Climate-driven transition in microbial deterioration and protection of stone surfaces at cultural heritage sites

As of September 2023, the World Heritage List recorded a total of 1154 cultural World Heritage Sites worldwide, including 897 cultural heritage sites (https://whc.unesco.org/en/list/). Over time, immovable cultural heritage exposed to the natural environment conditions experiences deterioration due to physical, chemical, and biological weathering processes^1,2,3,4. The Angkor monuments in Cambodia⁵, the Pharaonic sandstone monuments in Luxor, Egypt⁶, the statues of the Dazu Rock Carvings in China⁷, and many other stone cultural heritage sites exhibit such damage. This type of damage not only threatens the integrity of these ancient monuments, but also leads to irreparable losses in inherited culture. Therefore, mitigating the destruction of cultural heritage sites and developing protective strategies have garnered widespread global attention.

Biofilms are aggregates of microorganisms and their extracellular polymeric substances (polysaccharides, proteins, and environmental DNA)⁸. Studies have shown that microorganisms, particularly bacteria, play an important role in the formation of biofilms on the stone surfaces of heritage sites⁹. However, biofilms have been reported to exert two distinct effects on stone: biodeterioration and bioprotection⁸. Biodeterioration, ranging from material discoloration to structural damage, can result in a loss of the historical and aesthetic value of cultural heritage sites^10,11,12. This process can damage the rock structures through physical penetration, organic and inorganic acid corrosion, and the redox reactions of mineral cations^13,14,15. Conversely, bioprotection refers to the positive effects of biofilms on cultural heritage by which biofilms form protective layers on stone surfaces through biomineralization, such as calcium carbonate and calcium oxalate precipitation, which enhances resistance to environmental stress^16,17. Acting as physical barriers, biofilms can help protect stone surfaces from weathering, acid rain, and UV radiation¹⁸. However, microbial deterioration and protection are influenced by a combination of multidimensional factors, such as the microbial community, climatic conditions, and rock properties. The mechanisms underlying the functional transition of biofilms is not well understood, which limits the design and development of biointervention strategies.

A growing body of research suggests that climatic gradients may drive the transition of microbial functions between biodeterioration and bioprotection^19,20,21. The structure of microbial communities varies greatly across different heritage sites and is influenced not only by the type of stone but also by environmental factors such as temperature, precipitation, sunlight, and salinity^22,23,24. Generally, warm and humid climates provide favorable environmental conditions for the growth of most organisms^9,25. Precipitation influences the diversity and abundance of microbial communities²⁶, while an increase in salinity reduces microbial and enzyme activity²⁷. An increase in pollutant concentrations lowers the complexity and stability of microbial ecological networks²⁸. Although several studies have examined the effect of climatic conditions on the structure of microbial communities, cross-climatic investigations on microbial deterioration and protection remain scarce. Consequently, the effects of climate-driven transitions on these microbial functions are largely unexplored, and their underlying mechanisms are yet to be elucidated.

This study evaluated the dual roles of biofilms at cultural heritage sites by examining the relationship between the relative bioprotection rate of biofilms and climatic conditions. Specifically, 16S rRNA data from 91 sampling points from 10 World Heritage Sites across different climatic environments in East and South Asia were analyzed. By assessing the microbial community structures under varying climatic conditions, we explored the impact of climate on the dominant species, diversity, and network structure of microbial communities. Additionally, functional genes associated with biofilm protection and metabolic processes involved in deterioration were predicted to determine the biofilm protection rate. This rate was subsequently used to evaluate the impact of biological protection and deterioration on cultural heritage. Finally, the relationship between biofilm protection rate and climatic conditions was analyzed to elucidate the key mechanisms through which climatic gradients affect biofilm function. The results of this study provide new insights into the climate-driven transition of biofilm functions, promoting the sustainable conservation of stone cultural heritage sites by mitigating microbial deterioration and designing innovative biointervention strategies.

