes was not found in this study. Based on these data, it is hypothesized that the cascade is not caused by low oxygen stress, but by changes in the characteristic metabolic system of the intestinal environment under the heat tolerance.

In addition, a causal inference also raised the possibility of the genus Romboustia being a factor associated with increased body weight (Fig. 5d). The extent of the effect of thus genus was not measured in this study. Although causal inference was not positioned as an essential factor, it indicated an increase in the genus Flavonifractor, albeit at a low level, in the heat-test group (H_coagulans). As W. coagulans has recently attracted attention as a the beneficial bacterial candidate, it is assumed that the diversity of the host's intestinal flora is altered.

According to this study, the mechanism by which W. coagulans is involved in heat resistance is the one shown in Fig. 7. Thus, these mechanisms imply a potential for drug-free, probiotic-derived relief from heat stress and other forms of antioxidant stress. The understanding of the novel mechanism can be gained by studying in detail the relationship with the enteral bacterial flora as a basis.

Broiler management

All animal experiments were approved by the Institutional Animal Care and Use Committee of Chiba University (DOU27-131, DOU28-157, DOU30-331, DOU1-335) and carried out according to the guidelines of Chiba Prefectural Livestock Research Center⁴⁶. Broiler chicks (chunky) were conventionally maintained with feed available ad libitum available feed (CP, 22%; ME, 3,100 kcal/kg) in a battery brooder (width 85 cm x depth 85 cm x height 25 cm) for 10 days and thereafter separated into 3 groups until day 21 after day 10. Randomly selected chicks were maintained (n = 3 per cage), wherein the chicks were selected to adjust for the restricted feeding area (width 22 cm x depth 44 cm x height 45 cm). From Day 21 until day 49, each chick was maintained in another restricted feeding area (n = 1 per cage) (width 44 cm x depth 56 cm x height 65 cm). All chicks were orally administered 50 mg/L of doxycycline KS (Kyouritsu Seiyaku Co., Ltd.), belonging to the tetracycline class, in their water as an antibiotic until 6 days after hatching in their water. W. coagulans SANK70258 (Lacris-10) (Mitsubishi Chemical Corp.) was administered in the drinking water at a concentration of at least 10⁵ CFU/ml as a final concentration by ad libitum drinking for the first 10 days and thereafter at 0.005% concentration as a feed additive (at least 10⁵ CFU/g).

Preparation of faecal samples for HPLC

All the faecal samples obtained were transported at − 20°C and thereafter stored at − 60°C - −80°C until analysis. The faecal samples were prepared according to the manufacturer’s protocol⁴⁷ with some modifications. In brief, chick faeces (200–400 mg) were mixed with a 9-fold volume of Milli-Q water for 10 min. After centrifugation at 15,000 rpm, the supernatant was filtated with a 0.45 µm Millex-HA filter (SLHA033SS) (Millipore, USA). Finally, the filtated solutions were prepared for HPLC analysis of the HPLC instrument. To determine the concentrations of acetate, propionate, butyrate, and lactate, the content of frozen fresh fecal samples was examined by using an HPLC Prominence instrument (Shimadzu, Japan), which was equipped on an ion-exclusion column (Shim-pack SCR-102H) with an electrical conductivity detector (CDD-10A_VP) (Shimadzu, Japan).

The measurement conditions were adjusted as follows: mobile phase, 5 mM p-toluenesulfonic acid; buffer, 5 mM p-toluenesulfonic acid, 20 mM Bis-Tris, and 0.2 mM EDTA-4H; temperature, 40°C; flow rate, 0.8 ml/min.

