Abstract Background The exact cause of recurrent aphthous stomatitis is still unknown, making it a challenge to develop effective treatments. This study employs computational biology to investigate the molecular basis of recurrent aphthous stomatitis, aiming to identify the nature of the stimuli triggering these ulcers and the type of cell death involved. Methods To understand the molecular underpinnings of recurrent aphthous stomatitis, we used the Génie tool for gene identification, targeting those associated with cell death in recurrent aphthous stomatitis. The ToppGene Suite was employed for functional enrichment analysis. We also used Reactome and InteractiVenn for protein integration and prioritization against a PANoptosis gene list, enabling the construction of a protein-protein interaction network to pinpoint key proteins in recurrent aphthous stomatitis pathogenesis. Results The study’s computational approach identified 1,375 protein-coding genes linked to recurrent aphthous stomatitis. Critical among these were proteins responsive to bacterial stimuli, especially high mobility group protein B1 (HMGB1), toll-like receptor 2 (TLR2), and toll-like receptor 4 (TLR4). The enrichment analysis suggested an external biotic factor, likely bacterial, as a triggering agent in recurrent aphthous stomatitis. The protein interaction network highlighted the roles of tumor necrosis factor (TNF), NF-kappa-B essential modulator (IKBKG), and tumor necrosis factor receptor superfamily member 1A (TNFRSF1A), indicating an immunogenic cell death mechanism, potentially PANoptosis, in recurrent aphthous stomatitis. Conclusion The findings propose that bacterial stimuli could trigger recurrent aphthous stomatitis through a PANoptosis-related cell death pathway. This new understanding of recurrent aphthous stomatitis pathogenesis underscores the significance of oral microbiota in the condition. Future experimental validation and therapeutic strategy development based on these findings are necessary. Supplementary Information The online version contains supplementary material available at 10.1186/s12903-024-04917-z. Keywords: Aphthous stomatitis, Computational biology, Bacterial infections, Immunogenic cell death, Cell death Background Recurrent aphthous stomatitis, also known as canker sores or recurrent aphthae, represent the predominant type of lesion encountered by healthcare professionals dealing with oral ulcerative conditions, impacting an estimated 25% of individuals worldwide [[32]1, [33]2]. Characterized by numerous, recurrent round or oval ulcers with distinct margins, red haloes, and yellow or gray bases [[34]3], this condition significantly impacts the quality of life of affected individuals due to pain [[35]4]. Recurrent aphthous stomatitis might be associated with various local and systemic factors, such as infections (bacterial or viral), trauma, stress, nutritional deficiencies, systemic diseases, immunological disorders, or hereditary factors, although a definitive causal relationship has not been established [[36]5, [37]6]. A key clinical hurdle in preempting the onset and recurrence of oral ulcers lies in identifying the specific stimulus or agent responsible for the demise of oral mucosal keratinocytes. The enigmatic trigger behind recurrent aphthous stomatitis classifies it as an idiopathic condition. From a research perspective, the lack of a validated experimental model for this disease [[38]1] worsened the challenge, likely leading to the prevalent use of palliative over curative treatments. Computational biology is the use of computational techniques, algorithms, and models to analyze and elucidate biological data, simulate biological processes, and make predictions about biological systems. The domain of computational biology has proven its importance in unraveling complex biological enigmas by employing mathematical models and algorithms [[39]7]. Currently, the challenge in formulating new hypotheses concerning the stimuli or agents primarily implicated in recurrent aphthous stomatitis arises from the vast volume of available biomedical data. One of the key challenges for computational biology methods in the study of human diseases is to distill this information into a comprehensible format [[40]8]. The diversity and multiple causative factors of diseases, the challenge of identifying specific genes or proteins in certain locations, and the expenses associated with experimental research underscore the need for developing various computational biology methods to identify genes and proteins linked to a disease [[41]9]. Numerous studies