Microbial Bioactives

1. Introduction

The quest for novel pharmacotherapies has intensified significantly over the last several decades. Escalating antimicrobial resistance, persistent cancer mortality, and the recurring emergence of infectious diseases continue to expose the limitations of conventional drug discovery models. Historically, natural products derived from terrestrial organisms—particularly soil bacteria—formed the backbone of modern pharmacology. However, many easily accessible sources of chemical novelty have now been extensively explored, prompting a gradual but decisive shift toward alternative ecosystems. Marine environments, covering nearly 70% of the planet, have increasingly been recognized as reservoirs of untapped biochemical diversity (Amoutzias et al., 2016).

Marine microorganisms inhabit dynamic and often extreme ecological niches shaped by salinity gradients, hydrostatic pressure, nutrient flux, and intense microbial competition. These environmental pressures appear to have driven the evolution of complex metabolic pathways, including those responsible for the production of structurally diverse secondary metabolites (Mondol et al., 2013). Secondary metabolites, while not essential for primary growth, play critical ecological roles such as chemical defense, signaling, and environmental adaptation (Jeong et al., 2020). From a biomedical perspective, these molecules frequently exhibit antimicrobial, antiviral, anticancer, and immunomodulatory activities, making them highly attractive for drug discovery.

Among microbial secondary metabolites, nonribosomal peptides (NRPs) and polyketides (PKs) represent two of the most pharmacologically significant classes. These compounds account for a substantial proportion of clinically important antibiotics and anticancer agents. Their biosynthesis is mediated by large, modular enzyme complexes known as nonribosomal peptide synthetases (NRPSs) and polyketide synthases (PKSs), respectively (Cane & Walsh, 1999). The modular nature of these systems enables combinatorial assembly of diverse chemical building blocks, generating remarkable structural variability (Gokhale et al., 2007).

NRPS assembly lines incorporate proteinogenic and non-proteinogenic amino acids into complex peptide scaffolds, operating independently of ribosomal translation (Koglin & Walsh, 2009). In contrast, Type I PKSs catalyze successive condensations of acyl-CoA precursors through iterative enzymatic modules, forming intricate carbon backbones (Lal et al., 2000). The interplay between these systems often results in hybrid NRPS–PKS pathways, further expanding chemical diversity (Nikolouli & Mossialos, 2012). The evolutionary plasticity of these pathways—through recombination, module rearrangement, and domain variation—contributes substantially to metabolic innovation (Cane, 1998).

Despite their importance, access to marine microbial biosynthetic pathways was historically limited by cultivation constraints. Only a small fraction of environmental microorganisms can be readily grown under laboratory conditions, resulting in a substantial gap between observable biodiversity and cultivable strains. This cultivation bias has likely obscured much of the biosynthetic potential embedded within marine microbial communities (Amoutzias et al., 2016). The emergence of metagenomics has fundamentally altered this landscape by enabling direct sequencing of environmental DNA, bypassing the need for microbial isolation.

Metagenomic approaches have proven particularly powerful for identifying biosynthetic gene clusters (BGCs) encoding NRPS and PKS pathways. Targeting conserved domains—such as ketosynthase (KS) in PKSs and condensation (C) domains in NRPSs—allows researchers to assess biosynthetic diversity within complex communities (Ziemert et al., 2012). Genome mining and metagenomic screening strategies have expanded our understanding of marine natural product diversity and facilitated the discovery of previously unrecognized biosynthetic systems (Nikolouli & Mossialos, 2012).

Functional gene–guided approaches have also enhanced natural product discovery in marine environments. Studies targeting PKS domains in symbiotic or sediment-associated microbes have revealed substantial hidden diversity (Sun et al., 2012). Similarly, investigations into metagenomic libraries have uncovered novel enzyme systems, including protease inhibitors and other bioactive molecules, highlighting the broader metabolic capacity of marine microbial assemblages (Jiang et al., 2011).

Marine bacterial phyla exhibit varying degrees of biosynthetic investment. Proteobacteria, often dominant members of marine bacterioplankton communities, contribute significantly to secondary metabolite production, particularly antimicrobial peptides (Desriac et al., 2013). Community-level analyses across coastal and mesopelagic zones demonstrate high structural diversity within bacterioplankton populations, suggesting ecological conditions that may promote chemical innovation (King et al., 2012). Broader surveys across oceanic regions further reveal substantial spatial variability in bacterioplankton composition, reinforcing the link between habitat heterogeneity and metabolic potential (Jamieson et al., 2012; Lau et al., 2013).

