Metagenomic 16S rRNA amplicon and shotgun sequencing in investigation of granulomatous lesions in layer chickens: a case report

Abstract

Background

Granuloma lesions in poultry are a frequent pathological finding, representing a chronic inflammatory response to persistent infectious agents, most commonly bacteria or fungi. Accurate differentiation of the underlying cause requires additional diagnostic tests, such as acid-fast staining, fungal culture, or bacterial isolation and genetic identification. However, the pathogens often remain undetermined when conventional detection methods fail. The present case aimed to investigated granulomatous disease in multiple organs of layer chickens by applying metagenomic 16S rRNA amplicon and metagenomic shotgun sequencing.

Case presentation

A total of 35 deaths occurred in a flock of layer chickens, accompanied by a decrease in daily feed intake from 144 g to 104 g. Six carcasses from 38-weeks-old layer chickens were submitted for disease diagnosis. During necropsy, granuloma/neoplastic lesions were observed in the liver, ovary, proventriculus, pancreas and kidney. Histopathological examination revealed compartmentalized infiltration of lymphocytes and multinucleated giant cells in liver, ovary, proventriculus, pancreas and renal parenchyma. While Escherichia coli was isolated from the oviduct, no viral agents (IBV, CIAV, MDV, ALV, REV, or HEV) were detected by RT-PCR. Notably, some of the initial metagenomic shotgun results were recognized as different taxa due to the misclassification of certain reads. Following a re-analysis of these shotgun reads against the nucleotide database, both metagenomic 16S rRNA amplicon and metagenomic shotgun sequencing identified E. coli as the dominant species in liver samples.

Conclusion

The present case indicates that E. coli may contribute to granuloma lesions in layer chickens. While 16S rRNA amplicon sequencing provided more reliable bacterial identification for diagnostic purposes, metagenomic shotgun sequencing offers complementary insight by detecting broader microbial communities. Therefore, integrating both approaches alongside conventional diagnosis may improve the accuracy of diagnosis for granulomatous diseases in poultry.

Supplementary Information

The online version contains supplementary material available at 10.1186/s42523-026-00565-6.

Background

Granulomatous and neoplastic diseases in poultry are of considerable economic and health importance, as they reduce productivity and may spread rapidly within flocks [1]. Early diagnosis and prompt management are therefore critical to minimizing losses and controlling disease progression in commercial production systems. Traditionally, diagnosis relies on clinical and necropsy findings supported by histopathology and isolation/detection of causative agents [2]. However, these approaches may be limited by carcass conditions, prior antimicrobial administration, co-infections. or the opportunistic nature of some pathogens, making a definitive diagnosis difficult in certain cases [3].

While major avian granulomatous/neoplastic diseases have been linked to oncogenic viruses, growing evidence suggests that other factors contribute to tumor-like lesions through granulomatous, chronic inflammation and immune modulation [4, 5]. Metagenomic sequencing has emerged as a valuable tool for culture-independent diagnosis of causative agent. Specifically, metagenomic 16S rRNA amplicon sequencing analyzes the compositional information of bacterial species. In parallel, metagenomic shotgun sequencing provides insight into entire microbiota while identifying genetic feature such as virulence and antimicrobial resistance [6].

In the present case, the granuloma/neoplastic lesions were found in the liver, ovary, proventriculus, pancreas and kidneys of commercial layer chickens. A definitive diagnosis could not be reached using conventional methods. However, by applying metagenomic 16S rRNA amplicon and metagenomic shotgun sequencing, E. coli was identified as the dominant species within the granulomas. This highlights the potential contribution of E. coli to avian granulomatous disease. It also underscores the importance of integrating metagenomics with conventional pathology for more accurate diagnosis in cases of unknown etiology.

Case presentation

Samples

In February 2025, 35 deaths occurred in a single building on a layer chicken farm located in Pocheon-si, Gyeonggi-do. Although the laying rate was approximately 96%, daily feed intake had decreased from 114 g to 104 g. Six carcasses from 38-weeks-old layer chickens were submitted to the Animal and Plant Quarantine Agency (APQA, South Korea) for disease diagnosis. Postmortem examination, histopathological analysis, and bacterial and viral identification were performed in accordance with the Avian Disease Diagnosis Manual of the APQA (https://ebook.qia.go.kr/20250325_171417/) and standard protocols of Poultry Disease [2]. Relevant information (Additional file 1), including farm location, flock size, clinical symptoms, and diagnostic results, was recorded in the Korea Animal Health Integrated System (KAHIS; https://www.kahis.go.kr/).