Climate-driven variations of microbial communities at stone cultural heritage sites

To determine how climatic conditions affect microbial communities on stone surfaces, the microbial community compositions of 91 sampling points from ten heritage sites in East and South Asia under different climatic conditions were analyzed. Based on the sampling location, these microbial communities were classified into three groups: N, ST, and T, which corresponded to temperate, subtropical, and tropical climates, respectively. Non-metric multidimensional scaling (NMDS) based on Bray–Curtis distances reflected the β-diversity of microbial community composition across different climatic environments (stress = 0.184) (Fig. 1a). The results showed significant differences in microbial structures among the different climatic conditions (PERMANOVA: F = 4.308, P < 0.001). Notably, the ST group overlapped with both the N and T groups, whereas the N and T groups were distinctly separated. This indicated that bacterial communities different notably between temperate and tropical climates.

Chao and Shannon indices were used to measure the alpha diversity of different bacterial communities. Both indices exhibited an increasing trend with increasing precipitation (Figs. 1b, c). The increase in precipitation can provide sufficient water resources for microbial communities, thereby promoting the growth in the total species number and enhancing species richness within the community, ultimately increasing the overall species diversity²⁹. However, the relationship between bacterial alpha diversity and minimum temperature of the coldest month (MINTCM) was not linear (Figs. 1d, e). The highest Shannon and Chao indices were observed in samples with minimum temperature between 0 and 10 °C (approximately 25–30°N latitude), likely because temperature extremes increase interspecies competition and favor the growth of strains better adapted to the environment. In contrast, the abundance of other species with low-temperature adaptability decreased or even ceased to exist³⁰. These results suggest that the alpha diversity of the microbial community was strongly influenced by precipitation. The ranking of climatic factors based on their importance to the alpha diversity of the microbial community further supports this, as precipitation-related factors, including precipitation of the wettest quarter (PWETQ), precipitation of the warmest quarter (PWARQ), and mean annual precipitation (MAP), received the highest importance scores among all climatic factors (Supplementary Fig. 1).

In terms of community composition, Actinomycetota, Pseudomonadota, and Cyanobacteriota were the most widely distributed bacterial phyla on the stone surfaces (Fig. 1f). Their combined relative abundances exceeded 50%, making them the dominant bacterial phyla across all studied heritage sites. Climatic conditions significantly influenced the abundance of the dominant bacterial phyla (Supplementary Fig. 2). The relative abundances of Actinomycetota and Cyanobacteriota decreased with increasing precipitation and temperature, whereas Pseudomonadota accumulated along the precipitation and temperature gradients. Actinomycetota and Cyanobacteriota are typically adapted to drier environments^31,32. In particular, the abundance of Cyanobacteriota decreased as water availability decreases³². Conversely, Pseudomonadota are sensitive to water, and their abundance tends to show strong and consistent variation with precipitation³³.

To reveal potential bacterial interactions at the heritage sites under different climatic conditions, an operational taxonomic units (OTU)-based co-occurrence network was constructed (Fig. 2; Supplementary Table 1). Among these, microbial communities in tropical climates (T) exhibited the most complex network structures. Their core networks, composed of nodes (OTUs) and edges (correlations between OTUs), were the largest, with N = 1024 and L = 10,214. This indicated that bacterial communities in tropical climates have stronger interconnections than those in temperate climates. Studies have shown that in warm and humid environments, microbial communities tend to be more complex because they can sustain stronger ecological interactions³⁴. These stable environmental conditions provide ample water and suitable temperatures for bacterial growth and reproduction, thereby promoting interconnectivity. However, the Venn diagram showed that the subtropical bacterial community had the highest number of unique OTUs (Supplementary Fig. 3), which is likely related to the broader distribution range of the subtropical samples.