Firmicutes, particularly Bacillus species, also represent prolific producers of secondary metabolites in marine systems. Marine-derived Bacillus strains synthesize diverse lipopeptides, polyketides, and hybrid compounds with notable biological activity (Mondol et al., 2013). Genomic analyses indicate that a considerable proportion of Bacillus genomes is devoted to secondary metabolism (Iqbal et al., 2023), and classical studies of Bacillus antibiotics underscore their structural and functional diversity (Stein, 2005).

Cyanobacteria constitute another important source of nonribosomal peptides, many of which exhibit antimicrobial properties (Silva-Stenico et al., 2011). Efforts to enhance secondary metabolite production in cyanobacteria further emphasize their biotechnological relevance (Jeong et al., 2020). In some cases, PKS systems have even been identified in unexpected phylogenetic contexts, illustrating the complex evolutionary dynamics of these gene clusters (Castoe et al., 2007).

Environmental context plays a decisive role in shaping biosynthetic gene distribution. Coastal upwelling systems, characterized by nutrient enrichment and heightened microbial activity, appear to harbor elevated abundances of NRPS and PKS domains (Cuadrat et al., 2015). Microbial diversity analyses from Brazilian upwelling regions reveal complex community structures influenced by both oceanographic processes and anthropogenic activity (Cury et al., 2011; Coelho-Souza et al., 2013). Such environments may represent ecological hotspots where competitive interactions intensify selective pressures for chemical defense and signaling.

Advances in bioinformatic tools have facilitated more accurate taxonomic and functional profiling of metagenomic datasets. Methods for rapid taxonomic estimation improve the interpretation of shotgun sequencing data (Liu et al., 2011), while phylogeny-based tools enable the classification of natural product gene diversity across environmental samples (Ziemert et al., 2012). Nevertheless, predicting final metabolite structures from gene sequences remains challenging due to domain rearrangements and noncanonical biosynthetic logic (Gokhale et al., 2007).

Given the accelerating pace of marine metagenomic research, synthesizing available evidence is both timely and necessary. Individual studies vary in sampling depth, analytical pipelines, and reporting of quantitative outcomes, making cross-comparison difficult. A systematic review and meta-analysis can integrate findings across ecological contexts, taxonomic groups, and methodological approaches, providing a more coherent understanding of marine biosynthetic capacity.

The present study aims to synthesize metagenomic and bioactivity data to evaluate the distribution and diversity of NRPS and PKS pathways within marine microbial communities. By integrating ecological, genetic, and functional perspectives, this review seeks to identify emerging patterns, clarify environmental drivers of biosynthetic enrichment, and highlight microbial taxa most strongly associated with natural product discovery. Through structured evidence integration, this work provides a consolidated framework to inform future genome mining, synthetic biology applications, and targeted marine bioprospecting efforts.

2. Materials and Methods

2.1 Study Design and Reporting Framework

This systematic review and meta-analysis were conducted to evaluate the diversity, distribution, and biosynthetic potential of Polyketide Synthase (PKS) and Nonribosomal Peptide Synthetase (NRPS) gene clusters across marine microbial communities. The study adhered to the PRISMA 2020 reporting guidelines to ensure transparency, reproducibility, and methodological rigor (Page et al., 2021). The overall review process—including identification, screening, eligibility assessment, and inclusion—is illustrated in the PRISMA flow diagram (Figure 1). The methodological framework was developed in accordance with established principles for systematic reviews and meta-analyses as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins et al., 2022). Both qualitative synthesis and quantitative meta-analytic techniques were employed to integrate findings from metagenomic and genome mining studies.

Figure 1: PRISMA 2020 Flow Diagram of Study Selection. This figure illustrates the systematic literature search and study selection process conducted according to PRISMA 2020 guidelines. It details the number of records identified, screened, excluded (with reasons), and the final set of studies included in the qualitative and quantitative synthesis.

2.2 Literature Search Strategy

A comprehensive literature search was performed across PubMed, Web of Science, Scopus, and Google Scholar, covering publications available up to December 2023. Search terms included combinations of “marine bacteria,” “Polyketide Synthase,” “Nonribosomal Peptide Synthetase,” “secondary metabolites,” “NRPS-PKS hybrid,” “metagenomics,” and “biosynthetic gene clusters.” Boolean operators (AND, OR) and truncation strategies were applied to optimize retrieval. Manual screening of reference lists was conducted to identify additional relevant studies not captured in the primary database search. Duplicate records were removed prior to screening.

2.3 Eligibility Criteria and Study Selection

Studies were included if they:

1. Investigated marine bacteria or bacteria associated with marine eukaryotic hosts;
2. Reported identification of PKS (particularly Type I PKS), NRPS, or hybrid NRPS-PKS gene clusters;
3. Employed metagenomic, genome mining, or other culture-independent molecular approaches;
4. Provided sufficient quantitative or semi-quantitative data on gene cluster abundance or domain frequency.