Preparation of faecal samples for GC/MS/MS

To measure water-soluble metabolites, freeze-dried faecal samples were processed. A GC/MS/MS instrument and GCMS-TQ8030 triple quadrupole mass spectrometer (Shimazu, Japan) were used according to a previously described method⁴⁸ with some modifications. Since chicken excrement tends to be indigestible, some pretreatment steps were applied. The freeze-dried faeces were disrupted with zirconia beads using Micro Smash MS-100 (Tomy Seiko, Japan). Then, 5 mg of the faecal sample was suspended in 150 µL of Milli-Q water containing 15 µL of an internal standard (1 mM 2-isopropyl maleric acid), and then 150 µL methanol and 60 µL chloroform were added. These solutions were mixed and incubated at 37°C for 30 min with shaking at 1200 rpm. After centrifugation at 16,000 g for 5 min at room temperature, 250 µL of the supernatant (aqueous layer) was transferred to a new tube, and 60 µL of chloroform was added. After mixing, the solutions were incubated at 37°C for 30 min with shaking at 1200 rpm. Each solutions was centrifuged at 16,000 × g for 5 min at room temperature, and 250 µL of the supernatants was transferred to a new tube, and 200 µL of Milli-Q water was added to the tube. The solutions were centrifuged at 16,000 g for 5 min at room temperature, and 125 µL of the supernatant was evaporated to dryness using a vacuum evaporator system (CentriVap Centrifugal Vacuum Concentrator, Labconco, USA) for 20 min at 40°C and lyophilized using a freeze dryer (Taitec, Japan). The dried extracts were stored at -80°C until the subsequent pretreatment was performed immediately before the measurement. The dried extracts were first methoxylated with 40 µL of 20 mg/mL methoxyamine hydrochloride (Sigma-Aldrich, USA) dissolved in pyridine. The samples were incubated for 90 min at 30°C with shaking and were then sialylated with 20 µL of N-methyl-N-trimethylsilyl-trifluoroacetamide (MSTFA) (GL Science, Japan) for 30 min at 37°C with shaking. After derivatization, the samples were centrifuged at 16,000 × g for 5 min at room temperature, and the supernatant was transferred to a glass vial for GC–MS/MS measurement. GC–MS/MS analysis was carried out on a Shimadzu GCMS-TQ8030 triple quadrupole mass spectrometer (Shimadzu, Japan) with a capillary column (BPX5; SGE Analytical Science). The GC oven was programmed as follows: 60°C held for 2 min, increased to 330°C (15°C/min), and finally maintained at 330°C for 3.45 min. GC was performed in constant linear velocity mode at 39 cm/s. The detector and injector temperatures were 200°C and 250°C, respectively. The injection volume was set at 1 µL with a split ratio of 1:30. Data processing was performed using LabSolutions Insight (Shimadzu, Japan).

Enrichment analysis

Pathway analysis and enrichment analysis for the detected characteristic metabolites (p < 0.1; 1.5-fold change vs. the data of the heat-stress control group) were performed using MetaboAnalyst 5.0 (https://www.metaboanalyst.ca)^49–51.

DNA preparation from faecal samples

Total DNA was prepared from the faeces as previously reported⁵² with some modifications. Approximately 100 mg of lyophilized faeces was disrupted with zirconia beads using Micro Smash MS-100 (Tomy Seiko, Tokyo, Japan), and suspended in 600 µL of a solution containing 10 mM Tris-HCl (pH 8.0), and 1 mM EDTA (TE). We transferred 475 µL of faecal suspension into a 1.5-mL microcentrifuge tube. This suspension was incubated with 15 mg/mL lysozyme for 1 h at 37°C. Netxt, achromopeptidase (FUJIFILM Wako Pure Chemical Corp., Osaka, Japan) was added at a final concentration of 2,000 units/mL, and the sample was further incubated for 30 min at 37°C. Then, 25.7 µL of proteinase K and sodium dodecyl sulfate were added at a final concentrations of 1 mg/mL and 1%, respectively, and then the mixed solution was incubated at 55°C for 1 h. These enzymatic treatments were carried out with shaking (1,000 rpm). The resulting lysate was treated with phenol/chloroform/isoamyl alcohol, and DNA was precipitated with ethanol. The DNA pellet was rinsed with 70% ethanol, dried, and dissolved in 300 µL of TE. Then, RNase was added at a final concentration of 0.1 mg/mL, and the solution was incubated at 37°C for 30 min. Then, an equal volume of a 20% polyethylene glycol 6,000 solution containing 2.5 M NaCl was added. After incubation for 30 min on ice, DNA was pelleted by centrifugation, rinsed, and then dissolved in TE.

