Molecular

white


Introduction
The silkworm Bombyx mori is an important economic insect and serves as a model organism for biology and biotechnology research.Usually, silkworms are raised mainly for silk production, but they can also be used for medical resources and foodstuff (Liang et al. 2018, Altomare et al. 2020).China's annual production of cocoons ranks first in the world, and the output value of the silk industry is more than 30 billion dollars per year (Jiang et al. 2021).The quality and yield of cocoons depended not only on the silkworm strain but also on the rearing situation, including silkworm health, climatic conditions, and absorption of nutrients.Gut microbiota has been found to be closely related to digestive absorption, immune defense, growth and development, and silk production of B. mori (Cani 2014, Zhang et al. 2022, Yuan et al. 2023).Therefore, studying the intestinal microflora of silkworms and its relationship with the host's health and immune response can provide valuable insight for improving the health status and cocoon yield of silkworms.
Silkworms have been domesticated for about 5,000 years.Despite the efforts to maintain a clean rearing environment and provide high-quality food for silkworms, their intestines harbor a diverse and abundant community of microorganisms (Anand et al. 2010, Chen et al. 2018a, Hou et al. 2018).Over the past few decades, several studies have investigated the intestinal bacteria of B. mori based on culture-independent techniques.For instance, the effects of developmental stage (Sun et al. 2016, Hou et al. 2018), gender (Sun et al. 2016), rearing temperature (Sun et al. 2022), food sources (Dong et al. 2018), harmful substances (Li et al. 2016(Li et al. , 2021)), and virus infection (Sun et al. 2016, Kumar et al. 2019, Shi et al. 2021) on silkworm intestinal flora have been reported.Because of the close relationship between virus invasion and bacterial infection per os, the impact of virus invasion on intestinal microorganisms of silkworms has become a research hotspot.It has been reported that B. mori bidensovirus, cypovirus, and nucleopolyhedrovirus infection could alter the composition and diversity of intestinal microflora of the susceptible silkworm strains (Sun et al. 2016, Kumar et al. 2019, Shi et al. 2021).However, little progress has been made in understanding the differences in gut bacteria between the virus-resistant and susceptible strains of silkworm.
Bombyx mori nucleopolyhedrovirus (BmNPV) is a highly significant pathogen that poses a substantial threat to the sericulture industry.In nature, BmNPV mainly infects silkworms through oral infection.The occlusion body of BmNPV enters the digestive tract of the silkworm after being ingested by B. mori; then the polyhedrin is dissolved in the strong alkaline environment of the midgut, releasing thousands of occlusion-derived virus (one virion phenotype) and causing infection of the epithelial cells of the silkworm midgut (Blissard and Theilmann 2018).Generally, BmNPV infection causes damage to the intestinal cells of silkworms, which leads to the outbreak of intestinal bacterial diseases of the silkworm.On the other hand, BmNPV infection triggers the immune responses and the production of antiviral and antibacterial peptides in silkworms (Furukawa et al. 2007, Chen et al. 2018b, Dong et al. 2019), which may affect the composition and diversity of the intestinal flora of B. mori.It has been reported that midgut microbiota could modulate the resistance to baculovirus infection in Helicoverpa armigera (Yuan et al. 2021), while viral infection also resulted in changes in the gut microbiome of silkworms (Sun et al. 2016, Kumar et al. 2019, Shi et al. 2021).Therefore, virus infection can be regulated by the intestinal microbiota of the host, and in turn, the latter is also affected by the host immune response induced by the former.
In this study, we identified the differences in composition and diversity of fecal bacterial communities of BmNPV-resistant and susceptible strain silkworms, as well as the changes in fecal microbiota in silkworms in response to BmNPV infection using 16S rRNA gene PCR amplicons sequencing.Our findings suggested that different silkworm strains resulted in different gut bacterial compositions, and the dynamic changes of intestinal microbial population response to BmNPV infection also differed.This research will improve our understanding of the effects of viral infection on gut microflora assessed by feces, and also provide theoretical help for the breeding of resistant strains and healthy rearing of silkworms.Although feces have been used as a proxy material to analyze intestinal flora in some insect species, the microbiota in feces is not necessarily equivalent to that in the intestine.Therefore, objectively speaking, the conclusions obtained from this study using feces to assess the intestinal microbiota have some limitations.