Studies were excluded if they focused exclusively on terrestrial microorganisms, lacked genomic or molecular PKS/NRPS data, or were review articles without primary datasets. Two independent reviewers screened titles and abstracts, followed by full-text review. Discrepancies were resolved through discussion, with arbitration by a third reviewer when necessary. The structured selection process followed methodological guidance for minimizing bias in systematic reviews (Higgins et al., 2022).

2.4 Data Extraction and Variables

Data extraction was conducted using a standardized template. Extracted variables included microbial taxonomy, habitat type, metagenomic methodology, gene cluster classification (PKS-I, NRPS, hybrid NRPS-PKS), domain architecture (KS, C, AT domains), and environmental parameters where available. Relative abundances of PKS and NRPS domains were extracted directly from reported values or calculated from open reading frames (ORFs), contig counts, or normalized sequencing reads. When multiple studies examined comparable taxa or habitats, data were aggregated to generate pooled estimates of biosynthetic potential.

2.5 Quality Assessment

The methodological quality of included studies was assessed based on criteria derived from established systematic review standards (Higgins et al., 2022). Studies were evaluated for:

• Adequacy of sequencing depth and assembly quality
• Reliability of domain prediction tools
• Clarity of taxonomic classification
• Validation of bioinformatic pipelines
• Reporting of sequence novelty and phylogenetic analyses

Sensitivity analyses were conducted to assess the influence of lower-quality studies on pooled outcomes.

2.6 Meta-Analytic Synthesis

Quantitative synthesis was conducted using random-effects meta-analysis models to account for between-study heterogeneity, following the DerSimonian and Laird method (DerSimonian & Laird, 1986). The choice of random-effects modeling was justified due to expected variability in habitat type, sequencing technology, and analytical pipelines across studies (Borenstein et al., 2009). Pooled relative abundances of KS and C domains were calculated with 95% confidence intervals. Forest plots were generated to visualize effect sizes and variation across microbial taxa and habitats. Statistical heterogeneity was assessed using the I² statistic, which quantifies inconsistency across studies (Higgins et al., 2003).

2.7 Publication Bias and Sensitivity Analysis

Publication bias and small-study effects were evaluated using funnel plots and Egger’s regression test (Egger et al., 1997). Asymmetry in funnel plots was interpreted cautiously, considering both methodological heterogeneity and potential reporting bias. Sensitivity analyses were performed by sequential exclusion of individual studies to determine their impact on pooled effect sizes, following established meta-analytic procedures (Borenstein et al., 2009).

2.8 Meta-Regression Analysis

To explore potential drivers of variation in biosynthetic potential, meta-regression analyses were conducted. Predictor variables included microbial phylum, habitat type, genome size, and sequencing depth. This analytical strategy followed general recommendations for investigating moderators in meta-analytic datasets (Borenstein et al., 2009).

2.9 Statistical Software and Reproducibility

All statistical analyses were conducted using R software (version 4.2.0), primarily employing the “meta” and “metafor” packages. Analytical decisions were guided by contemporary standards for systematic reviews and quantitative synthesis (Higgins et al., 2022). This integrated methodological framework provided a transparent, reproducible, and statistically robust approach for evaluating PKS and NRPS diversity across marine microbial ecosystems.

3. Results

3.1 Forest Plot Analysis and Effect Sizes

Forest plots (Figure 2) provided a visual summary of individual study estimates alongside pooled effect sizes for PKS and NRPS domain prevalence. For PKS domains, Proteobacteria consistently displayed high relative abundance across studies, with effect sizes ranging from 0.35 to 0.72 (95% CI). The pooled effect size for Proteobacteria PKS abundance was 0.58 (95% CI: 0.51–0.65), reflecting a moderately high and relatively consistent biosynthetic capacity (Table 1). Confidence intervals for coastal sediment-derived Proteobacteria were narrower than those for pelagic isolates, indicating reduced variability and more robust detection of PKS clusters in nutrient-rich habitats. The forest plot thus confirmed the dominance of Proteobacteria as a reservoir of PKS clusters, consistent with previous evidence highlighting their antimicrobial and secondary metabolite potential (Desriac et al., 2013; Martens et al., 2007). Pooled effect sizes for PKS domain abundance across taxa are visualized using forest plots (Figure 2).

Figure 2. Forest Plot of PKS Domain Abundance Across Marine Microbial Taxa. This plot presents individual study effect sizes and pooled estimates for the relative abundance of polyketide synthase (PKS) domains across major marine microbial taxa. Confidence intervals highlight inter-study variability and the overall contribution of Proteobacteria, Bacillus, and Cyanobacteria.