Climate-driven microbial functions of biofilms

Biofilms on stone surfaces may exert both protective and detrimental effects, which are controlled by microbial metabolic functions. To assess the influence of climatic conditions on microbial metabolic function, we analyzed the functional genes related to negative or positive metabolic processes in biofilms (Supplementary Table 2)³⁵. Microbial metabolic functions were closely correlated with climatic factors (Fig. 3a). Correlations between positive and negative metabolic functions and climatic factors were consistent. These functions were significantly positively correlated with temperature or precipitation extremes (such as minimum temperature (MINTCM) and precipitation of the wettest quarter (PWETQ)), and negatively correlated with temperature seasonality (TSEA) (P < 0.001). This relationship may occur because higher temperatures and humid environments typically enhance microbial growth and metabolic activity, thereby promoting metabolic processes³⁶. However, when temperature fluctuations are large, they may affect the stability of microbial metabolism, inhibiting the normal progression of certain metabolic functions³⁷. Notably, the maximum temperature of the warmest month (MAXTWM) was negatively correlated with assimilatory nitrate reduction and denitrification but positively correlated with oxalate biosynthesis. Although these three metabolic processes are considered to have positive functions, they exhibit different trends under the influence of different climatic factors. This suggests that the climatic factors dynamically affect biofilm functions.

To quantitatively assess the influence of bacterial communities on cultural heritage materials, the “biofilm protection ratio” that characterizes the dual role of biofilm was analysed^8,38. The “biofilm protection ratio” is the ratio between the sum of natural bioprotection effect genes and the sum of biodeterioration effect genes. A value greater than 1 indicates that the biofilm plays a protective role, meaning that the non-colonized areas exhibit a more severe deterioration pattern than biofilm-colonized areas. According to this study, the bioprotection ratio was closely related to climatic factors (Figs. 3b, c). Specifically, MINTCM and MAP showed a positive correlation with the bioprotection ratio, with correlation coefficients of R = 0.67 and R = 0.44, respectively (Figs. 3b and 3c, P < 0.001). This indicated that the protective role of biofilms is driven by climatic conditions. Biofilms exhibit strong protective effects at heritage sites in warm and humid tropical and subtropical environments.

Key functional genes that control the protective role of biofilms

Climate drives the acceleration of the microbial geochemical cycle affecting the deterioration or protection of stone, as evidenced by the significant correlation between the N, S and C cycles and climatic factors (Supplementary Fig. 4). To explore the mechanisms by which microbial communities contribute to the protection of stone cultural heritage sites, we analyzed the variations in functional genes related to nitrogen (N), sulfur (S), and carbon (C) cycles under different climatic conditions (Fig. 4), focusing on their effects on the biofilm protection ratio. These functional genes regulate key metabolic pathways and drive biofilms formation to protect or degrade stone surfaces under different climatic conditions.

The abundance of nitrogen cycle genes was significantly correlated with the minimum annual temperature (MINTCM) (P < 0.001, Fig. 4a). The abundance of dissimilatory nitrate-reduction genes (napAB, narGHI, nrfAH, and nirBD) increased with increasing temperature (R = 0.39, P < 0.001). These genes facilitate the conversion of nitrate into ammonium (NO₃⁻ → NO₂⁻ → NH₄⁺)³⁹, which reduced the risk of acid corrosion of stone caused by nitrate accumulation and enhanced the protective effect of biofilm on heritage⁴⁰. In addition, the abundance of assimilated nitrate reduction genes (nasAB and nirA) and denitrification-related genes (nirKS, norBC, and norZ) decreased with increasing temperature. However, its abundance is much lower than that of the dissimilatory nitrate reduction gene, which makes biofilms in tropical environments more likely to exert protective effects. This finding shows that the effect of biofilms on stone is a dynamic proces dominated by dissimilatory nitrate reduction.

Moreover, assimilatory sulfate reduction and dissimilatory sulfate reduction play positive roles in the S cycle of microorganisms, whereas sulfur oxidation plays a negative role. The abundance of genes involved in assimilatory sulfate reduction (cysJI and sir) increased with increasing minimum annual temperatures (R = 0.24, P < 0.05). In warmer environments, microorganisms preferentially utilize sulfate (SO₄²⁻) to synthesize sulfur-containing organic compounds, thus minimizing the corrosive effects of sulfuric acid on stone surfaces. Conversely, genes associated with negative effects, such as sulfur oxidation, were negatively correlated with temperature (Fig. 4b). In hot and humid environments, the detrimental metabolic processes of biofilms were weaker.