During necropsy, organs showing gross lesions were collected and fixed in 10% neutral buffered formalin for more than 12 hours, followed by paraffin embedding. Tissue sections were cut at a thickness of 5 μm, and stained with hematoxylin and eosin (H&E) for microscopic examination.

The liver, kidney, and oviduct samples showing signs of bacterial infection were aseptically collected using sterile cotton swabs and cultured on 5% sheep blood agar (Seong-nam Co., South Korea), MacConkey agar, and LB broth (Additional file 1). Cultures were incubated under both aerobic and anaerobic condition at 37 °C. Bacterial colonies were isolated and identified using the VITEK II system (bioMérieux, Marcy I’Etoile, France) according to the manufacturer’s instructions.

For viral agent detection, the trachea, thymus, liver, spleen, kidney, cecal tonsil, and granuloma/neoplastic tissues from all six carcasses were collected and pooled by organ. Total nucleic acids were extracted from homogenized tissues using the Maelstrom 4000 DNA/RNA automated extraction system (TANBead, Taiwan Advanced Nanotech Inc., Taoyuan City, Taiwan) according to the manufacturer’s instructions. Reverse transcription PCR (RT-PCR) was used to detect viral agents, based on previous studies (Additional file 1). Infectious Bronchitis Virus (IBV) was tested in the trachea, liver, kidney, and cecal tonsil; Chicken Infectious Anemia Virus (CIAV), Reticuloendotheliosis Virus (REV), Marek’s Disease Virus (MDV) and Avian Leukosis Virus (ALV) was tested in the trachea, thymus, liver, kidney, cecal tonsil, and granuloma/neoplastic tissues; and Hepatitis E virus (HEV) was tested in the liver [7–9]. For RT-PCR targeting IBV, ALV, and HEV, field samples previously confirmed positive through routine diagnostic procedures at APQA manual were used as positive controls. These included tracheal tissue for IBV (KAHIS sample ID 23AQ002), liver tissues for ALV (23AD027), and HEV (24AD013) which were verified by pathological examination and sequencing before use. CIAV, REV, and MDV were confirmed using BlackCheck CIAV/MDV/REV Multi Detection Kit (Ventech Science, South Korea). Although IBV-positive RT-PCR products are typically sequenced following established protocol [8], no IBV-positive samples were identified in the present case.

Metagenomic 16S rRNA amplicon and metagenomic shotgun sequencing of the six liver samples were conducted by a commercial service provider (Sanigen Co., South Korea), following protocols described in previous studies [10, 11]. Briefly, DNA was extracted using Qiagen DNeasy Blood & Tissue Kit (Qiagen, MD, USA) and sequencing libraries were prepared by using TruSeq Nano DNA library Prep kit (Illumina, CA USA). The 16S rRNA amplicon sequencing was conducted to amplify the V3–V4 region of the bacterial 16S rRNA genes using Forward (5’-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG-3’) and reverse (5’-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC-3’) primers. The PrimesSTAR HS DNA polymerase (Takara, Japan) was used in PCR amplification. The PCR product was purified using Agencourt AMPure XP beads (Beckman Coulter, USA). Following the use of an Illumina Nextera XT index kit and quantification of the constructed library using a Bioanalyzer 2100 (Agilent, USA) and Qubit4.0 (Thermo Fisher Scientific, USA), deep sequencing was conducted via Illumina MiSeq platform (Illumina) with a paired-end (2 ×300 bp) sequencing mode. Taxonomic classification was performed using the SILVA reference database, and result visualization was conducted using QIIME2 v.2022.2. For the metagenomic shotgun sequencing, library fragment size was determined by the Tapestation4200 (Agilent Technologies, CA, USA) and quantity determined via KAPA Library Quantifcation Kit (KAPA Biosystems). Libraries with insert size of 400 ~ 600 bp were constructed for each individual and sequenced using the Nextseq2000 (Illumina, CA, USA). Taxonomic classification was performed using Kraken2 v2.1.2 and species abundance was estimated by Bracken v2.55 for shotgun reads. The database was constructed using the reference sequences from the National Center for Biotechnology Information (NCBI) RefSeq database. Filtered reads were assembled by SPAdes v3.15.2 with ‘−meta’ option. To ensure assembly quality and relevance to downstream analysis, contigs with a depth less than 5X or length less than 1,000 bp were removed.