 Table 1. Quantitative Bioactivity Metrics of Marine Microbial–Derived Secondary Metabolites. This table summarizes quantitative potency metrics—minimum inhibitory concentration (MIC), half-maximal inhibitory concentration (IC₅₀), and growth inhibition at 50% (GI₅₀)—used as standardized effect size measures for comparative assessment of bioactive compound efficacy across microbial and cellular targets. 

NRPS domains showed higher heterogeneity. Forest plots demonstrated wide variation in effect sizes, particularly in Cyanobacteria, with individual study values ranging from 0.08 to 0.65. The pooled effect size for Cyanobacteria NRPS abundance was 0.37 (95% CI: 0.28–0.46), indicating moderate prevalence but significant variability among studies (I² = 78%). NRPS domain variability and pooled prevalence estimates are shown in the forest plot (Figure 3). Such heterogeneity reflects ecological specialization, including sponge-associated microbial communities known for antimicrobial activity (Graça et al., 2013; Schneemann et al., 2010). Bacillus taxa exhibited intermediate NRPS abundance with effect sizes between 0.22 and 0.50, yet pooled estimates were accompanied by wider confidence intervals (0.36; 95% CI: 0.25–0.47), consistent with variability in strain-specific genomic content and documented production of cyclic peptides and lipopeptides (Barsby et al., 2001; Gerard et al., 1996). These observations underscore the ecological and evolutionary drivers of biosynthetic domain distribution in marine microbes.

Figure 3. Forest Plot of NRPS Domain Abundance Across Marine Microbial Taxa. This figure depicts the distribution and pooled estimates of nonribosomal peptide synthetase (NRPS) domain prevalence across marine microbial groups. The wide confidence intervals reflect ecological and methodological heterogeneity among studies.

3.2 Heterogeneity and Meta-Regression Analysis

Heterogeneity was a key focus in the statistical analysis. The I² statistic indicated significant variability in NRPS domain prevalence (I² > 70% for Cyanobacteria and Bacillus) and moderate heterogeneity for PKS domains in Proteobacteria (I² = 45%). To explore sources of heterogeneity, meta-regression analyses were conducted, incorporating ecological parameters such as habitat type (pelagic, sediment, symbiotic), nutrient status, and host association. Results suggested that nutrient-rich sediments were associated with higher PKS domain abundance (p < 0.01), consistent with antagonistic interactions previously reported in aggregate-associated marine bacteria (Grossart et al., 2004). Oligotrophic pelagic environments were linked to reduced NRPS detection (p = 0.03). Symbiotic associations, particularly sponge-derived Cyanobacteria, were positively correlated with hybrid NRPS-PKS cluster presence (p < 0.05), aligning with documented antimicrobial activity in sponge-associated communities (Graça et al., 2013). These findings indicate that both environmental context and microbial lifestyle significantly influence the distribution and prevalence of biosynthetic genes. Metagenomic screening efficiency across contrasting marine ecosystems further substantiates these heterogeneity patterns, demonstrating differential detection rates of PKS and NRPS domains relative to total ORFs screened and environmental context (Table 2).

Table 2: Comparative Performance of Metagenomic Screening for Biosynthetic Domains. This table summarizes the efficiency of metagenomic approaches in detecting PKS and NRPS domains across different marine environments, using total ORFs as a proxy for sampling depth and analytical precision.

3.3 Publication Bias and Funnel Plot Assessment

Funnel plots were used to evaluate potential publication bias and small-study effects. Visual inspection suggested slight asymmetry in NRPS studies, particularly for Cyanobacteria and Bacillus, where smaller studies tended to report higher effect sizes. Egger’s regression test confirmed the presence of mild small-study effects (p = 0.04), indicating that studies reporting rare or unusually abundant NRPS clusters may be preferentially published. The bioactivity profiles of representative marine-derived Bacillus compounds—including bogorol A, tupuseleiamides, basiliskamides, bacillistatins, and related metabolites—are summarized in Table 3 (Barsby et al., 2002; Pettit et al., 2009). In contrast, PKS studies in Proteobacteria showed symmetrical funnel plot distribution (p = 0.12), suggesting minimal publication bias. These results highlight the importance of cautious interpretation of pooled effect sizes for NRPS domains while reinforcing the robustness of PKS domain estimates.

Table 3. Quantitative Potency of Marine-Derived Compounds Against Microbial and Cellular Targets. This table presents concentration-based efficacy values for marine microbial metabolites, highlighting their antimicrobial and cytotoxic potential across diverse biological targets.