Chemolithoautotrophic organisms can fix CO₂ through the Calvin–Benson–Bassham (CBB) and Wood–Ljungdahl (WL) pathways^41,42,43. Genes associated with these cycles account for 0.25%–0.90% of the entire dataset. The genes associated with Carbon fixation were not significantly correlated with MINTCM (P > 0.05). Microorganisms involved in the C cycle can thrive in both cold and warm environments. On nutrient-poor rock surfaces, carbon fixation plays an important role in the initial formation of biofilms by providing bioavailable organic carbon. However, from the perspective of cultural heritage preservation, this process may have potential negative impacts. The organic matter produced by carbon fixation can create favorable conditions for other microorganisms, such as algae and lichens, which may accelerate the degradation of heritage materials.

Furthermore, several metabolic pathways with potential protective functions on the stone surface, including biomineralization, oxalate biosynthesis and metal resistance-related functional gene abundance, were positively correlated with MINTCM (P < 0.01). Biomineralization-related genes may promote the precipitation of carbonate minerals, enhancing the connection of micro-pores in the stone¹⁸. The oxalate biosynthetic pathway can form an insoluble calcium oxalate layer, providing a protective film that resists further chemical erosion⁴⁴. Genes related to metal resistance help mitigate microbial stress caused by heavy metals, thereby stabilizing the biofilm structure and alleviating the deterioration process⁴⁵. Therefore, in warmer areas, the protective effect of biofilms may be further enhanced.

Climate is vital for microbial communities and functions of stone heritage

The microbial communities on stone cultural heritage sites under different climatic conditions were predominantly composed of Actinomycetota, Pseudomonadota, and Cyanobacteriota, each playing a distinct ecological role. Actinomycetota and Pseudomonadota, primarily associated with filamentous bacteria⁴⁶, were observed on sandstone surfaces such as those at Dazu Rock Carvings. Studies have shown that these microbial groups are influenced by rainfall⁴⁷, which can modify the conditions on stone surfaces and impact microbial colonization. Cyanobacteriota were present in all samples as photosynthetic autotrophs⁴⁸. Notably, the genus Chroococcidiopsis within Cyanobacteriota is commonly found in extreme and arid habitats and was relatively abundant in samples from Beishiku Grottoes¹². Their widespread distribution across various habitat types suggests strong adaptability to arid conditions. Network analysis revealed the potential bacterial interactions at heritage sites under different climatic conditions. Pechlivanis et al. found that microbial communities in tropical regions exhibited lower diversity but maintained a denser network, which is consistent with the conclusion of this study⁴⁹.

In recent years, the beneficial role of biofilms on stone heritage, that is, bioprotection, has garnered increasing attention. Our findings suggest that biofilms in tropical environments may offer greater protection (Fig. 5). This observation could be linked to the high expression levels of genes related to positive metabolic pathways in microbial communities under tropical climate. In particular, genes associated with nitrate reduction (napAB and narGHI) and sulfate reduction (cysJI and sir) contribute to mitigating material degradation by consuming soluble salts³⁵. The dissolution and crystallization of salt can exert physical stress to stone, resulting in erosion, cracking, or pulverization⁵⁰. Nitrate reduction converts nitrate to nitrite or ammonium (NH₄⁺), while sulfate reduction converts sulfate to hydrogen sulfide. Both processes help to alleviate the damage to stone by altering the salt concentration in the surrounding environment.

It should be emphasized that, although this study indicates that these genes may be related to heritage bioprotection or biodeterioration, the underlying mechanisms still needs further study and verification. For example, while assimilatory sulfate reduction primarily supports biosynthetic processes⁵¹, the hydrogen sulfide (H₂S) produced during dissimilatory sulfate reduction—a volatile and highly reactive gas— has the potential to exacerbate the deterioration of stone surfaces⁵². When specific metal ions, such as iron, are present, H₂S can be oxidized to sulfuric acid (H₂SO₄), which then reacts with the calcium in the rock to form gypsum (CaSO₄·2H₂O). This leads to the accumulation of acidic substances, ultimately resulting in surface crust formation, granular disintegration, increased porosity, and loss of cohesion⁵³. Therefore, although sulfate reduction consumes sulfate ions, its products, especially H₂S produced during dissimilatory sulfate reduction, may still react with minerals and accumulate in poorly ventilated or humid environments, thereby triggering localized stone deterioration.