16S rRNA sequencing analysis

16S rRNA gene analysis was carried out by PCR amplification with the barcoded forwards primer 27F-mod (5′-AGRGTTTGATYMTGGCTCAG-3′) and reverse primer 338R (5′TGCTGCCTCCCGTAGGAGT-3′) primers targeting the hypervariable regions (V1-V2) of the 16S rRNA gene according to a previous report⁵². The PCR reaction mix comprised 4 µL of DNA template, 0.5 µL of ExTaq polymerase (Takara Bio Inc., Kusatsu, Japan), 5 µL of ExTaq buffer, 5 µL of dNTPs, and 10 pmol each of the forwards and reverse primers and brought to a total volume of 50 µL with DNA-free water. Amplicons were generated with 30 cycles of 98 ℃ for 30 s, 55 ℃ for 45 s, and 72 ℃ for 2 min were purified using AMPure XP (Beckman Coulter Inc., Brea, CA, USA). Multiplexed amplicon sequencing was performed under the conditions described below. According to the manufacturer’s instructions, the samples were sequenced using a 454 GS FLX/Junior. Reads with an average quality value < 25 or inexact matches to both universal primers were filtered out. Finally, filtered reads were used for further analysis. These reads were then sorted according to average quality value and grouped into operational taxonomic units (OTUs) using UCLUST (http://www.drive5.com/). The sequence identity threshold was 96% for the 454 data. The representative sequences of the generated OTUs were subjected to a homology search against the databases mentioned above using the GLSEARCH program for taxonomic assignments. For an assignment at the phylum, genus, and species levels, sequence similarity thresholds of 70%, 94%, and 97% were applied, respectively. Indices for α-diversity, namely community richness (Chao1) and diversity (Shannon and Simpson), were calculated, and indices for β-diversity were also estimated using UniFrac analysis with weighted and unweighted principal coordinate analysis (PCoA). Alla16S rRNA gene data sets were deposited in the GenBank Sequence Read Archive database.

Correlation analyses

The relative values of dominant and/or characteristic faecal metabolites were visualized by constructing a correlation heatmap after the Pearson correlation coefficient was calculated among the fecal metabolites by the R software as previously described ^29,53,54.

Machine learning

Feature factors of metabolome data were extracted by supervised ML, Randomforest ⁵⁵, one of ML with bagging (bootstrap aggregating) ⁵⁶, and XGBoost ⁵⁷, one of ML with extreme gradient boosting ⁵⁶, respectively. Even with a small number of observations, a fixed subsample can be generated by random sampling using the bootstrap method, and the selected explanatory variables can be used to calculate the best classifying factors and their threshold values. In this study, the advantage of this ML characteristic was utilized to extract one characteristic group of bacteria and metabolites under the limited experimental conditions. The characteristic components were selected using the R software packages "randomForest" (https://cran.r-project.org/web/packages/randomForest/index.html) and “xgboost” (https://cran.r-project.org/web/packages/xgboost/index.html). Finally, these extracted feature factors were visualized as a bubble chart using the python modules “numpy” and “matplotlib” (https://matplotlib.org/stable/index.html).

Estimation of causal inference and structural equation across the animal species

To assess structural universality across the animal species, Direct LiNGAM, which is a linear non-Gaussian acyclic model and a non-Gaussian independent component analysis for estimating causal relationships⁵⁸, was performed as statistical causal inference for a group of factors selected by machine learning. DirectLiNGAM was established with Python code on the website (https://github.com/cdt15/lingam) as previously described ^53,59. In this study, the following packages were used on Python version 3.11.2: Cartopy 0.21.1, graphviz 0.20.1, libarchive 0.4.7, lingam 1.7.1, numpy 1.23.4, pandas 1.5.3, and pygam0.8.0. Directed acyclic graphs (DAGs) based on data calculated by Direct LiNGAM were visualized (Force Atlas with Noverlap) by Gephi (version 0.9.2)(https://gephi.org). The modularity class was automatically discerned by Gephi.