The obtained sequencing data were submitted to the NCBI with the BioSample accession number SAMN53296137 (shotgun), and SAMN53307544 (16S rRNA).

Diagnostic assessment

Necropsy findings revealed that the livers were fragile (n = 2), and exhibited features of fatty (n = 1) and green liver (n = 1). Splenomegaly (n = 2) and nephromegaly (n = 2) were also observed. One layer chicken showed intraperitoneal hemorrhage, while others presented with ascites (n = 2), egg yolk peritonitis (n = 3), and salpingitis accompanied by intraluminal abscess (n = 1). Granuloma/neoplastic lesions were noted in several layer chickens (Fig. 1.A to C), including yellow nodules in the liver (n = 1), as well as tumors in the ovary (n = 3), proventriculus (n = 1), pancreas (n = 1) and kidney (n = 1).

Fig. 1.

The gross findings (A to C) and histopathological examination (D to I) of layer chickens. The granuloma/neoplastic lesions were observed in liver (A), ovary and kidney (B), and pancreas (C). The liver exhibited hepatocellular caseous necrosis surrounded by granulomatous lymphocyte infiltration (D), and syncytial cell formation (E). The caseous necrosis with granulomatous lymphocyte infiltration, and connective tissue prolifertation were observed in the ovary (F), proventriculus (G), pancrease (H) and renal parenchyma (I). The white arrowhead indicate granuloma/neoplastic lesions, asterisk denote areas of caseous necrosis, and black arrow indicate syncytial cells. Scale bars are shown in the bottom right (D to I)

Histopathological examination revealed congestion and lymphocyte infiltration in the tracheal submucosa. The livers exhibited perivascular inflammatory cells infiltration, compartmentalized granulomatous lymphocyte infiltration, connective tissue proliferation, syncytial cell formation, hepatocellular caseous necrosis, and fibrotic deposition (Fig. 1.D and E). In the spleen, fibrin deposition and reticulocyte proliferation were evident. Granulomatous lymphocyte infiltration, caseous necrosis and connective tissue proliferation were also observed in the ovary, proventriculus, pancreas and renal parenchyma (Fig. 1.F to I).

Pathogen detection

Bacterial isolation was attempted from four livers, one kidney, and one oviduct samples. No colonies were observed from the livers and kidney under either aerobic or anaerobic culture conditions. However, aerobic culture of the oviduct yielded Escherichia coli, which was confirmed through biochemical identification.

RT-PCR-based viral detection for IBV, CIAV, REV, MDV, ALV, and HEV returned negative results for all tested organ tissues.

Metagenomic results

To further investigate the etiology of the hepatic granuloma/neoplastic lesions, metagenomic shotgun sequencing was conducted using pooled liver samples for six layer chickens (Fig. 2.A). A total 88,632,432 reads were obtained, with 30,852,113 reads retained after quality trimming. Of these, 80.894% were bacteria, 15.771% eukaryota, 2.434% virus, and 0.901% categorized as others, excluding the host reads (Additional file 2). Homology-based classification (BLAST) of the bacterial reads identified a total of 251,005 reads, comprising 30 phyla, 63 classes, 159 orders, 372 families, and 1,125 genera. The most abundant bacterial species were Paraglaciecola sp. L1A13 (28.794%), Staphylococcus aureus (12.626%), and E. coli (10.766%). Other species were detected in minor proportions (<2.5%). From the 48,936 eukaryotic reads, a total of 77 reads matched fungal species, spanning three phyla, 11 classes, 15 orders, 27 families, and 48 genera. The dominant fungal species identified were Fusarium verticillioides and F. venenatum. Among 7,551 viral reads, 36 viral species were classified across 9 phyla, 12 classes, 13 orders, 21 families, and 25 genera. Avian endogenous retrovirus EAV-HP was the most abundant (41.110%). Initial microbiome composition was reported as follows: Paraglaciecola sp. L1A13 (23.032%), S. aureus (10.099%), E. coli (8.611%), F. verticillioides (1.627%), and Avian endogenous retrovirus EAV-HP (0.993%).