Silkworm Preparation and Sample Collection
The BmNPV (T3 strain), BmNPV-resistant strain A35 and susceptible strain P50 silkworm were maintained in our lab, and all larvae were reared with fresh mulberry leaves as described (Wang et al. 2017, Cao et al. 2021).Individual fifth-instar third-day larvae of A35 and P50 silkworm were peroral droplet feeding with 5 μl of BmNPV suspended in sterile water (1.0 × 10 6 polyhedra/ml) using a micropipette, respectively, and the control samples were treated with sterile water.Silkworms that did not completely inhale the droplets or spit them out were abandoned.Then, all larvae that were effectively fed droplets were fed with fresh mulberry at 24 ± 1 °C.At 24 h postinfection (hpi), each larva was cleansed with 70% ethanol for surface sterilization and then placed into a sterile 50-ml centrifuge tube to collect feces at 24 ± 1 °C.When the silkworm excreted one feces, we immediately collected the feces in a sterile environment and froze it in liquid nitrogen.Five feces were collected from each larva in about 30 min (pooled as one sample) of healthy and BmNPVinfected silkworms of each strain (10 replicates for each group), respectively.All samples were frozen with liquid nitrogen immediately after collection.

DNA Extraction and 16S rDNA Sequencing
The total genomic DNA of each sample was extracted using MagPure Soil DNA LQ Kit (Magen, China) following the manufacturer's instructions.The concentration of DNA was verified with NanoDrop 2000 spectrophotometer (Thermo Fisher) and agarose gel electrophoresis.The genome DNA was used as a template for PCR amplification of the V3-V4 region of bacterial 16S rRNA genes, with 343-F (5ʹ-TACGGRAGGCAGCAG-3ʹ), 798-R (5ʹ-AGGGTATCTAATCCT-3ʹ) primers and Tks Gflex DNA Polymerase (Takara, China).After being amplified and purified for 2 rounds, the final products were quantified using Qubit dsDNA assay kit (Life Technologies, Q32854, United States).Equal amounts of purified amplicon were pooled and sequenced on the Illumina MiSeq platform (OE Biotech Company, China) with 2 paired-end reads.

Sequence Processing
Paired-end reads were preprocessed using Trimmomatic software (V 0.35) to cut off ambiguous bases (N) and low-quality sequences with average quality scores below 20 or lengths below 50 bp (Bolger et al. 2014).After trimming, paired-end reads were assembled using FLASH software (V 1.2.7) (Reyon et al. 2012).Sequences were performed further denoising as follows: reads with ambiguous, homologous sequences or below 200 bp were abandoned.Reads with 75% of bases above Q20 were retained using QIIME software (V 1.8.0) (Caporaso et al. 2010).Then, all clean reads were subjected to primer sequences removal and clustered to generate operational taxonomic units (OTUs) using VSEARCH software (V 2.4.2) with a 97% threshold value (Rognes et al. 2016).The representative read of each OTU was selected using the QIIME package.All representative reads were annotated and blasted against the Silva database (Version 132) using the RDP classifier (confidence threshold: 70%) (Wang et al. 2007).

Data Analysis
The Chao1 and "observed species" indices calculated in QIIME were used to assess species diversity among samples.Principal coordinate analysis (PCoA) was conducted using R language based on a weighted unifrac algorithm.Wilcoxon test was used to analyse the differences between the 2 groups.PICRUSt2 software was used to perform functional prediction analysis.The phylogenetic trees on the basis of representative sequences of the top 50 OTUs were inferred using MEGA11 with the neighbor-joining method (1,000 bootstrap tests).