In addition, structural equation modeling (SEM) for confirmatory factor analysis (CFA) was mainly performed for a group of components with a large number of DAGs. SEM was conducted using the package “lavaan”^60,61 of R software as previously described ^53,59. The analysis codes were referred to the website (https://lavaan.ugent.be). Since CFA needs a hypothesis, the groups selected by association analysis were utilized as factors for a latent construct of metabolites and microbiota. The models as hypotheses were statistically estimated using maximum likelihood (ML) parameter estimation with bootstrapping (n = 1000) by the functions ‘lavaan’ and ‘sem’. Model fit was assessed by the chi-squared p value (p > 0.05, nonsignificant), comparative fit index (CFI/cfi) (> 0.9), Tucker–Lewis index (TLI/tli) (> 0.9), goodness-of-fit index (GFI/gfi) (> 0.95), adjusted goodness-of-fit index (AGFI/agfi) (> 0.90), root mean square error of approximation (RMSEA/rmsea) (< 0.05), and standardized root mean residual (SRMR/srmr) (< 0.08) as indices of good model fit ⁶². In selecting the optimal structural equation, when the above conditions were met, an equation with little difference between the values of GFI and AGFI was adopted. Next, the model with a low value of Akaike information criterion (AIC) was adopted in the case with many candidate models. The path diagrams of the good model were visualized using the package “semPlot” of R software ⁶³. In addition, the visualization of the degree of influence for each animal species was performed by Chord diagram using the package “circlize” of R software. To avoid the effects of experimental conditions on the difference, all data were calculated based on the median value (M) of data for the same component and sorted to 0 (< M) and 1 (> M).

Immunoblots

50 mg of tissue was transferred in 150 µL of ice-cold PBS-T containing protease inhibitor: 10µL/mL of PMSF (SigmaP7626) (Sigma, USA), 10µL/mLsodium orthovanadate (Sigma S6508) (Sigma, USA), and 20µL/mL of protease inhibitor cocktail (Thermo Scientific Halt Protease Inhibitor Singl-Use Cocktail #78430) (Thermo Scientific, USA). The samples were homogenized on ice using using sonicator Ultrasonic processor XL2020 (MISONIX, USA), and centrifuged at 4°C, 10,000 × g for 5 minutes and to collect the supernatant. The protein concentrations of the supernatants were determined using Thermo Scientific Pierce BCA Protein Assay kit #23227 (Thermo Scientific, USA).

The crude protein solution were separated by electrophoresis on 7.5% sodium dodecyl sulfate polyacrylamide gels and then transferred to PVDF membranes (GE Healthcare Life science, Sweden). The membranes were blocked with PVDF Blocking Reagent (TOYOBO, Japan) at room temperature for 1 hour. The primary antibodies were rabbit polyclonal TRPA1 antibody (19124-1-AP)(proteintech group, Japan) diluted 1:1500 or rabbit polyclonal HIF-1 alpha antibody (bs-20399R)(Bioss andibodies, USA) diluted 1:1500 and the secondary antibody was horseradish peroxidase-conjugated goat anti-rabbit IgG antibodies diluted 1:12000 (Cell Signaling Technology, USA). GAPHDH was detected using rabbit polyclonal GAPDH antibody (10494-1-AP) (proteintech group, Japan) diluted 1:10000. Immunoblots were incubated with primary antibodies at 4°C overnight and peroxidase-conjugated secondary antibodies at room temperature for 1 hour. The protein signals were detected using Western Lightning TM ECL Pro (Perkin-Elmer Life Sciences, USA). The chemiluminescence scanner C-DiGiT (Scrum, Japan) and Image Studio was used to measure the protein signals.

Analysis of malondialdehyde concentration

The concentrations of malondialdehyde (MDA) in the liver and white adipocyte tissues were determined using an MDA assay kit, Bioxytech MDA-586 (Percipio Bioscience, CA, USA) according to the manufacturer’s protocol as previously described ⁴¹. In brief,

50 mg of tissue was homogenised in BHT-containing PBS-T (containing 0.05% tween 20) using sonicator Ultrasonic processor XL2020 (MISONIX, USA), incubated at 45°C for 60 minutes, and centrifuged at 10,000 × g for 10 minutes. Then the supernatant was measured at 586 nm.