Fig. 2.

Taxonomic classification of total microbiome by megagenomic shotgun sequencing (A), and bacterial species by 16S rRNA amplicon sequencing (B) from liver samples of layer chickens. The results were represented by circular distribution charts and bar graphs. The indicated numbers denote the relative abundance (%) of each species. The underlined yellow mark indicated taxa that were re-analyzing the shotgun reads by homology-based classification (BLAST) using nucleotide database (ntDB)

However, upon re-analysis of five most abundant taxa using the nucleotide database, some reads were found to align with host chicken (Gallus gallus) genomes or with other microbiome species, indicating possible misclassification. Consequently, these were designated as ‘Unclassified host and others reads.’ After this adjustment, the re-calculated microbial composition showed E. coli from 8.611% to 8.242%, and Avian endogenous retrovirus EAV-HP from 0.993% to 0.229. Notably, S. aureus, Paraglaciecola sp. L1A13 and F. verticillioides showed almost no matches, with most reads re-classified as host-derived or belonging to other microbial taxa (Additional file 3).

A total 415,650 reads were obtained via metagenomic 16S rRNA sequencing using the same pooled liver samples, and 298,155 reads were survived after filtering process (Fig. 2.B). BLAST of bacterial reads classified with 17 phyla, 27 classes, 63 orders, 96 families, and 154 genera. The most abundant taxa were Escherichia-Shigella sp. (59.250%), Lactobacillus sp. (10.674%), and L. salivarius (7.243%) among total 217 taxa. Reads assigned to Staphylococcus sp. represented only 0.117% of the total, indicating a low relative abundance in liver microbiota (Additional file 4).

Discussion and conclusions

At the APQA, poultry disease investigations follow a standardized diagnosis manual consisting of: ①gross necropsy, ②histopathological examination, and ③isolation/identification of causative agents. While this approach provides reliable results in most cases, causative agents were sometimes not recovered due to postmortem autolysis, prior antimicrobial treatment, or differences of host immunity and pathological responses [3]. In such cases, conventional bacteriological culture or PCR detection may be insufficient. Accordingly, metagenomic analysis, which can identify the entire microbiota present in subjects, is increasingly being applied to the diagnosis of animal disease [12, 13].

Tumoral diseases in chickens are generally divided into two categories: infectious cases, most often induced by oncogenic viruses [1, 14], and sporadic non-infectious cases of unknown etiology [15]. Viral tumors are of major economic concern because they are contagious and highly pathogenic. Marek’s disease, avian lymphoid leukosis, and reticuloendotheliosis are the three principal oncogenic viral disease in poultry. All are characterized by lymphoid proliferation and immunosuppression. MD causes lymphoid infiltration and tumors in multiple organs, with subclinical effects on growth and egg productivity; AL, especially subgroup J in chicken, causes myeloid neoplasms and reduced egg production; and RE produces chronic lymphoid neoplasms, immunosuppression, and feathering abnormalities [1, 14]. In South Korea, routine vaccination against MD in commercial chicks has limited its prevalence [15], with sporadic cases shown mainly in unvaccinated native breeds or spent layer hens. Preventive strategies against AL and RE rely primarily on eradication and genetic control, as vaccines are unavailable. In the present case, however, PCR analysis excluded all three oncogenic viruses, supporting the likelihood of a non-viral etiology.

Tumorigenesis may also result from both extrinsic factors (chemicals, microbial infections, and environmental factors) and intrinsic factors (genetics, hormones, age, and immune status) [5]. In recent years, bacterial contributions to tumoral processes have gained attention. Bacteria can alter the host microenvironment, promote chronic inflammation, induce DNA damage, and disrupt immune regulation across various organs, not limited to the digestive systems. Bacterial colonization of tumors may directly modulate tumor cell behavior, facilitate immune evasion, and contribute to tumor dynamics via bacterial toxins as paracrine effects [16].