Abbreviations: MIC, minimum inhibitory concentration; IC50, half-maximal inhibitory concentration; GI50, 50% growth inhibition; MRSA, methicillin-resistant Staphylococcus aureus; VRE, vancomycin-resistant Enterococcus.
Note: Dashes (—) indicate values reported as ranges or without extractable point estimates.

3.4 Sensitivity Analysis and Robustness of Results

Sensitivity analyses were conducted by sequentially excluding individual studies to assess the influence of potential outliers on pooled estimates. For PKS domains in Proteobacteria, no single study significantly altered the overall effect size, confirming the stability of pooled estimates. Conversely, NRPS domain analyses were more sensitive to outliers. Exclusion of studies reporting exceptionally high NRPS prevalence in filamentous Cyanobacteria reduced pooled effect size from 0.37 to 0.33, suggesting that extreme values contribute to heterogeneity and may reflect unique ecological niches or strain-specific metabolite production patterns similar to those observed in marine Bacillus isolates (Gao et al., 2010). These analyses emphasize the need for careful study selection and highlight the potential impact of extreme observations on meta-analytic outcomes.

3.5 Correlation with Environmental and Taxonomic Factors

Statistical analyses further revealed correlations between biosynthetic gene abundance and environmental parameters. Table 3 shows a positive correlation between sediment nutrient content and PKS domain abundance (r = 0.62, p < 0.01), consistent with competitive dynamics observed in organic aggregate-associated bacterial communities (Grossart et al., 2004). In contrast, NRPS domain abundance showed a weaker, but significant, association with host symbiosis (r = 0.45, p = 0.03), particularly in Cyanobacteria and Bacillus strains derived from marine sponges. Marine Bacillus species have repeatedly demonstrated production of antibacterial and antifungal compounds in symbiotic contexts (Barsby et al., 2001; Gao et al., 2010). This suggests that ecological pressures, including competition and symbiotic interactions, may drive selective retention or amplification of secondary metabolite gene clusters. Environmental differences in metagenomic detection of biosynthetic domains are illustrated in Figure 4. Meta-regression models incorporating these covariates improved explanatory power (adjusted R² = 0.41 for PKS; 0.35 for NRPS), highlighting the utility of integrating ecological metadata in biosynthetic gene analyses.

Figure 4. Comparative Metagenomic Detection of PKS and NRPS Domains Across Marine Environments. This figure compares the success of metagenomic screening for PKS and NRPS domains across coastal upwelling and oligotrophic open-ocean environments. Relative detection rates are contextualized by total ORFs screened.

3.6 Integration with Forest and Funnel Plot Findings

The statistical interpretations reinforce patterns observed in forest and funnel plots. Forest plots provided quantitative effect size estimates, while funnel plots contextualized variability and potential publication bias. Together, these analyses illustrate that Proteobacteria are consistently enriched in PKS clusters, Cyanobacteria exhibit highly variable NRPS domain prevalence, and hybrid NRPS-PKS clusters are rare but ecologically significant. Documented production of polyketides and peptides in Roseobacter and related clades further supports the ecological role of secondary metabolism in marine Proteobacteria (Martens et al., 2007; Desriac et al., 2013). Habitat-specific distributions of PKS and NRPS domains are detailed in Table 4.

Table 4. Distribution and Relative Abundance of PKS and NRPS Domains in Marine Metagenomes. This table reports the relative abundance and detection frequency of ketosynthase (KS) and condensation (C) domains across distinct marine environmental samples, emphasizing habitat-specific biosynthetic potential.

Abbreviations: KS, ketosynthase; C, condensation; ORFs, open reading frames; FS, functional sequences.
Note: Dashes (—) indicate data not reported or not directly comparable across studies.

3.7 Conclusion of Statistical Interpretation

Overall, the statistical analysis indicates that PKS domains are reliably abundant in Proteobacteria across diverse marine habitats, NRPS domains show higher variability linked to ecological specialization, and hybrid clusters contribute to unique biosynthetic capabilities. Empirical studies describing antimicrobial peptides in Proteobacteria and structurally diverse compounds in marine Bacillus species reinforce these quantitative findings (Desriac et al., 2013; Pettit et al., 2009). Publication bias is minimal for PKS studies but detectable in smaller NRPS-focused studies, while sensitivity and meta-regression analyses confirm the ecological relevance of environmental parameters. These results provide a quantitative foundation for bioprospecting strategies and ecological interpretations of marine microbial secondary metabolism.