have documented the effective application of these methods in prioritizing genes associated with diseases and suggesting potential therapeutic targets [[42]10–[43]12]. However, these techniques have not been utilized in researching the etiology of recurrent aphthous stomatitis. To enhance our comprehension of the molecular underpinnings of recurrent aphthous stomatitis and to hypothesize about the nature of the triggering stimulus for the lesions, we employ data mining techniques to gather an extensive list of protein-coding genes linked to this condition. Subsequently, we analyze the biological properties of these proteins and prioritize a group of them using Venn diagrams and protein-protein interaction networks. Focusing on the important proteins within the interaction networks (those with the greatest number of connections), we delineate the pathways of oral keratinocyte death, providing a basis for a theoretical model for ulcer development. Methods General design Utilizing Génie for literature-based gene identification, we extracted genes associated with cell death in aphthous stomatitis, aiming to understand the stimulus triggering lesions. Functional enrichment was performed using ToppGene Suite. Protein integration and prioritization through Reactome and InteractiVenn, compared with a PANoptosis gene list, helped construct a network to identify key proteins (Fig. [44]1). Fig. 1. [45]Fig. 1 [46]Open in a new tab Comprehensive computational pipeline to elucidate stimulus nature and cell death mechanism in recurrent aphthous stomatitis. This figure outlines our step-by-step approach to studying recurrent aphthous stomatitis, starting with the Génie web server that ranks protein-coding genes associated with cell death in this condition. We then used ToppGene to characterize these genes based on gene ontology, which helped identify potential stimuli that trigger ulcer formation. Reactome further allowed us to examine specific types of cell death, improving our understanding of how these stimuli affect oral mucosa keratinocytes. Finally, using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database and the Integrated Value of Influence (IVI) application, we built a protein interaction network that revealed key proteins involved in the disease’s pathology. Alongside these tools, PubMed identifiers (PMIDs) provide unique reference numbers for tracking scientific articles, while the ToppGene Suite offers extensive capabilities for gene list analysis and prioritization. Voronoi diagrams show us how biological patterns can emerge, and Reactome aids in the visualization and analysis of complex biological pathways. The InteractiVenn tool allows for straightforward analysis of element lists using Venn diagrams, enhancing our grasp of data intersections. Protein-protein interaction maps give insight into how proteins interact within a cell, essential for understanding cellular functions and the impact of disruptions on disease. The STRING database compiles a wide array of protein interactions from experimental and predictive data, and the IVI app identifies the most influential nodes in our network, highlighting critical targets for potential therapy. This array of sophisticated tools and databases enables a detailed exploration of the molecular dynamics involved in recurrent aphthous stomatitis, offering insights into its underlying mechanisms and potential treatments Literature-based gene identification We used the Gene Network Inference with Ensemble of Trees (Génie) online tool [[47]13] to prioritize genes relevant to cell death in aphthous stomatitis from a vast array of scientific literature. Understanding the mechanisms and types of cell death [[48]14, [49]15] is crucial for tracing the potential triggers, especially in recurrent aphthous stomatitis where the initiating factors remain unidentified. Génie utilizes natural language processing to examine connections between genes and biomedical topics across MEDLINE/PubMed abstracts, focusing specifically on cell death related to this condition. Our search included a targeted selection of three pertinent abstracts (PubMed identifiers PMIDs 3858774 [[50]16], 25861801 [[51]17], and 37170213 [[52]18]). We set stringent criteria with a cut-off p-value of 0.01 for abstracts and a false positive discovery rate (FDR) of 0.01 for genes, aiming to ensure that only significant coding genes were identified. Fisher’s exact test was used to analyze gene-topic relationships. Functional enrichment Following the gene identification, functional enrichment analysis was conducted based on the gene ranking generated