Overview of the 16S rDNA Sequencing-Derived Dataset
After quality control, a total of 2,548,398 clean tags were obtained, and the number of reads from 40 samples ranged from 45,242 to 86,390.After removing chimeric sequences, a total of 1,890,313 valid tags were obtained for further analysis.The number of valid tags ranged from 28,426 to 69,207, and the average length of valid tags ranged from 408.67 to 416.85 bp.The total number of valid tags in P50 control (Pc) group, P50 infection (Pv) group, A35 control (Ac) group, and A35 infection (Av) group were 486,934, 449,155, 489,421, and 464,803, respectively (Supplementary Table S1).And a total of 18,443 OTUs were identified, the count of unique OTUs in 40 samples was between 1,197 and 4,703, and the percentage of OTUs annotated at order and genus ranged from 92.68% to 98.41% and from 39.91% to 68.34%, respectively (Supplementary Table S2).Venn diagrams showed that the number of core OTUs of all 40 samples was 177, and the number of unique OTUs of each sample ranged from 1,020 to 4,526 (Fig. 1A).However, there were 6,096 core OTUs between 4 groups, and the number of unique OTUs of Pc, Pv, Ac, and Av groups was 1,493, 1,632, 1,846, and  1,519, respectively (Fig. 1B).The richness curves revealed that the sequencing depth of the bacterial community sufficiently covered the bacterial diversity in the gut of the silkworm (Supplementary Fig. S1A).And the diversity index of Av group was significantly less than that of Ac group and Pv group (Supplementary Fig. 1B).All of the raw data were deposited in the Sequence Read Archive database of NCBI under the Bioproject accessions PRJNA796270.
When comparing control samples of resistant and susceptible strains (Ac vs. Pc), we identified 652 differential OTUs and 58 differential genera, while the differential OTUs and genus of BmNPVinfected samples of resistant and susceptible strains (Av vs. Pv) were 777 and 55, respectively (Table 1).And the BmNPV infection resulted in 625 differential OTUs and 45 differential genera in Av vs. Ac, while 590 differential OTUs and 46 differential genera in Pv vs. Pc (Table 1).

Taxonomic Composition of Bacteria from Resistant and Susceptible Strains of B. mori Response to BmNPV Infection
Among all of the 40 samples, a total of 34 bacterial phyla (including 889 genera) were detected, and the number of phyla/ genus of Pc, Pv, Ac, and Av groups were 32/745, 33/727, 32/766, and 32/722, respectively (Supplementary Table S3).The 10 most represented phyla were the Proteobacteria, Bacteroidetes, Firmicutes, Actinobacteria, Gemmatimonadetes, Acidobacteria, Fusobacteria, Epsilonbacteraeota, Spirochaetes, and Tenericutes (Supplementary Fig. S2).Among them, the Proteobacteria were dominant in Pc and Pv groups, with the average relative abundances of 39.39% and 33.54%, respectively.While the Bacteroidetes were dominant in Ac and Av groups with the average relative abundances of 35.06% and 34.27%, respectively (Table 2).At the same time, the Proteobacteria were the second most abundant phylum in Ac and Av groups and represented 34.29% of the OTUs in Ac group and 31.84% of those in Av group.The Bacteroidetes were the second most abundant bacterial taxa in Pc and Pv groups and comprised 30.57% and 33.52% of the corresponding communities (Table 2).

Diversity of Bacterial Communities in the Gut of BmNPV-Infected Resistant and Susceptible Silkworm
Among 328 bacterial families detected from susceptible strain silkworm P50, BmNPV infection resulted in a marked change 3. Diversity of bacterial communities in the gut of Bombyx mori.A) Principal coordinate analysis (PCoA) of weighted UniFrac distances of microbial communities from healthy and infected silkworms of P50 and A35 strain.B) Heat map showing the differences of bacteria families of healthy silkworms between P50 and A35.Asterisks indicate significant differences between mean values of the P50 and A35 groups, *P < 0.05; **P < 0.01 (Wilcoxon test).

Functional Enrichment Profiles of Intestinal Bacteria
To predict the functional clusters of OTUs among different groups, the Clusters of Orthologous Groups (COG) and Kyoto Encyclopedia of Genes and Genomes (KEGG) database were used.A total of 4,474 COG items were obtained from Pc, Ac, Pv, and Av groups, and most bacteria were clustered into amino acid transport and metabolism, transcription, carbohydrate transport and metabolism, and cell wall/ membrane/envelope biogenesis (Fig. 5A).The count of different COG items between Ac vs. Pc, Av vs. Pv, Pv vs. Pc, and Av vs. Ac were 229, 47, 143, and 23, respectively.However, no COG was shared by Pv vs. Pc and Av vs. Ac (Fig. 5B).Among the top 15 COG, the relative abundance of bacteria from 4 items (COG2207, COG2814, COG0583, and COG1609) was significantly higher in Pc group than Ac group, while the relative abundance of bacteria from the remaining 11 items showed the opposite trend (Fig. 5C).Moreover, BmNPV infection resulted in a significant decrease in the relative abundance of bacteria clustered to the top 15 COG, both in P50 and A35 silkworms (Fig. 5C).
At KEGG level 3, the relative abundance of bacteria associated with the shigellosis signal pathway in Pc group was significantly higher than that in Ac group, while it was contrary to the bacteria associated with glycosylphosphatidylinositol-anchor biosynthesis (Fig. 6A).After BmNPV infection, the relative abundance of bacteria related to spliceosome signal pathway significantly decreased in A35 strain (Fig. 6B), and the similar case for bacteria related to hematopoietic cell lineage, pancreatic secretion, biosynthesis of type II polyketide backbone and products pathway occurred in P50 strain (Fig. 6C).However, BmNPV invasion caused a remarkable increase in the relative abundance of bacteria related to DDT degradation (Fig. 6C).