Statistical analyses

Data for colon faecal bacterial flora and metabolites concentrations were analyzed as follows: the Shapiro-Wilk test was used to evaluate the Gaussian distribution. In addition, the Bartlett’s test was also used to evaluate equal variances and to select parametric and non-parametric analyses. Appropriate methods dependent upon data sets were selected as follows: for parametric analysis, ANOVA followed by Tukey’s post hoc test was used; as non-parametric analysis, Kruskal-Wallis one-way analysis of variance was performed, followed by the Steel-Dwass test was performed. These calculation data were prepared by using the R software (version 3.6.2, 4.0.5, and 4.2.2) and GraphPad Prism software (version 9.1.2 and 9.2.0) (GraphPad Software, USA), and Microsoft Office (version 16.66.1). Data in pathway analysis were calculated and visualized using MetaboAnalyst 5.0^49–51. Significance was declared when p < 0.05, and a tendency was assumed at 0.05 ≤ p < 0. 20. Data are presented as the means ± SEs.

Data availability. Raw files of the bacterial V1–V2 16S rRNA data are deposited in the DNA Data Bank of Japan (DDBJ), under NCBI Bio-Project accession numbers PRJDB10703 (PSUB013670)(BioSample accession numbers: SAMD00261459-SAMD00261461, SAMD00253848-SAMD00253854). In case that you need information that is not listed here, please contact the corresponding authors.

Acknowledgements

This research was partly supported by a Grant-in-Aid from the Chiba Prefectural Federation of Small Business Association, based on the foundation for manufacturing (Mono-Zukuri) of the Ministry of Economy, Trade, and Industry in Japan (Grant ID 25121200402). We are grateful to Ms. Shoko Iwata (Chiba University) for sampling and preparing of the feed. We would like to extend special thanks to Dr. Toshiya Katsuragi, Koga, Watanabe, Mr. Keita Kashiwa, Mr. Takayuki Maeda (Mitsubishi Chemical Corp.) for special advice, Mr. Masayuki Waki for special coordination of staff and faculty, and Mr. Keiji Takahashi, Hiroko Okada, Mr.Tadahiko Fujisaki, Mr. Yasuhiro Takahashi, Mr. Yoshiaki Tada, and Mr. Harukazu Gotoh (Chiba Prefectural Livestock Research Center) for special advice and management of breeding facilities, sampling, and preparation of feed. In addition, we thank to Mrs. Miki Asakura and Mrs Ayumi Ito (RIKEN IMS) for providing technical support for GCMS/MS experiments.

Author contributions

Conceived and designed the experiments: K.I, H.M.,Y.T., and H.O. Performed the animal experiments: K.I.,H.M.,M.M, C.I.,F.H., and T.I. Analyzed the data: K.I., H.M., M.M., C.I., Y.N, W.S., T.S., A.K, N.T., T.N., and J.K. Contributed reagents/materials/analysis tools: K.I., W.S., C.S., Y.T., J.K., and H.O. Wrote the paper: K.I., H.M., W.S., J.K., and H.O. Revised the manuscript: M.M., C.I., Y.N., W.S., T.S., A.K., N.T., R.M., R.Y., H.K., and H.O.

Competing interests

The authors declare no competing interests.