by Génie. This analysis assesses the frequency of gene appearances across established biological process categories [[53]19]. Using the ToppGene Suite (version 2023-05-03, 20,649 genes in the biological process category) set to default settings, we conducted an enrichment analysis. This process helped us select coding genes that are highly overrepresented, enhancing our understanding of the pathways and broader processes affected by these proteins. Our primary aim was to identify proteins that might elucidate the nature of the stimuli potentially causing ulcer formation. Protein integration and prioritization strategy To explore the pathological significance of these proteins, we utilized the Reactome Knowledgebase [[54]20], which provides detailed insights into a wide range of human biological processes, including those related to genetic and acquired diseases. For protein prioritization, we used InteractiVenn [[55]21] to compare the Génie rankings with an independently established PANoptosis gene list by Song et al. [[56]22]. , which includes genes involved in various cell death pathways such as pyroptosis, apoptosis, and necroptosis. By matching proteins, we constructed a protein-protein interaction network using the STRING database [[57]23], aiming to determine if the prioritized proteins form a biologically cohesive community. To identify the most influential nodes in this network, we employed the Integrated Value of Influence (IVI) shiny application [[58]23], which helps pinpoint key proteins based on their interactive significance. Results Integrated bioinformatics reveals involvement of an external biotic stimulus in cell death of recurrent aphthous stomatitis Génie generated a statistically significant ranking of 1,375 protein-coding genes from a dataset comprising 20,396 genes and 1,183,931 gene-abstract linkages, connected to 551,555 distinct PMIDs (Supplementary dataset file [59]1). Notably, high mobility group protein B1 (HMGB1), toll-like receptor 2 (TLR2), and toll-like receptor 4 (TLR4) emerged as the top genes. In the first positions, proteins that participate in response against bacteria stand out (Table [60]1). Table 1. Génie-generated ranking (top five positions) Rank Gene symbol Protein name (Uniprot ID) Description* 1 HMGB1 High mobility group protein B1 ([61]P09429) Promotes host inflammatory response to sterile and infectious signals and is involved in the coordination and integration of innate and adaptive immune responses. 2 TLR2 Toll-like receptor 2 ([62]O60603) This protein is a cell-surface protein that can form heterodimers with other TLR family members to recognize conserved molecules derived from microorganisms known as pathogen-associated molecular patterns (PAMPs). Activation of TLRs by PAMPs leads to an up-regulation of signaling pathways to modulate the host’s inflammatory response. This gene is also thought to promote apoptosis in response to bacterial lipoproteins. Cooperates with LY96 to mediate the innate immune response to bacterial lipoproteins and other microbial cell wall components. 3 TLR4 Toll-like receptor 4 ([63]O00206) 4 LY96 Lymphocyte antigen 96 ([64]Q9Y6Y9) Cooperates with TLR4 in the innate immune response to bacterial lipopolysaccharide (LPS), and with TLR2 in the response to cell wall components from Gram-positive and Gram-negative bacteria. 5 CASP12 Inactive caspase-12 ([65]Q6UXS9) Caspases (cysteinyl aspartate proteases) are involved in the signaling pathways of apoptosis, necrosis and inflammation. These enzymes can be divided into initiators and effectors. May function as a negative regulator of inflammatory responses and innate immunity. May reduce cytokine release in response to bacterial lipopolysaccharide during infection. [66]Open in a new tab UniProt is the world’s leading high-quality, comprehensive and freely accessible resource of protein sequence and functional information. *We obtained descriptions from The GeneCards human gene database ([67]https://www.genecards.org/). To understand the context in which the protein collective might participate, we performed an enrichment analysis using the ToppGene Suite. Using this methodology, we categorized gene products, or proteins, into biological processes they are involved in. Table [68]2 presents the top five pertinent processes. Notably, the biological processes related to response to biotic stimuli, response to other organisms, and response to external biotic stimuli are particularly noteworthy, aligning with the established association of noxa in recurrent aphthous stomatitis. Supplementary dataset file [69]2 has the complete categorization. Table 2. Top five biological processes associated with recurrent aphthous stomatitis. Regulation of cell death and response to a biotic stimulus stand out Biological process Description* q-value FDR B&H** Proteins count Regulation of programmed cell death Any process that modulates the frequency, rate or extent of programmed cell death, cell death resulting from activation of endogenous cellular processes. 