Phylogenetic Analysis of Predominant Microflora in Resistant and Susceptible Silkworm
To reveal the evolutionary relationship of gut-predominant bacteria of silkworms, the top 50 OTUs from each sample were chosen to construct phylogenetic trees with the neighbor-joining method.The results showed that the top 50 OTUs from Pc group were clustered into 5 categories (A-E), category A included Bacteroidia, category B included Alphaproteobacteria and Gammaproteobacteria, category C included Bacilli and Clostridia, categories D and E included Fusobacteriia and Actinobacteria, respectively.Category A was divided into subcategories A1 and A2, with A1 including Muribaculaceae and A2 including Rikenellaceae, Porphyromonadaceae, Bacteroidaceae, and Prevotellaceae.Similarly, category B was divided into 2 branches, with B1 enriched with Enterobacteriaceae, Neisseriaceae, Moraxellaceae, Pasteurellaceae, Burkholderiaceae, and Rhodanobacteraceae, and B2 including Caulobacteraceae, Sphingomonadaceae, and Micropepsaceae.Category C was also bifurcated into 2 groups (C1 and C2), with C1 including Lactobacillaceae and Streptococcaceae and C2 including Ruminococcaceae (Fig. 7A).The top 50 OTUs from Ac group differed from those of Pc group; thus, the phylogenetic trees of the 2 groups were also different.The phylogenetic tree of Ac group consisted of categories A, B, C, E, F, and G.The composition of the first 4 categories was similar to that of the Pc group, while categories F and G included Desulfovibrionia and Gemmatimonadetes, respectively.The other detail differences were that the A2 subcategory of Ac tree lacked Porphyromonadaceae while increased with Tannerellaceae, B2 subcategory of Ac tree increased with Devosiaceae and Xanthobacteraceae, and C2 subcategory of Ac tree increased with Lachnospiraceae (Fig. 7B).After BmNPV infection, the top 50 OTUs from Pv group differed a little from those of Pc group (42 out of 50 were the same), which resulted in the similarity of the phylogenetic tree between the 2 groups.The only detailed differences were that the B2 subcategory of Pv tree lacked Micropepsaceae while it increased with Acetobacteraceae (Fig. 7C).Although the top 50 OTUs from Av group differed a lot from those of Ac group (35 out of 50 were the same), the structure of the phylogenetic trees of the 2 groups was similar.The only detailed differences were that the Av tree lacked category G while it increased with category D, and the B2 subcategory of Av tree lacked Devosiaceae and Xanthobacteraceae (Fig. 7D).