Oxalate biosynthesis and biomineralization are the crucial processes to protect the stone heritage from erosion³⁵. Genes associated with these processes were identified in each sample. The biosynthetic pathway of oxalate is particularly important in the formation of insoluble calcium oxalate crystals. These crystals can form a protective barrier on the surface of the stone and effectively resist the chemical erosion of environmental factors such as acid rain and ultraviolet light⁵⁴. In addition, genes related to biomineralization can promote the precipitation of carbonate calcium, which can fill pores to strengthen the microstructure of stone⁵⁵. Gene markers related to metal ion absorption and regulation can be used as an indicator of ion extraction in rocks^56,57. Metabolism genes of siderophores indicate that bacteria promote the absorption of iron released during rock dissolution by generating siderophores, thereby enhancing the leaching of mineral elements⁵⁶. Metal resistance genes also help reduce heavy metal stress, stabilize biofilm structure, and slow down biofilm deterioration^45,57.

Sustainable conservation strategies for cultural heritage

Our study reveals the dual role of biofilms in the protection and deterioration of cultural heritage sites, with climate being a key factor influencing biofilm function. Specifically, biofilms in tropical climates may offer protective effects, but they can also contribute to deterioration through biochemical processes (such as sulfate and nitrate reduction) and physical processes, including the retention of capillary water, which accelerates abiotic weathering. In contrast, biofilms in cold and dry temperate climates tend to exhibit stronger deteriorative effects. Therefore, biointervention strategies should be tailored to the specific climatic conditions of heritage sites. In hot and humid areas, while biofilms may have protective benefits, it is essential to monitor and maintain them in their original state to balance their protective and deteriorative effects. In these regions, microbial community-directed regulators, such as the inhibition of acid-producing bacteria or the inoculation of bacteria with nitrate-reducing functions, could be developed using environmental DNA detection. Additionally, the use of nanocoating materials with strong antibacterial properties and good compatibility may offer effective protection for stone heritage sites in these climates. Therefore, dynamically managing biofilms based on environmental conditions and adopting customized protection strategies are considered critical for future cultural heritage protection.

It is important to evaluate the metabolic processes that may affect the stone heritage to assess the impact of bacterial communities on stone deterioration. According to the calculation results of biofilm protection rate, it was found that the bacterial community may tend to protect in the tropical environment. However, even so, the bacterial community that inhabits the surface of the rock matrix retains the possibility of a negative impact on the stone. This study attempts to quantify the dual role of biofilms by predicting the metabolic capacity of bacterial communities. This method enables us to determine whether biofilms are more inclined to protect or deteriorate. However, due to the complexity of the natural environment and microbial communities, specific bioprotection or biodeterioration effects need to be further studied.

This study investigates the impacts of climatic environments on microbial communities in stone cultural heritage, with a focus on the structural and functional differences across climatic zones. The composition of bacterial communities in temperate, subtropical and tropical regions was analyzed using 16S rRNA data from stone surfaces of East Asian and South Asian cultural heritage sites. The findings demonstrate variations in microbial diversity driven by climatic factors such as temperature and precipitation.

Additionally, this study predicted functional genes associated with biofilm protective and degradative metabolic processes, and employed the relative biofilm protective rate to assess the relationship between biodeterioration and bioprotection at cultural heritage sites under different climatic environments. As the climate shifts from temperate to tropical zones, the role of biofilms shifts towards the protective function for cultural heritage. This transition may be modulated by climate-driven variations in functional genes linked to assimilatory nitrate reduction (napAB, narGHI, nrfAH, and nirBD) and assimilatory sulfate reduction (cysJI and sir). The expression of these pathway-related genes inhibits the accumulation of soluble salts and bioacids, thereby generating protective effects.