2.11E-248 583 Regulation of apoptotic process Any process that modulates the occurrence or rate of cell death by apoptotic process. 3.69E-237 563 Response to biotic stimulus Any process that results in a change in state or activity of a cell or an organism (in terms of movement, secretion, enzyme production, gene expression, etc.) as a result of a biotic stimulus, a stimulus caused or produced by a living organism. 1.03E-198 529 Response to other organism Any process that results in a change in state or activity of a cell or an organism (in terms of movement, secretion, enzyme production, gene expression, etc.) as a result of a stimulus from another living organism. 1.10E-192 516 Response to external biotic stimulus Any process that results in a change in state or activity of a cell or an organism (in terms of movement, secretion, enzyme production, gene expression, etc.) as a result of an external biotic stimulus, an external stimulus caused by, or produced by living things. 2.49E-192 516 [70]Open in a new tab *We obtained biological definitions from QuickGo ([71]https://www.ebi.ac.uk/QuickGO/).**q-value FDR (False Discovery Rate) Benjamini and Hochberg (B&H). A q-value is a p-value that has been adjusted for the FDR. The Benjamini-Hochberg procedure is a method for correcting for multiple p-values. A q-value threshold of 0.05 yields a FDR of 5% among all features called significant. The q-value is the expected proportion of false positives among all features as or more extreme than the observed one Functional interactions reveal programmed cell death pathways We delved into the Génie ranking using the Reactome platform to graphically depict biological pathways with comprehensive mechanistic detail (as shown in Fig. [72]2A). Through the Voronoi diagram-based pathway overrepresentation analysis, it was revealed that in the broader programmed cell death pathway, the sub-pathways, visually represented as integrated cells, primarily included regulated necrosis (encompassing pyroptosis) and apoptosis. A comparative analysis with an independent PANoptosis-related protein database demonstrated that Génie’s compilation encompassed a significant portion of relevant proteins (n = 150), especially notable in apoptosis (52%), pyroptosis (59%), and necroptosis (100%), as depicted in Fig. [73]2B. The STRING interaction network suggested a high probability of these proteins forming a biologically cohesive community, as evidenced by their high enrichment value. Within this network, the IVI tool successfully pinpointed several pivotal hub nodes, prominently including tumor necrosis factor (TNF), NF-kappa-B essential modulator (IKBKG) and tumor necrosis factor receptor superfamily member 1A (TNFRSF1A), as illustrated in Fig. [74]2C. The 3 proteins are listed in Table [75]3 with their molecular functions. Fig. 2. [76]Fig. 2 [77]Open in a new tab Insight into programmed cell death in recurrent aphthous stomatitis through computational biology. This figure provides a view into how programmed cell death, including regulated necrosis, pyroptosis, and apoptosis, plays a role in recurrent aphthous stomatitis. (A) Through Reactome, a Voronoi diagram analysis reveals a notable overrepresentation of these cell death processes. (B) Additionally, an independent dataset analysis points to a significant involvement of proteins associated with external biotic stimuli, primarily those engaged in pyroptosis, apoptosis, and necroptosis, collectively known as PANoptosis. (C) We further constructed an interaction network that integrates proteins common to both the Génie and PANoptosis datasets (n = 150), uncovering a robust interactome with verifiable biological connectivity, evidenced by a significant protein-protein interaction (PPI) enrichment p-value. Prominent proteins within this network include TNF (tumor necrosis factor, Uniprot ID: [78]P01375), IKBKG (NF-kappa-B essential modulator, Uniprot ID: [79]Q9Y6K9), and TNFRSF1A (tumor necrosis factor receptor superfamily member 1A, Uniprot ID: [80]P19438). The Uniprot IDs or primary (citable) accession numbers are stable identifiers for proteins. For a more detailed view, the original network is available at the STRING database [81]https://string-db.org/cgi/network?taskId=bfq6HqRrmUrE&sessionId=bIO DrGwGklyC Table 3. Function of key nodes associated with PANoptosis in recurrent aphthous stomatitis Node Gene symbol Protein name (Uniprot ID) Description* 1 TNF Tumor necrosis factor ([82]P01375) Multifunctional proinflammatory cytokine. It is predominantly produced and secreted by macrophages. Its functional mechanism involves binding to receptors TNFRSF1A/TNFR1 and TNFRSF1B/TNFBR. It plays a crucial role in orchestrating a diverse array of biological activities, encompassing cell proliferation, differentiation, apoptosis, lipid metabolism, and coagulation processes. 