Stocker, T. F. et al. IPCC-2013: climate change 2013: the physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change., 1535 (2013).
Koteswara Rao, K. et al. Projections of heat stress and associated work performance over India in response to global warming. Sci Rep 10, 16675 (2020).
Delworth, T. L., Mahlman, J. D. & Knutson, T. R. Changes in heat index associated with CO2-induced global warming. Chimatic Change 43, 369–386 (1999).
Cramer, M. N., Gagnon, D., Laitano, O. & Crandall, C. G. Human temperature regulation under heat stress in health, disease, and injury. Physiol Rev 102, 1907–1989 (2022).
Gathiram, P., Wells, M. T., Brock-Utne, J. G., Wessels, B. C. & Gaffin, S. L. Prevention of endotoxaemia by non-absorbable antibiotics in heat stress. Journal of Clinical Pathology 40, 1344–1368 (1987).
Holloway, K. A., Rosella, L. & Henry, D. The Impact of WHO Essential Medicines Policies on Inappropriate Use of Antibiotics. PLoS One 11, e0152020 (2016).
Becker, C. A., Aghalari, A., Marufuzzaman, M. & Stone, A. E. Predicting dairy cattle heat stress using machine learning techniques. J Dairy Sci 104, 501–524 (2021).
Laporta, J. ADSA Foundation Scholar Award: Early-life exposure to hyperthermia: Productive and physiological outcomes, costs, and opportunities. J Dairy Sci 104, 11337–11347 (2021).
Ishida, S., Osawa, T., Yamaguchi, S., Yamazaki, T. & Hagiya, K. Suitable indicator of heat stress for genetic evaluation of heat tolerance in Holstein cows in Japan. Anim Sci J 93, e13688 (2022).
West, J. W. Effects of Heat-Stress on Production in Dairy Cattle. J. Dairy Sci. 86, 2131–2144 (2003).
Brugaletta, G., Teyssier, J. R., Rochell, S. J., Dridi, S. & Sirri, F. A review of heat stress in chickens. Part I: Insights into physiology and gut health. Front Physiol 13, 934381 (2022).
Goel, A., Ncho, C. M. & Choi, Y. H. Regulation of gene expression in chickens by heat stress. J Anim Sci Biotechnol 12, 11 (2021).
He, Y., Maltecca, C., Tiezzi, F., Soto, E. L. & Flowers, W. L. Transcriptome analysis identifies genes and co-expression networks underlying heat tolerance in pigs. BMC Genet 21, 44 (2020).
Dauria, B. D. et al. Genetic effects of heat stress on milk fatty acids in Brazilian Holstein cattle. J Dairy Sci 105, 3296–3305 (2022).
Siddiqui, S. H., Kang, D., Park, J., Khan, M. & Shim, K. Chronic heat stress regulates the relation between heat shock protein and immunity in broiler small intestine. Sci Rep 10, 18872 (2020).
Bagath, M. et al. The impact of heat stress on the immune system in dairy cattle: A review. Res Vet Sci 126, 94–102 (2019).
Schneiders, G. H. et al. Heat stress reduces sexual development and affects pathogenesis of Eimeria maxima in meat-type chickens. Sci Rep 10, 10736 (2020).
Pardue, S. L., Thaxton, J. P. & Brake, J. Role of ascorbic acid in chicks exposed to high environmental temperature. J Appl Physiol (1985) 58, 1511–1516 (1985).
Wang, H. L. et al. Dihydromyricetin attenuates heat stress-induced apoptosis in dairy cow mammary epithelial cells through suppressing mitochondrial dysfunction. Ecotoxicol Environ Saf 214, 112078 (2021).
Khan, A. Z. et al. Selenium-enriched probiotics improve hepatic protection by regulating pro-inflammatory cytokines and antioxidant capacity in broilers under heat stress conditions. J Adv Vet Anim Res 6, 355–361 (2019).
Wasti, S., Sah, N., Lee, C. N., Jha, R. & Mishra, B. Dietary supplementation of alpha-lipoic acid mitigates the negative effects of heat stress in broilers. PLoS One 16, e0254936 (2021).
Salem, H. M. et al. Heat stress in poultry with particular reference to the role of probiotics in its amelioration: An updated review. J Therm Biol 108, 103302 (2022).
Ringseis, R. & Eder, K. Heat stress in pigs and broilers: role of gut dysbiosis in the impairment of the gut-liver axis and restoration of these effects by probiotics, prebiotics and synbiotics. J Anim Sci Biotechnol 13, 126 (2022).
Cramer, T. A. et al. Effects of probiotic (Bacillus subtilis) supplementation on meat quality characteristics of breast muscle from broilers exposed to chronic heat stress. Poult Sci 97, 3358–3368 (2018).