Discussion
There are abundant and diverse microbes living in the gut of insects in a symbiotic relationship; some of them have even developed coevolutionary associations with hosts (Dillon and Dillon 2004).These microorganisms have been reported to play beneficial roles to insects, such as helping host digestion and nutrient absorption, strengthening the immune system, resistance to pathogen and toxic substances, ecological balance maintenance, and environmental stress tolerance (Dillon et al. 2005, Engel and Moran 2013, Chen et al. 2020, Sun et al. 2022).The intestinal fluid dilution culture method was widely used to investigate the interactions between intestinal microorganisms and their hosts.Based on API-20E identification system or PCR of bacterial colonies, researchers isolated and identified plenty of culturable bacteria.However, these methods did not work for most predominant gut bacteria because of unculturability or low abundance in the insect intestinal tract.With the development of modern biotechnology, high-throughput sequencing of prokaryotic biomarker 16S rRNA genes is becoming more and more popular, which can identify both culturable and unculturable bacteria.
In the present study, culture-independent 16S rRNA sequencing was performed to investigate gut bacterial communities of BmNPVresistant and susceptible strain silkworm and the response of intestinal microflora of silkworm to BmNPV infection.A large amount of high-quality valid tags and annotated OTUs provided adequate detailed information about the gut microbes of B. mori.Rarefaction analysis showed that all of the richness curves reached plateaus, indicating that the sequencing approach was carried out sufficiently and the bacterial diversity in the silkworm gut could be assessed effectively.Here, 34 bacterial phyla were identified from the frass of healthy and BmNPV-infected P50 and A35 strain silkworms.Likewise, 29 and 32 phyla were identified from the healthy and BmBDV-infected silkworms, respectively (Kumar et al. 2019).However, only 16 and 14 phyla were identified from the healthy and BmCPV-infected silkworms, respectively (Sun et al. 2016).These findings suggested that different virus infections had various influences on the intestinal microorganisms in silkworms.
Herein, the Proteobacteria phylum dominated all of the communities from 4 groups, followed by Bacteroidetes, Firmicutes, and Actinobacteria.Similar situations were reported in the previous studies; for instance, bacteria of the Proteobacteria and Firmicutes phyla were also present in Mythimna separata (He et al. 2013), Firmicutes, Proteobacteria, and Actinobacteria were dominant phyla in healthy and BmCPV-infected silkworm (Sun et al. 2016), Firmicutes, Proteobacteria, Actinobacteria, and Bacteroidetes were dominant phyla in control and BmBDV infected silkworm (Kumar et al. 2019), while Firmicutes, Proteobacteria, Cyanobacteria, Bacteroidetes, and Actinobacteria were mainly phyla in the midgut of H. armigera with and without baculovirus infection (Yuan et al. 2021).These results suggested that the intestinal flora of different Lepidoptera insects share similar characteristics.However, there were still differences in the composition and diversity of bacterial microbiota in response to virus infection, indicating that the changes of intestinal microflora in insects infected with viruses could be affected by both the kind of virus and the races and species of insect.
Virus infection of the silkworm can also lead to changes in gut bacteria at the family level.In the previous study, 77 bacterial families were detected in BmCPV-infected Daizo strain silkworm, fewer compared to the healthy silkworm (94 families) (Sun et al. 2016), while infection with BmBDV for 48 and 144 h resulted in a reduction of 5 and 142 taxa at the family level in the gut microbiota of silkworm, respectively (Kumar et al. 2019).Here, it was found that BmNPV infection led to a decrease of 8 and 16 taxa at the family level in fecal microflora of P50 and A35 strain silkworms, respectively.These results suggested that virus infection could reduce the diversity at the family level in the intestinal microflora of silkworms.
The impact of gut microbiota on host-pathogen interactions are diverse.Microbiota can protect hosts from virus-induced diseases through direct or indirect mechanisms, but it can also promote viral infection (Wilks and Golovkina 2012).In Schistocerca gregaria, species-rich gut bacteria protect the host against pathogen invasion (Dillon et al. 2005).While in Spodoptera exigua, SeMNPV infection resulted in decrease in the transcription level of immune-relative genes and increase in bacterial load in the gut of the larva.As a result, gut microbiota enhanced the virulence, pathogenicity and dispersion of SeMNPV (Jakubowska et al. 2013).In this study, we found that BmNPV infection of P50 silkworm caused significant changes in the relative abundance of bacteria in 17 families, including an increase of 7 families and a decrease of 10 families.While BmNPV infection of A35 silkworm caused significant changes in the relative abundance of bacteria in 11 families, including an increase of 3 families and a decrease of 8 families.On the other hand, only 2 increased families and 4 decreased families were observed in BmNPV-infected Dazio race silkworms (Shi et al. 2021).In terms of BmCPV and BmBDV, viral infection led to a decrease in bacterial diversity in the gut of silkworms (Sun et al. 2016, Kumar et al. 2019).Although these alterations in gut microbiota may be disparate in different insect and viral species, virus infection can affect the balance of intestinal flora and alter the host immunity, leading to changes in host resistance to pathogens.
Since the intestinal microbes are present at the sites where the virus entered the host, they may alter the result of the infection and the host's resistance to the virus.In Aedes aegypti, the endogenous gut microbiota stimulates the host's Toll immune pathway to modulate the dengue virus infection (Xi et al. 2008).Likewise, baculovirus infection caused the upregulation of genes in Toll and IMD pathways and elimination of the gut microbiota in H. armigera, and consequently, intestinal flora minified the expression of prophenoloxidase to facilitate HearNPV infection (Yuan et al. 2021).Here, we found that bacterial diversity in healthy A35 silkworms was higher than that of P50 strain at order, family and genera level.Previous studies stated that the diversity of gut bacteria prevented the host from pathogen infection, and some bacteria could enhance host resistance to pathogens (Dillon et al. 2005, Ramirez et al. 2014).Therefore, we hypothesized that the composition and diversity of intestinal flora of silkworms might be a nonnegligible factor leading to the difference in baculovirus resistance among silkworm varieties.
Feces are frequently used as proxies for gut microbiota because they are naturally collected, noninvasive and can be sampled repeatedly (Tang et al. 2020), especially in studies of humans or other research subjects with poor access to gut contents.Using feces as materials to analyze intestinal flora has been reported in some insect species, including Protaetia brevitarsis (Xuan et al. 2022), Apis mellifera ligustica (Zhang et al. 2021), and 123 species of caterpillars in Lepidoptera (Hammer et al. 2017).However, it has been demonstrated that the fecal and luminalassociated microbiota are 2 distinct microbial niches (Rangel et al. 2015, Ringel et al. 2015).In the previous study, up to 4,594 unique OTUs and a total of 434 bacterial families were detected from gut samples of healthy and BmNPV-infected Jingsong × Haoyue strain silkworms, and BmNPV infection resulted in 448 differential OTUs and 17 differential families (Shi et al. 2021).While in this study, up to 4,703 unique OTUs and a total of 356 bacterial families were identified from feces samples of healthy and BmNPV-infected P50 and A35 strain silkworms, BmNPV infection resulted in 590 and 625 differential OTUs, 21 and 14 differential families in P50 and A35 strain silkworms, respectively.The deviations in the above results may be caused by the difference in silkworm strains or by the difference in feces and gut samples.Therefore, there are some deficiencies in the evaluation of intestinal microbiota using feces in this study that we cannot ignore.And more work needs to be done to investigate the differences in microflora between feces and intestinal contents of these 2 silkworm strains.
Taken together, the differences in composition and diversity of fecal microbiota between BmNPV-resistant and susceptible strain silkworm and the changes of fecal microflora between the 2 strains of the silkworm in response to BmNPV infection were revealed in the present study.BmNPV-resistant strain silkworm possessed higher bacterial diversity than the susceptible strain, and BmNPV infection decreased the diversity of intestinal flora assessed by feces in both strains.This understanding can shed light on the broader impacts of viral infections on host-microbiota interactions.However, more in-depth work needs to be done to elucidate the effects of intestinal bacteria on BmNPV invasion and to develop BmNPV-resistant silkworm strains based on gut microbiota.