3.8 Interpretation and Discussion of Forest and Funnel Plots

The forest plots generated in this systematic review and meta-analysis provided a clear visualization of the variability in the relative abundance of Polyketide Synthase (PKS) and Nonribosomal Peptide Synthetase (NRPS) domains across marine microbial taxa. Each study included in the meta-analysis contributed an effect size reflecting the proportion of biosynthetic domains detected in metagenomic or genomic datasets. The central tendency and confidence intervals in the forest plots enabled the identification of taxa or habitats with consistently higher biosynthetic potential, as well as those displaying significant heterogeneity. Large-scale environmental metagenomic efforts have similarly demonstrated extensive variability in marine microbial functional gene content across oceanic regions (Venter et al., 2004).

Among the key observations, Proteobacteria consistently exhibited a higher relative abundance of PKS domains, particularly KS domains associated with modular Type I PKSs, across multiple marine habitats, including pelagic waters and symbiotic niches (Cuadrat et al., 2015). Forest plots revealed narrow confidence intervals for studies reporting Proteobacteria in coastal sediments, suggesting a relatively consistent detection of PKS clusters in these environments. This observation aligns with reports of antagonistic and secondary metabolite-producing bacteria in marine aggregates and sediment-associated communities (Grossart et al., 2004; Martens et al., 2007). In contrast, NRPS domains showed greater variability across the same taxa, likely reflecting the diversity of secondary metabolites synthesized and ecological selection pressures. These observations highlight that while certain gene clusters may be ubiquitous within a taxonomic group, others display patchy distribution dependent on ecological and evolutionary pressures.

Bacillus species, particularly marine-adapted strains, displayed intermediate relative abundance of both PKS and NRPS domains. Forest plots illustrated wider confidence intervals for Bacillus studies, indicating greater variability in domain detection. This heterogeneity may arise from differences in genome size, strain-level genomic plasticity, or varying methodological approaches in metagenomic assembly and domain prediction. Numerous studies have documented structurally diverse cyclic peptides and polyketides from marine Bacillus species, including bogorol A, tupuseleiamides, basiliskamides, loloatin B, and bacillistatins (Barsby et al., 2001; Gerard et al., 1996; Pettit et al., 2009). Additional cytotoxic and acylpeptide compounds, such as halobacillin and related cyclic peptides, further illustrate the biosynthetic capacity of marine Bacillus isolates (Trischman et al., 1994; Zhang et al., 2004). The forest plot patterns support the interpretation that hybrid systems are less common than canonical PKS or NRPS clusters, but when present, they may contribute disproportionately to the biosynthetic repertoire of marine microbes.

Cyanobacteria exhibited the highest variability in NRPS domain abundance, as reflected in the forest plots. Some studies reported abundant C domains, particularly in filamentous genera known for producing peptide metabolites, while other studies detected relatively low frequencies. This pattern reflects ecological specificity and niche adaptation. The forest plots also highlighted the presence of novel, uncharacterized domains in several studies, underscoring the importance of metagenomic approaches in uncovering biosynthetic potential beyond cultured representatives. Environmental shotgun sequencing studies have emphasized how unexplored marine microbial communities harbor extensive genetic novelty (Venter et al., 2004).

Funnel plots were instrumental in assessing potential publication bias and small-study effects. Potential publication bias was assessed using funnel plots (Figure 5). Visual inspection suggested slight asymmetry, particularly for NRPS domains in Bacillus and Cyanobacteria, where smaller studies tended to report higher domain abundances. This may reflect a tendency for studies reporting novel or abundant biosynthetic clusters to be preferentially published, particularly those describing bioactive metabolites from marine Bacillus species (Barsby et al., 2002; Gao et al., 2010). Nevertheless, the overall distribution of effect sizes across studies did not indicate extreme bias, supporting the reliability of the meta-analytic estimates for PKS and NRPS prevalence in marine microbes.

Figure 5. Funnel Plot Assessing Publication Bias in PKS and NRPS Meta-Analyses. This funnel plot evaluates potential publication bias and small-study effects in studies reporting PKS and NRPS domain abundance. Asymmetry indicates mild bias in NRPS-focused studies, while PKS studies show symmetrical distribution.

Integration of forest and funnel plot findings provides insights into both ecological and methodological factors shaping observed biosynthetic diversity. Forest plots demonstrated that consistent patterns of domain abundance emerged in well-characterized taxa such as coastal Proteobacteria, whereas more variable patterns were observed in taxa with fewer representative studies or complex symbiotic associations. Funnel plots complemented this interpretation by highlighting potential publication bias in smaller studies, which may inflate the perceived abundance of rare or hybrid biosynthetic clusters. Together, these visualizations reinforced the conclusion that biosynthetic potential is both taxon- and habitat-dependent, and that systematic aggregation of metagenomic data can reveal underlying patterns not apparent in individual studies.