E. coli, an opportunistic bacterium, has been associated with chronic inflammatory states and has been detected within neoplastic tissues of host [17, 18]. In the present case, E. coli were isolated from the oviduct, but no bacterial colonies were isolated from the livers. This discrepancy highlights the limitations of conventional bacteriology, especially in cases involving prior antimicrobial exposure or fastidious organisms. Moreover, as E. coli transmission in the vaginal pathway is primarily progresses through the cloaca [19], it may not be directly associated with liver lesions, and thus metagenomic analysis was applied in present case to investigate the liver granuloma. The results revealed that both metagenomic 16S rRNA amplicon and metagenomic shotgun sequencing identified E. coli as a dominant species in liver samples, strongly suggesting its involvement.

While E. coli is ubiquitous in poultry farms, its pathogenic role may be expressed under immunosuppressive conditions influenced by other pathogens and environmental factors. This bacterium can enter via fecal-oral, respiratory, and vaginal pathway. Because poultry lack a diaphragm, E. coli can reach the air sacs through the trachea, leading to airsacculitis and subsequent spread to adjacent organs, such as the liver [19]. Moreover, a possible route of infection is that airsacculitis may subsequently allow translocation to the kidney and other organs via hematogenous dissemination [20].

Coli-granulomatous lesions caused by E. coli are characterized by granuloma formation in various visceral organs, particularly the liver, of affected poultry. Although E. coli-induced granuloma lesions occur infrequently at the individual level, severe outbreaks within flocks can result in mortality rates approaching 75% [21], suggests a potential for intra-flock transmission. In poultry, granuloma lesions develop when the immune system fails to eliminate infectious agents with affected organs. Histologically, these lesions are characterized by coagulative and caseous necrosis, surrounded by epithelioid cells, multinucleated giant cells, lymphocytes, and macrophages. However, they can be difficult to distinguish morphologically from leukosis tumors [21–23]. Previous studies have reproduced similar pathological lesions through the experimental inoculation of E. coli isolated from granuloma lesions into chickens [23]. Furthermore, pyogranulomatous inflammation caused by E. coli has been directly associated with E. coli-induced granulomas in the lungs and air sacs in layer chickens [24]. In addition, several cases of E. coli-associated granulomatous disease have been reported in multiple chicken breeds. These were characterized by multifocal or nodular hepatic lesions confirmed histologically and accompanied by E. coli isolation from the affected organs [4, 19, 25–27]. The liver plays a critical role in nutrient metabolism, detoxification, and immune defense. In poultry it is a major target of both infectious and non-infectious disease, and hepatic lesions primarily reflect systemic pathology [28]. In the present case, therefore, metagenomic analysis was performed only on the liver, as it is key organ for poultry disease diagnosis. However, other granuloma/neoplastic lesions observed in the ovary, proventriculus, and kidney were not analyzed, which represents a limitation of this case. Further metagenomic analysis of these organs would provide a more comprehensive understanding of the disease process. Nevertheless, rapid diagnosis and management are crucial for the poultry industry, and it is important to quickly identify the pathogens of chicken disease through targeted analysis of key organs. Therefore, E. coli is regarded as the most likely pathogens contributing to the granuloma of liver, and potentially to those in the ovary, proventriculus, and kidney as well.