This study provides an exploration to assess the dual role of stone heritage biofilms. Although DNA-based studies mainly reveal potential rather than actual functional capabilities, our data provide valuable insights into the metabolism of bacterial communities on stone heritage. It forms the foundation for the development of sustainable preservation strategies that account for the interaction between microbial activity and climatic conditions.

Bioinformatic data of cultural heritage

In this study, high-throughput sequencing data of 91 samples from 10 stone cultural heritage sites in Asia (mainly East Asia and South Asia) were compiled and analyzed (Fig. 6). Each sampling point is from an approximately open environment. Each of heritage site has a unique climate: Beishiku Grottoes (Sample: BSK1-9), Tiantishan Grottoes (Sample: TT1-3), and Maijishan Grottoes (Sample: MJ1-5) in China have temperate climates; Dazu Rock Carvings (Sample: DZ1-14), Leshan Giant Buddha (Sample: LS1-10), Feilai Peak in Hangzhou (Sample: XH1-20), and Leizhou Stone Dog sculptures (Sample: LZ1-5) in China have subtropical climates; the Preah Vihear Temple (Sample: BWX1-8), the Royal Palace of Angkor Thom in Cambodia (Sample: WG1-7), and the Historic Stone Ruins of Tamil Nadu (Sample: India1-10) in India have tropical climates. These sampling points belonging to the ‘Grottoes’ are close to the outdoor environment, so their environmental conditions can be considered similar to those of the open-air sites. Among them, the samples from Dazu were collected and tested by our team, while the data from other heritage sites were from NCBI database. Detailed sample information is shown in Supplementary Table 4.

Water and temperature play an important role in the colonization and deterioration of microbial communities^25,26. Appropriate levels of water and temperature contribute to the growth and metabolism of the microbial community. Therefore, 21 bioclimatic variables related to temperature and precipitation, with a resolution of 2.5 km, were obtained from the WorldClim database (www.worldclim.org), following references from other literature²⁴. Temperature-related variables included the mean annual temperature (MAT), temperature seasonality (TSEA), mean diurnal temperature range (MDTR), isothermality (ISO), maximum temperature of the warmest month (MAXTWM), minimum temperature of the coldest month (MINTCM), annual temperature range (TRANGE), mean temperature of the wettest quarter (TWETQ), mean temperature of the driest quarter (TDQ), mean temperature of the warmest quarter (TWARQ), and mean temperature of the coldest quarter (TCQ). The precipitation-related variables include mean annual precipitation (MAP), precipitation of the wettest month (PWETM), precipitation of the driest month (PDM), precipitation seasonality (PSEA), precipitation of the wettest quarter (PWETQ), precipitation of the driest quarter (PDQ), precipitation of the warmest quarter (PWARQ), and precipitation of the coldest quarter (PCQ). The human influence index (HII) and altitude (ALT) were also included. Detailed environmental information is provided in Supplementary Table 5.

The 14 samples from Dazu were taken from the statues of Dazu Rock Carvings in Chongqing, China. Biofilm samples containing a small amount of weathered sandstone were carefully collected from the biofilm-sandstone interface using a sterile surgical knife. Each sample weighed approximately 2 g and was stored in a sterile centrifuge tubes. Following the manufacturer’s instructions, DNA was extracted from all samples using the YH-Soil FastPure soil DNA extraction kit (product number: T09-96; mJYH Biotech, Shanghai, China). The quantity and quality of the extracted DNA was assessed using a Nanodrop instrument, and the extraction quality was verified through 1.2 % agarose gel electrophoresis. PCR amplification was conducted using universal primers: 338 F (5′-ACTCCTACGGGAGGCAG-3′) and 806 R (5′-GGACTACHVGGGTWTCTAAT-3′) to target the V3–V4 region of the 16S rRNA gene⁵⁸. The PCR product was extracted from a 2% agarose gel and purified using the gel extraction kit (AXYGEN Co., China), following the manufacturer’s instructions. The purified amplicons were quantified using Qubit 4.0 (Thermo Fisher Scientific, USA). Amplicons were then pooled in equimolar amounts and paired-end sequenced on an Illumina MiSeq PE300 platform (Illumina, USA), according to standard protocols provided by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China).