Fig. 1 .
Fig. 1.Venn diagrams showing the number of OTU distributed among samples.A) Veen analysis of the number of OTU in each sample, B) Veen analysis of the number of OTU in 4 groups.Pc/Ac: healthy samples of P50/A35 strain, Pv/Av: BmNPV-infected samples of P50/A35 strain.

Fig. 4 .
Fig. 4. BmNPV infection alters the composition of intestinal bacteria in silkworms.A) Heat map showing the changes of bacteria families in the host gut after BmNPV infection in P50 silkworm.B) Heat map showing the changes of bacteria families in the host gut after BmNPV infection in A35 silkworm.Asterisks indicate significant differences between mean values of the healthy and infected silkworm in each strain, *P < 0.05; **P < 0.01 (Wilcoxon test).

Fig. 5 .
Fig. 5. COG functional enrichment profiles of intestinal bacteria.A) COG functional enrichment analysis of gut bacteria of 4 groups.B) Veen analysis of the number of different COG items between Ac vs. Pc, Av vs. Pv, Pv vs. Pc, and Av vs. Ac.C) Heat map showing the relative abundance of bacteria from the top 15 COG items in 4 groups.

Fig. 6 .
Fig. 6.KEGG functional enrichment profiles of intestinal bacteria.A) Heat map showing the differences of the relative abundance of bacteria from healthy silkworms at KEGG level 3. B) Heat map showing the changes in the relative abundance of bacteria post BmNPV infection in A35 silkworm at KEGG level 3. C) Heat map showing the changes in the relative abundance of bacteria post BmNPV infection in P50 silkworm at KEGG level 3. Asterisks indicate significant differences between the mean values of the 2 groups, *P < 0.05; **P < 0.01 (Wilcoxon test).

Table 2 .
Average relative abundances of the 10 most abundant phyla/family within Bombyx mori gut communities Fig. 2. Relative abundance of the top 30 families in all samples.