Five years of wastewater surveillance reveal COVID-19–Associated amplification and seasonal shifts in bla_KPC and bla_NDM

Abstract

Antimicrobial resistance (AMR) has intensified in recent years, raising concerns about the potential impact of the COVID-19 pandemic on AMR dynamics. However, the influence of pandemic on the enviromental occurrence and dissemination of antibiotic-resistant genes (ARGs) remains under investigation. Wastewater-based epidemiology (WBE) provides an integrated view of community-level AMR dynamics, yet long-term assessments spanning the pandemic period remain limited. This study investigated the occurrence, temporal patterns, and seasonal behavior of two clinically relevant carbapenemase genes, bla_KPC and bla_NDM, in raw wastewater from the Attica region over a five-year period (July 2021–September 2025). A sensitive and specific 3-plex qPCR assay was developed and analytically validated for the simultaneous quantification of bla_KPC, bla_NDM, and 16S rRNA, the latter serving as a stable bacterial load indicator.

Across the monitoring period, bla_KPC was consistently detected, showing a steady increase and reaching its highest concentrations in 2023, while bla_NDM exhibited intermittent but pronounced surges, particularly in 2023. Comparison with community SARS-CoV-2 viral loads suggest temporal association with AMR signatures in wastewater. Seasonal analysis showed non-uniform patterns, with atypical warm-season peaks of both ARGs in 2023–2024 and a return to classical autumn–winter peaks in 2025, suggesting a gradual restoration of pre-pandemic seasonality.

To our knowledge, this is among the first long-term WBE studies to characterize post-COVID-19 shifts in ARG occurrence and seasonality. The results demonstrate that the pandemic period was associated with elevated and fluctuating ARG levels in the urban wastewater system, emphasizing the need for sustained AMR monitoring. These findings address critical knowledge gaps in environmental AMR surveillance and support the integration of WBE into public health strategies aimed at mitigating AMR dissemination.

Graphical abstract

Highlights

• •Wastewater surveillanceenables monitoring of community-level AMR dynamics.
• •Five-year monitoring assessed seasonal patterns of two prevalent ARGs.
• •bla_KPC and bla_NDM increased markedly in 2023, then declined afterward.
• •SARS-CoV-2 and both ARGs showed opposite community trends.
• •ARGs peaked in warm seasons in 2023 and 2024, reversing in 2025.

1. Introduction

Antimicrobial resistance (AMR) is recognized as a critical global health threat, undermining healthcare systems and turning previously manageable infections into life-threatening conditions [1]. Antibiotic- resistant bacteria (ARB) and antibiotic-resistant genes (ARGs) occur naturally in the environment as part of bacterial defense mechanisms, even in areas with minimal human influence [2]. However, human-driven stressors -including chemical pollution from heavy metals and microplastics [3], [4], [5], [6], climate change, and other socioeconomic factors [7] further promote their spread. Lately, the potential impact of the COVID-19 on AMR has been increasingly investigated. Extensive antibiotic overuse, often without confirmed bacterial co-infections, was observed during the pandemic, with overprescription varying between high and low-income countries [8], [9]. However, the effect of the COVID-19 pandemic on AMR appears to be geographically heterogenous, with several regions reporting increases in specific resistant organisms [10], [11], [12], while other observed stable or even declining AMR trends [11], [13], emphasizing the importance of region-specific surveillance. In Greece, findings from a tertiary care hospital in Athens demonstrated a significant post-pandemic increase in carbapenemase-producing Enterobacterales (CPEs) and an altered carbapenemase epidemiology, with increased antimicrobial resistance rates to multiple antibiotics in 2022 and 2023 compared to their corresponding with 2020 [14]. This local epidemiological context, together with increased hospitalization pressure, ICU burden, broad-spectrum antibiotic use, and potential disruption of infection prevention and control practices during the pandemic, provides a rationale for investigating COVID-19-associated changes in carbapenemase ARGs in wastewater from the Attica region.

A key mechanism is the endogenous or acquired production of β-lactamases, which inactivate β-lactam antibiotics [15]. Carbapenems serve as last-line drugs against multidrug-resistant (MDR) Gram-negative infections; therefore, carbapenemases -transmitted mainly via horizontal gene transfer- pose a major clinical challenge, especially in intensive care units (ICUs) [16], [17]. The most clinically significant carbapenemases include Klebsiella Pneumoniae carbapenemases (KPC), New Delhi metallo-β-lactamase (NDM), and oxacillinase-48 (OXA-48)-like enzymes ([18]. Ιn Greece, both KPC- and NDM- producing strains have long been endemic, with KPC predominating [14].

ARGs spread rapidly across healthcare, environmental, animal and food systems, underscoring the need for coordinated “One Health” approach to AMR mitigation (WHO,2024). Urban and hospital wastewater are major environmental reservoirs of ARBs and ARGs [19], [20]. Wastewater treatment plants (WWTPs)- with high microbial densities and conditions that promote gene exchange- serve as critical interfaces facilitating ARG selection, transfer and release into the environment [20]. Wastewater-based epidemiology (WBE) has emerged as a valuable tool for monitoring community health and is increasingly used to assess pathogen prevalence and transmission dynamics [21], [22], [23], [24], [25]. WBE is also considered a promising complement to current AMR surveillance strategies, offering near real-time population-level data [26]. Previous studies have explored ARG occurrence, seasonal trends, correlation with antibiotic levels and their persistence across WWTP treatment stages [27], [28], [29], [30], [31], [32], [33], [34]. Limited work has addressed links between wastewater ARGs and clinical isolates [35], [36].