High-throughput amplicon sequencing analysis

The raw amplicon sequencing dataset was primarily analyzed using the open-source microbiome ecological quantitative insight tool, QIIME. Fastp software was used to remove primer sequences, after which Trimmomatic was used to filter out low-quality sequences with length <150 bases or a 20-base wide moving window average quality score <20⁵⁹. The original paired-ended sequencing data from each study were first combined using FLASH⁶⁰. Subsequently, the USEARCH software (v11)⁶¹ was applied with the ‘fastx_uniques’ and ‘unoise3’ functions (default parameters) for dereplication and denoising (error correction) of the sequence.

Since the data set contains sequencing sequences from different regions of the 16S rRNA gene, analysis at the single nucleotide difference level, such as zero-radius operational taxonomic units (zOTU), was not possible and the sequences after quality control cannot be directly compared. Therefore, these fragments were aligned to the full-length sequence of the 16S rRNA gene (SILVA 138 database) using a closed reference process, with the ‘closed_ref ‘function to set a 97 % identity threshold⁶². Sequences that did not match any entries in the database were assigned to the “unclassified” or “norank” categories. The matched full-length sequences and their annotations were used as representative sequences and taxonomic classifications for subsequent studies. However, the use of different primers for different taxa introduces an objective bias in detection efficiency, which may affect the detection results of diversity, relative abundance, or specific groups.”

Statistical analysis

Alpha diversity analysis is used to investigate the diversity of species in a localized, homogeneous habitat and is an important component of microbiome diversity studies. This approach included measures of species richness (Chao and Ace) and diversity (Shannon and Simpson). In this study, one-way analysis of variance was used to compare significant differences in microbial diversity metrics between different groups. NMDS is an indirect gradient analysis method based on dissimilarity or distance matrices and is commonly used in microbiome research to display community beta diversity. This was performed using the “Vega” package in R. Spearman correlation between species and environmental factors was calculated using the “psych” package in R. The importance ranking of environmental factors was conducted using the random forest algorithm, with analysis completed in Python. Spearman correlation coefficients between OTUs were calculated using the “psych” package in R to obtain a correlation matrix. OTUs with correlation coefficients (r) greater than |0.9| and P-values less than 0.01 were retained, and a co-occurrence network were generated. The network was imported into Gephi for visualization using the Fruchterman-Reingold layout, and the nodes were colored according to the modules and phyla. The topological parameters of the network, including average node degree, clustering coefficient, average path length, and modularity, were calculated.

Microbial community functional prediction

The metagenomic function was predicted using PICRUSt2 (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) based on OTU representative sequences⁶³. PICRUSt2 is a software suites that includes a series of tools, including HMMER, which aligns OTU representative sequences with reference sequences. While the microbiome functions predicted by the PICRUSt2 should be considered as potential, it can still provide important insights for research⁶⁴. Functional genes associated with carbon fixation, and nitrogen and sulfur metabolism were identified using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Based on the current literature^18,65,66, functional genes related to rock interaction were selected, including those associated with metal uptake/ resistance, oxalate biosynthesis, biomineralization, inorganic P dissolution, and organic P mineralization. The metabolic processes potentially involved in the negative and positive effects on the stone heritage are primarily summarized in Table 1. A complete list of the surveyed genes is provided in Supplementary Table 2.

The “biofilm relative bioprotective ratio” is defined as the ratio of the sum of the biofilm-protective effects to the sum of the biodeterioration effects. A value greater than 1 indicated that the biofilm had a protective effect. The following formula was used for the calculations:

where ({sum }_{i=1}^{n}{x}_{i}) represents the sum of the relative abundances of genes with positive effects in the biofilm, and ({sum }_{j=1}^{m}{y}_{j}) represents the sum of the relative abundances of genes with negative effects in the biofilm; xi and yi denote the relative abundances of individual genes.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.