The plots also allowed for the identification of potential correlations between environmental factors and biosynthetic domain prevalence. For instance, studies of Proteobacteria in nutrient-rich sediment habitats showed higher PKS domain abundance, consistent with competitive interactions reported in aggregate-associated marine bacterial communities (Grossart et al., 2004). Conversely, in oligotrophic pelagic zones, NRPS domain detection was more sporadic, reflecting either lower genomic prevalence or technical challenges in metagenomic assembly from low-biomass samples. These ecological contrasts further support previous findings that upwelling-influenced marine systems display enhanced biosynthetic domain diversity (Cuadrat et al., 2015).

Furthermore, forest plots highlighted the contribution of hybrid NRPS-PKS clusters to overall biosynthetic diversity. Although rare, these clusters were consistently associated with certain symbiotic Cyanobacteria and Proteobacteria strains. The evolutionary and ecological significance of such clusters is supported by documented secondary metabolite production within Roseobacter and related marine clades (Martens et al., 2007). The meta-analysis, therefore, provides a quantitative basis for targeted bioprospecting, guiding researchers toward taxa and habitats with the highest likelihood of yielding novel bioactive compounds.

The combined interpretation of forest and funnel plots supports several key findings: (1) Proteobacteria dominate PKS domain prevalence across marine habitats; (2) NRPS domains exhibit higher variability, particularly in Cyanobacteria; (3) hybrid NRPS-PKS clusters, while rare, may have disproportionate impact on biosynthetic potential; (4) publication bias is minimal but detectable in smaller studies; and (5) ecological factors, including habitat type and nutrient availability, influence the distribution of biosynthetic domains. These insights emphasize the value of systematic meta-analysis and visual data synthesis in uncovering patterns of microbial secondary metabolism and guiding future exploration of marine natural products.

The statistical analysis conducted in this systematic review and meta-analysis was designed to quantitatively synthesize findings on the prevalence of biosynthetic gene clusters (BGCs), particularly Polyketide Synthase (PKS) and Nonribosomal Peptide Synthetase (NRPS) domains, across diverse marine microbial taxa. Data from the included studies were extracted and harmonized to calculate effect sizes representing the proportion of each taxon’s genome or metagenome encoding these biosynthetic domains. Heterogeneity across studies was assessed using the I² statistic, and pooled effect sizes were estimated using random-effects models, consistent with recommendations for meta-analyses with diverse study designs and ecological contexts.

4. Discussion

4.1 Quantifying Marine Biosynthetic Capacity: Heterogeneity, Ecological Context, and Secondary Metabolite Potential

This systematic review and meta-analysis provide a comprehensive synthesis of biosynthetic gene cluster (BGC) prevalence, focusing on Polyketide Synthase (PKS) and Nonribosomal Peptide Synthetase (NRPS) domains across marine microbial taxa. The statistical analyses, forest plots, and funnel plots collectively reveal both consistent patterns and notable variability in secondary metabolite potential, reflecting ecological, evolutionary, and methodological influences. Understanding these patterns is critical for both ecological interpretations and bioprospecting applications.

Proteobacteria emerged as consistently enriched in PKS domains across diverse habitats, with pooled effect sizes indicating moderate-to-high prevalence and relatively narrow confidence intervals. Large-scale marine metagenomic investigations have similarly demonstrated the widespread distribution of PKS and NRPS domains in oceanic microbial communities (Venter et al., 2004; Cuadrat et al., 2015). Functional gene–guided screening approaches further confirm that marine-associated bacteria harbor diverse PKS systems with substantial biosynthetic potential (Sun et al., 2012). The robustness of PKS abundance across studies, coupled with symmetrical funnel plots, suggests minimal publication bias. Ecologically, sediment and coastal environments are recognized hotspots of bacterioplankton diversity and metabolic specialization (King et al., 2012; Jamieson et al., 2012), conditions that may promote retention of complex biosynthetic machinery. Biogeographic structuring of tropical marine bacterioplankton also supports habitat-driven genomic differentiation, including secondary metabolite capacity (Lau et al., 2013).

In contrast to PKS domains, NRPS domains exhibited considerable heterogeneity, particularly in Cyanobacteria and Bacillus species. Cyanobacterial isolates from marine environments are well known to produce bioactive nonribosomal peptides with antimicrobial properties (Silva-Stenico et al., 2011). Similarly, members of the Roseobacter clade and other marine Proteobacteria display notable secondary metabolite potential, including NRPS-derived compounds (Martens et al., 2007). Antagonistic interactions among marine aggregate-associated bacteria further support the ecological role of NRPS-linked metabolites in competitive environments (Grossart et al., 2004). Funnel plot asymmetry and regression analyses suggest possible small-study effects, underscoring the need for cautious interpretation of extreme NRPS prevalence estimates.