2 IKBKG NF-kappa-B essential modulator ([83]Q9Y6K9) Regulatory subunit of the inhibitor of kappaB kinase (IKK) complex, which activates NF-kappaB resulting in activation of genes involved in inflammation, immunity, cell survival, and other pathways. Could be implicated in NF-kappa-B-mediated protection from cytokine toxicity. 3 TNFRSF1A Tumor necrosis factor receptor superfamily member 1A ([84]P19438) This receptor, widely present in cell membranes, specifically interacts with tumor necrosis factor (TNF). Upon activation, it recruits the adapter molecule FAS-associated death domain protein (FADD), which in turn brings caspase-8 into proximity with the receptor. This interaction forms the death-inducing signaling complex (DISC), facilitating the proteolytic activation of caspase-8. This event triggers a cascade of subsequent caspases, ultimately leading to the mediation of apoptosis. [85]Open in a new tab *We obtained descriptions from The GeneCards human gene database ([86]https://www.genecards.org/) While an external biological stimulus may collectively emerge as a plausible triggering factor, many of the identified proteins are involved in various pathophysiological processes. It is known that the same protein can serve different functions depending on the context, interaction partners, or its isoforms [[87]24]. Therefore, the possibility that other factors are involved must also be considered. Discussion Contemporary research into the etiopathogenesis of recurrent aphthous stomatitis has predominantly centered on inflammatory cells and overall systemic health, often overlooking the crucial role of oral mucosal epithelial cells. These cells form the primary defense barrier in the oral cavity, safeguarding against a myriad of external stimuli. In our study, we concentrated on deciphering the specific nature of the stimuli involved in recurrent aphthous stomatitis and the resultant cellular death mechanisms in oral keratinocytes. Leveraging computational biology to analyze extensive biomedical literature, our findings suggest that the triggering stimulus for the lesions in recurrent aphthous stomatitis is likely an external biotic factor. Furthermore, the mode of cell death in these instances appears to be immunogenic, possibly aligning with mechanisms characteristic of PANoptosis. Central to molecular medicine is the correlation of genes and proteins with their respective diseases, a concept well-established in the field [[88]25]. A multitude of tools have been developed to facilitate the association of genes with complex diseases. Among these, Génie has demonstrated utility, exemplified by its application in identifying risk genes for non-small cell lung cancer [[89]26] and oral lichen planus [[90]12]. Utilizing Génie, we identified approximately 1,400 genes linked to recurrent aphthous stomatitis, with a notable emphasis on genes involved in response to external biotic stimulus causing programmed cell death. At the forefront of our findings are toll-like receptors (TLRs) 2 and 4, which are pivotal in recognizing various bacteria when expressed on epithelial cells [[91]27]. TLRs, a group of membrane receptors, play essential roles in both immune regulation and maintaining epithelial barrier integrity [[92]28]. Investigations into TLR expression in recurrent aphthous stomatitis revealed that healthy oral mucosal epithelial layers typically do not express TLRs, rendering them unresponsive to pathogen-associated molecular patterns (PAMPs). However, in recurrent aphthous stomatitis, this standard TLR distribution is disrupted, with a significant presence of TLRs on the epithelial cell surface, deviating from the normal TLR architecture observed in healthy oral mucosa [[93]28, [94]29]. QuickGO (term GO:0043207) defines an external biotic stimulus as one originating from or produced by living entities. This concept is particularly relevant to the pathogenesis of recurrent aphthous stomatitis, where