The dual application of metagenomic 16S rRNA amplicon and metagenomic shotgun sequencing provided valuable, but not identical, insights. Shotgun metagenomics, while capable of simultaneously detecting bacteria, viruses, and fungi, and offering genetic characterization such as virulence and antimicrobial resistance factors, was hampered by misclassification as different or phantom taxa in Kraken. This was particularly evident at large sequencing depths, suggesting that 16S rRNA amplicon sequencing remains a useful tool [29]. In the present case, misclassification occurred as tens of millions of reads were generated. The S. aureus reads, which initially appeared at high abundance, were re-classified as host-derived or other microbial sequences, except for a single read, when re-analyzed and compared via nucleotide database. Similarly, in the case of Paraglaciecola sp. L1A13 and F. verticillioides, no reads matched these species upon re-analysis. The incomplete removal of host-derived sequences during the filtering process likely contribute to this misclassification in the metagenomic shotgun analysis. Although more than 99% of host reads can typically be removed, still some of proportion of residual host sequences often remains due to the imperfect performance of host decontamination tools. When host genomes (e.g., Gallus gallus) are not included in the reference database used for taxonomic classification, these residual reads may be erroneously assigned to the closest available reference organisms. In the present case, such reads were misclassified reflecting sequence similarity rather than true biological presence. In the case of E. coli, the relative abundance was reduced after re-analysis, but a substantial number of reads were still detected in the shotgun results, which was consistent with the 16S rRNA amplicon sequencing and bacterial isolation from oviducts. These findings suggest that while shotgun sequencing is powerful for broad-spectrum pathogen detection, 16S rRNA amplicon sequencing currently provides greater accuracy for bacterial identification in diagnostic contexts. In addition, the BLAST contrast verification process using nucleotide database for shotgun reads is still manual, and re-analysis of tens of millions of reads will required time and budget-consuming processes. Therefore, the integration of both methods is recommended: 16S rRNA amplicon sequencing for reliable species-level bacterial classification, and shotgun sequencing as an auxiliary approach for detecting co-infections with viruses or fungi, and identifying genetic features such as virulence factors or resistance genes.

In conclusion, the present case highlights the possible involvement of E. coli in granuloma lesions in liver of layer chickens, and also suggests a potential association with similar pathological changes in the ovary, proventriculus and kidney. The comparison of sequencing platforms underscores both the strengths and weaknesses of current approaches. Metagenomic 16S rRNA amplicon sequencing remains more reliable for diagnostic purposes regarding bacterial agents, whereas metagenomic shotgun sequencing showed limited suitability for routine diagnosis due to misclassification of taxa. This case has inherent limitations. Metagenomic analysis was performed only on liver samples, and findings in other affected organs were inferred rather than directly confirmed. Furthermore, because bacterial or viral agents could not be cultivated or isolated from the liver specifically, the reproducibility and virulence of pathogens in layer chickens remain restricted. Despite these constraints, metagenomic shotgun sequencing still provides valuable complementary information by enabling simultaneous detection of broader microbial communities, including bacteria, viruses, and fungi. Moving forward, the combined application of metagenomic 16S rRNA amplicon and metagenomic shotgun sequencing, and conventional disease diagnosis will enhance diagnostic accuracy and contribute to a deeper understanding of poultry disease, particularly granulomatous disease that may arise from diverse etiologies such as bacterial, viral, fungal, or other extrinsic and intrinsic factors.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

LB broth

Luria-Bertani broth

IB(V)

Infectious Bronchitis (virus)

CIA(V)

Chicken Infectious Anemia (virus)

MD(V)

Marek’s disease (virus)

AL(V)

Avian leukosis (virus)

HEV

Avian hepatitis E virus

E. coli

Escherichia coli

S. aureus

Staphylococcus aureus

F. verticillioides

Fusarium verticillioides

F. venenatum

Fusarium venenatum

L. salivarius

Lactobacillus salivarius

Author contributions

Conceptualization: Yun CS, Kim J, Moon JS; Data curation: Yun CS, Kim J; Formal analysis: Yun CS, Kim J, Kwon H; Methodology: Yun CS, Kim J; Software: Yun CS, Kim J; Investigation: Yun CS, Kim J, Kwon H; Writing-original draft: Yun CS; Writing-review & editing: Yun CS, Kim J, Kwon H, Her M, Moon JS.

Funding

This research was supported by a grant from the Animal and Plant Quarantine Agency, Ministry of Agriculture, Food and Rural Affairs of the Republic of Korea (N-1543084-2025-29-01). The funding body had no role in the design of the study, collection, analysis, interpretation of data, and in writing the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The present study does not require animal ethic approval because the layer’s bodies were submitted to an Animal and Plant Quarantine Agency for diagnosis by veterinarians and animal owners with their consents.

Consent for publication

Not applicable

Competing interests

The authors declare no competing interests.