Song, J. et al. Effect of a probiotic mixture on intestinal microflora, morphology, and barrier integrity of broilers subjected to heat stress. Poult Sci 93, 581–588 (2014).
Miyamoto, H. Symbiotic Microorganisms (Eds. by Hiroshi Ohno), Dojin Bioscience series 27, Kagaku-Dojin (Kyoto, Japan), 247–258 (2016).
Nakayama, O. & Sakaguchi, K. Studies on the spore-bearing lactic acid forming bacilli: part I. Journal of the Agriculural Chemistry Society of Japan 23, 513–517 (1950).
Sasaki, K. et al. Bacillus coagulans SANK 70258 suppresses Enterobacteriaceae in the microbiota of ulcerative colitis in vitro and enhances butyrogenesis in healthy microbiota. Appl Microbiol Biotechnol 104, 3859–3867 (2020).
Ito, K. et al. Noninvasive fecal metabolic profiling for the evaluation of characteristics of thermostable lactic acid bacteria, Weizmannia coagulans SANK70258, for broiler chickens. J Biosci Bioeng 134, 105–115 (2022).
Watanabe, Y., Fukuda, R. & Tsutsumi, T. Effects of Intake of Bacillus coagulans SANK 70258,“Spore-forming Lactic Acid Bacteria” on Bowel Movement in Healthy Volunteers -A Randomized, Double-blind, Placebo-controlled, Parallel-group Study-. Japanese Pharmacology and Therapeutics 46, 1549–1558 (2018).
Miyake, T. et al. Cold sensitivity of TRPA1 is unveiled by the prolyl hydroxylation blockade-induced sensitization to ROS. Nat Commun 7, 12840 (2016).
Kurganov, E., Saito, S., Tanaka Saito, C. & Tominaga, M. Requirement of extracellular Ca(2+) binding to specific amino acids for heat-evoked activation of TRPA1. J Physiol 595, 2451–2463 (2017).
Cellini, B., Montioli, R., Paiardini, A., Lorenzetto, A. & Voltattorni, C. B. Molecular Insight into the Synergism between the Minor Allele of Human Liver Peroxisomal Alanine:Glyoxylate Aminotransferase and the F152I Mutation. J Biol Chem 284, 8349–8358 (2009).
Varalakshmi, P. & Richardson, K. E. The effects of vitamin B-6 deficency and hepatectomy on the synthesis of oxalate from glycolate in the rat. Biochimica et Biophysica Acta 757, 1–7 (1983).
Brown, S. D. et al. Twenty-one genome sequences from Pseudomonas species and 19 genome sequences from diverse bacteria isolated from the rhizosphere and endosphere of Populus deltoides. J Bacteriol 194, 5991–5993 (2012).
Zhang, X., Wang, D. & Liu, J. Hypoxia-inducible factor-1alpha is involved in the response to heat stress in lactating dairy cows. J Therm Biol 112, 103460 (2023).
Brugaletta, G. et al. Effect of Cyclic Heat Stress on Hypothalamic Oxygen Homeostasis and Inflammatory State in the Jungle Fowl and Three Broiler-Based Research Lines. Front Vet Sci 9, 905225 (2022).
Klumpen, E. et al. Reactive oxygen species (ROS) and the heat stress response of Daphnia pulex: ROS-mediated activation of hypoxia-inducible factor 1 (HIF-1) and heat shock factor 1 (HSF-1) and the clustered expression of stress genes. Biol Cell 109, 39–64 (2017).
Baird, N. A., Turnbull, D. W. & Johnson, E. A. Induction of the heat shock pathway during hypoxia requires regulation of heat shock factor by hypoxia-inducible factor-1. J Biol Chem 281, 38675–38681 (2006).
Hatano, N. et al. Hypoxia-inducible factor-1alpha (HIF1alpha) switches on transient receptor potential ankyrin repeat 1 (TRPA1) gene expression via a hypoxia response element-like motif to modulate cytokine release. J Biol Chem 287, 31962–31972 (2012).
Miyamoto, H. et al. Thermophile-fermented compost as a possible scavenging feed additive to prevent peroxidation. J Biosci Bioeng 116, 203–208 (2013).
Rosenberg, L. E., Lilljeqvist, A. & Hsia, Y. E. Methylmalonic aciduria: metabolic block localization and vitamin B 12 dependency. Science 162, 805–807 (1968).
Pillai, N. R. et al. Liver transplantation in propionic and methylmalonic acidemia: A single center study with literature review. Mol Genet Metab 128, 431–443 (2019).
Fernandes, C. G. et al. Experimental evidence that methylmalonic acid provokes oxidative damage and compromises antioxidant defenses in nerve terminal and striatum of young rats. Cell Mol Neurobiol 31, 775–785 (2011).
Gressner, O. A., Lahme, B., Siluschek, M. & Gressner, A. M. Identification of paraxanthine as the most potent caffeine-derived inhibitor of connective tissue growth factor expression in liver parenchymal cells. Liver Int 29, 886–897 (2009).