the oral microbiome plays a pivotal role [[95]30]. The immune response implicated in epithelial damage in recurrent aphthous stomatitis is believed to be triggered by oral microbiota. Research indicates that bacterial activities may cause oral keratinocyte lysis, and shifts in the normal oral microbiome can prompt ulcer formation, altering the microbial balance [[96]31, [97]32]. Hypotheses suggest mucolytic enzyme-secreting microorganisms could damage the mucosal pellicle, resulting in oral ulcers [[98]33]. Post-ulcer formation, specific Streptococcus species and related antigens, like S-glucosyltransferase, might infiltrate the oral mucosa, attracting cytotoxic T lymphocytes that target oral keratinocytes [[99]32–[100]34]. Salivary bacterial diversity in recurrent aphthous stomatitis patients differs from healthy individuals, correlating with disease severity. Notable microbial changes include reduced Streptococcus salivarius and increased Acinetobacter johnsonii [[101]35–[102]37], Escherichia coli [[103]36, [104]37], and Bacteroidetes [[105]36]. Additionally, Prevotella, Veillonella, and Streptococcus are linked to lesion onset and progression [[106]38, [107]39]. Our analysis of saliva samples from recurrent aphthous stomatitis patients reveals complex biological processes underpinning tissue damage, immune responses, and intracellular protein alterations, including Neisseria meningitidis presence [[108]40]. Despite advances in techniques like 16s rRNA or 16s rDNA sequencing, the key microorganisms in recurrent aphthous stomatitis pathogenesis remain elusive [[109]35, [110]36]. Although the precise initiating agents of the lesions are unidentified, focusing on microorganisms, particularly bacteria, appears to be a promising direction. Our computational biology pipeline suggests that the destruction of oral keratinocytes in the context of recurrent aphthous stomatitis occurs in an immunogenic-type cell death, some type of mechanism within PANoptosis. PANoptosis represents a sophisticated and dynamically regulated programmed cell death pathway that amalgamates key elements of pyroptosis, apoptosis, and necroptosis. This process can be triggered by a range of factors, including bacterial and fungal infections, injuries such as acute lung injury, acute respiratory distress syndrome, ischemia-reperfusion, and organ failure, or intrinsic cellular anomalies like tumorigenesis [[111]41, [112]42]. The formation and activation of the PANoptosome are pivotal in these processes. Therefore, while our results suggest a bacterial trigger, other stimuli cannot yet be ruled out. The human body possesses a diverse array of sensors attuned to detect danger associated molecular patterns (DAMPs), PAMPs, and other risk factors, playing a crucial role in initiating PANoptosis [[113]43]. A comprehensive analysis by Jian et al. [[114]44]. presents that PANoptosis is a crucial innate immune response in oral tissues, activated in reaction to microbial infections. This includes potential bacterial triggers associated with recurrent aphthous stomatitis. It is widely accepted that an unidentified antigen stimulates oral keratinocytes, leading to the production of pro-inflammatory cytokines and chemotaxis of leukocytes, ultimately resulting in the development of recurrent aphthous stomatitis [[115]45–[116]47]. Keratinocytes, functioning as potent immune cells, secrete various cytokines, chemokines, and antimicrobial peptides/proteins, thereby recruiting and activating immune cells like dendritic cells, Th1, Th2, and Th17 cells [[117]48]. Typically, oral keratinocytes produce a cytokine profile similar to epidermal keratinocytes, including interleukin-1 (IL-1), interleukin-6 (IL-6), interleukin-8 (IL-8), transforming growth factor beta (TGF-β), basic fibroblast growth factor (FGF2), and TNF (41). TNF, a central player in our functional interaction network for recurrent aphthous stomatitis cell death, is a crucial cytokine implicated in its etiopathogenesis [[118]28] and known to trigger PANoptosis [[119]41, [120]49]. As a key regulator of cell survival, apoptosis, and necroptosis, TNF signaling is vital in understanding necroptosis and its role in inflammation from infections or injuries [[121]41, [122]50]. Considering the significant roles of toll-like receptors TLR2 and TLR4 (Table [123]1), it’s plausible that bacteria, innate immune responses, and PANoptosis are intricately involved in the etiopathogenesis of recurrent aphthous stomatitis. Our study’s limitations include reliance on a computational biology pipeline using database references without conducting an additional