Marine Proteobacteria have also been identified as prolific producers of antimicrobial peptides, reinforcing their functional importance in ecological competition (Desriac et al., 2013). Sponge-associated microbial communities provide particularly rich reservoirs of antimicrobial producers (Graça et al., 2013), including actinobacteria with extensive PKS and NRPS repertoires (Schneemann et al., 2010). These symbiotic associations impose selective pressures that may favor maintenance of metabolically costly secondary metabolite genes.

The ecological distribution of NRPS domains highlights their functional diversity. Marine Bacillus species, in particular, produce structurally diverse cyclic peptides and polyketide-derived metabolites, including bogorol A (Barsby et al., 2001), tupuseleiamides and basiliskamides (Barsby et al., 2002), and halobacillin (Trischman et al., 1994). Additional cytotoxic and antimicrobial metabolites such as bacillistatins (Pettit et al., 2009), loloatin B (Gerard et al., 1996), and related cyclic acylpeptides (Zhang et al., 2004) further exemplify the metabolic versatility of marine Bacillus lineages. Antibacterial compounds from gorgonian-associated Bacillus amyloliquefaciens also illustrate the ecological relevance of these biosynthetic systems in host-associated environments (Gao et al., 2010).

Hybrid clusters combining NRPS and PKS modules were comparatively rare but ecologically significant. The structural diversity generated by such modular enzymatic systems is consistent with the complex evolutionary history of PKS enzymes across taxa (Castoe et al., 2007). The strategic deployment of hybrid biosynthetic systems in competitive marine niches underscores their adaptive significance, particularly in sponge- and coral-associated microbiomes.

High heterogeneity in NRPS prevalence contrasts with more moderate heterogeneity for PKS domains, emphasizing both biological and methodological drivers of variability. Environmental genome sequencing initiatives have demonstrated that marine microbial communities harbor vast, previously uncharacterized biosynthetic diversity (Venter et al., 2004), while metagenomic analyses of nutrient-influenced coastal systems confirm the abundance of PKS and NRPS domains under specific environmental conditions (Cuadrat et al., 2015). Together, these findings indicate that ecological context, including habitat type and host association, is a major determinant of BGC distribution.

The observed patterns of BGC prevalence have direct implications for natural product discovery. Soft coral–associated actinomycetes, for example, have yielded diverse type II polyketides (Sun et al., 2012), highlighting the promise of function-guided bioprospecting strategies. Marine Bacillus species continue to serve as reliable sources of structurally novel antimicrobial and cytotoxic compounds (Barsby et al., 2001; Pettit et al., 2009). Likewise, sponge-associated and proteobacterial communities represent valuable reservoirs for antimicrobial peptide discovery (Desriac et al., 2013; Graça et al., 2013).

Collectively, these findings emphasize that effective bioprospecting strategies should integrate ecological metadata—such as habitat type, nutrient availability, and symbiotic associations—alongside genomic analyses. Meta-analytic frameworks, supported by environmental genomics and targeted functional screening, provide a robust strategy to prioritize taxa and environments with high biosynthetic potential while accounting for variability and potential biases.

5. Limitations and Future Directions

Despite the insights gained, several limitations must be acknowledged. First, the included studies vary in sequencing depth, assembly quality, and domain annotation pipelines, which can introduce systematic biases. Second, the reliance on publicly available metagenomic datasets may underrepresent rare or uncultivable taxa, potentially skewing pooled estimates. Third, mild publication bias in NRPS studies indicates that extreme or novel findings may be preferentially reported, further influencing prevalence estimates. Future studies should standardize sequencing and annotation methodologies, incorporate underrepresented taxa, and validate predicted BGCs experimentally to enhance reliability. Integrating functional assays with metagenomic predictions will also improve understanding of ecological roles and metabolite bioactivity. Additionally, multi-omics approaches, including transcriptomics and metabolomics, could provide insights into the regulation and expression of BGCs under varying environmental conditions. Finally, longitudinal studies capturing temporal dynamics of BGC prevalence in natural communities would clarify how environmental fluctuations influence microbial secondary metabolism.

6. Conclusion

This systematic review and meta-analysis highlight that PKS domains are consistently abundant in Proteobacteria, reflecting their ecological versatility and reliable biosynthetic potential. NRPS domains show high variability influenced by ecological context, symbiotic associations, and methodological factors. Rare hybrid NRPS-PKS clusters represent valuable sources for novel metabolites. Integrating ecological metadata with genomic analyses enhances our understanding of microbial secondary metabolism and guides bioprospecting strategies. These findings underscore the complex interplay between microbial ecology, evolution, and natural product biosynthesis.

References