Ito Kayo, Y. T., Honda Fuyuko, Okada Hiroko, Takahashi Keiji and Waki Masayuki. Heat Adaptation Technology by Precise Nutrition Management for Each Fattening Stage in Broilers. Bulletin of the Chiba Prefectural Livestock Research Center No.19, 7–15 (2019).
Tonda, R. M., Rubach, J. K., Lumpkins, B. S., Mathis, G. F. & Poss, M. J. Effects of tannic acid extract on performance and intestinal health of broiler chickens following coccidiosis vaccination and/or a mixed-species Eimeria challenge. Poult Sci 97, 3031–3042 (2018).
Sato, K. et al. Obesity-Related Gut Microbiota Aggravates Alveolar Bone Destruction in Experimental Periodontitis through Elevation of Uric Acid. mBio 12, e0077121 (2021).
Xia, J., Psychogios, N., Young, N. & Wishart, D. S. MetaboAnalyst: a web server for metabolomic data analysis and interpretation. Nucleic Acids Res 37, W652-660 (2009).
Xia, J. & Wishart, D. S. Web-based inference of biological patterns, functions and pathways from metabolomic data using MetaboAnalyst. Nat Protoc 6, 743–760 (2011).
Xia, J. & Wishart, D. S. Metabolomic data processing, analysis, and interpretation using MetaboAnalyst. Curr Protoc Bioinformatics Chap. 14, Unit 14 10 (2011).
Kim, S. W. et al. Robustness of gut microbiota of healthy adults in response to probiotic intervention revealed by high-throughput pyrosequencing. DNA Res 20, 241–253 (2013).
Miyamoto, H. et al. A potential network structure of symbiotic bacteria involved in carbon and nitrogen metabolism of wood-utilizing insect larvae. Science of the total Environment https://doi.org/https://doi.org/10.1016/j.scitotenv.2022.155520 (2022).
Miyamoto, H. et al. An agroecological structure model of compost – soil - plant interactions for sustainable organic farming. ISME Communications (in press).
Breiman, L. Random Forests. Machine Learning 45, 5–32 (2001).
Miyamoto, H. & Kikuchi, J. An evaluation of homeostatic plasticity for ecosystems using an analytical data science approach. Comput Struct Biotechnol J 21, 869–878 (2023).
Chen, T. & Guestrin, C. XGBoost. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 785–794 (2016).
Shimizu, S., Hoyer, P. O., Hyvarinen, A. & Kerminen, A. A Linear Non-Gaussian Acyclic Model for Causal Discovery. Journal of Machine Learning Research, 2003–2030 (2006).
Miyamoto, H. et al. Computational estimation of sediment symbiotic bacterial structures of seagrasses overgrowing downstream of onshore aquaculture. Environ Res https://doi.org/10.1016/j.envres.2022.115130, 115130 (2022).
Rosseel, Y. lavaan: An R Package for Structural Equation. Journal of Statistical Software 48, 1–36 (2012).
Rosseel, Y. et al. Latent Variable Analysis. R package version0.6-9 (2021).
Hooper, D., Coughlan, J. & Mullen, M. R. Structural equation modelling: guidelines for determining model fit. Electron J Business Res Methods 6, 53–60 (2008).
Epskamp, S., Stuber, S., Nak, J., Veenman, M. & Jorgensen, T. D. Path Diagrams and Visual Analysis of Various SEM Packages' Output R package version1.1.2 (2019).

Table 1 is available in the Supplementary Files section

(Not answered)

DatafinalLac.xlsx
Source data
HeatSupplementaryInformation.pdf

Download PDF

Version 1

posted

You are reading this latest preprint version

Weizmannia coagulans functions as a heat-stress protective probiotic via modulation of the pyridoxal-oxalate interaction

Status:

Version 1

Abstract

Figures

Introduction

Results

Scheme and breeding results of this study

Metabolomics of the intestinal contents and liver

Extraction of key metabolites and development of network models

Effects on the fecal bacterial population

Discussion

Methods

Broiler management

Preparation of faecal samples for HPLC

Preparation of faecal samples for GC/MS/MS

Enrichment analysis

DNA preparation from faecal samples

16S rRNA sequencing analysis

Correlation analyses

Machine learning

Estimation of causal inference and structural equation across the animal species

Immunoblots

Analysis of malondialdehyde concentration

Statistical analyses

Declarations

References

Table

Additional Declarations

Supplementary Files

Status:

Version 1