Despite growing interest, ARG surveillance in environmental matrices remains methodologically challenging due to the absence of standardized analytical methods [37]. Moreover, few studies have assessed how the COVID-19 pandemic influenced ARG distribution, and most seasonal analyses rely on pre-pandemic samples, leaving current spatiotemporal trends poorly characterized.

In this context, the present study aimed to evaluate the impact of COVID-19 pandemic on the occurrence and seasonal patterns of two clinically relevant carbapenemase genes, bla_KPC and bla_NDM in urban wastewater samples from the city of Attica. Target gene selection was based on their long-term clinical occurence in Greece. To our knowledge, this is among the first studies to use wastewater surveillance to examine post-COVID-19 shifts in ARG seasonality. More importantly, this work is integrated within the “One Health” framework, utilizing WBE as a holistic, non-invasive tool to monitor high-priority clinical targets within the urban aquatic environment. Urban wastewater represents a critical interface that bridges human epidemiological trends with environmental resistance signatures.

2. Materials & methods

2.1. Wastewater sampling and processing

24-h composite, flow proportional raw wastewater samples were collected from the Wastewater Treatment Plant (WWTP) of Attica, located on the island of Psyttaleia. The facility includes primary sedimentation, activated sludge with biological nitrogen and phosphorus removal, and secondary sedimentation [38] and serves a large proportion of the population in the Athens metropolitan area and its suburbs.

A total of 47 raw wastewater samples were collected monthly from July 2021 to September 2025 and analyzed for two prevalent ARGs (bla_KPC and bla_NDM) and the universally conserved bacterial marker 16 s rRNA. This sampling period encompasses the peak of the COVID-19 pandemic, the implementation of associated restriction measures and their gradual relaxation in the post-COVID-19 period. Samples were collected in pre-cleaned 1 L high-density polyethylene bottles, transported to the laboratory at 4 °C, and processed immediately upon arrival.

Wastewater concentration and nucleic acid extraction were performed using the “Wizard Enviro Total Nucleic Acid” Kit (Promega, United States) following the manufacturer's instructions. Briefly, 500 μL of protease solution was added to 40 mL wastewater samples, followed by incubation for 30 min at room temperature to facilitate protein digestion. Samples were then centrifuged at 3000 xg for 10 min to separate particulate matter, and the resulting supernatant was transferred into two clean sterile centrifuge tubes. Wastewater concentration was subsequently performed according to the Wizard® Enviro TNA kit protocol using PureYieldTM Binding Columns to capture and concentrate nucleic acids from the sample matrix. Total Nucleic acid extraction step was then carried out to obtain the TNA eluates, from which DNA was used for downstream analysis.

2.2. qPCR analysis

A 3-plex qPCR protocol was developed and optimized for the detection and quantification of the three genes. The oligonucleotide primers and probe sets (Integrated DNA Technologies, USA) were based on guidelines from Centers for Disease Control and Prevention (CDC) with slight modifications and target evolutionarily conserved regions of the genome (Table S1). The optimized PCR reaction mix consisted of the following: 4.0 μL of PCR buffer (5×), 2 μL of Mg²⁺ (25 mM), 0.4 μL dNTPs (10 mM), 0.5 μL of BSA (10 μg/ μl), 0.6 Ll of primer mix (10μΜ each), 1 μL of probe mix (10μΜ each), 0.3 μL of Taq DNA polymerase (Promega, USA) (5 units/μl), 2 μL of DNA template and nuclease- free H₂O at a final volume of 20 μL. The Real-time qPCR assay was performed in the Touch CFX96™ Real-Time PCR Detection System (Bio- Rad, USA) (92 °C/2 min, 40 cycles of 95 °C/30s, 60 °C/30s and 72 °C/30s). Negative and positive controls were used in each run to evaluate the performance of the assay.

Three synthetic DNA oligonucleotides were designed and used for the analytical validation of the assay (gBlocks; Integrated DNA Technologies, USA; Table S2). Each synthetic control contained the target sequences of bla_KPC (126 bp), bla_NDM (125 bp) and 16 s rRNA (200 bp), including all primer and probe binding sites. The synthetic oligos were reconstituted with Tris−EDTA (TE) buffer to a stock concentration of 10 ng/μL. 10-fold dilutions of the synthetic controls were prepared to generate standard curves for the three targets, assess the analytical performance of the assay and quantify sample copies numbers per reaction using Eq. (1).

$$\frac{\mathit{copies}}{\mathit{reaction}}={10}^{\left(\frac{\mathit{Cq}-\mathit{intercept}}{\mathit{slope}}\right)}$$ (1)

2.3. Development and analytical validation of the 3-plex qPCR assay

A specific and sensitive 3-plex qPCR assay was developed and optimized for the detection and quantification of bla_KPC, bla_NDM and 16S rRNA genes. Prior to application to raw wastewater samples, extensive optimization experiments were preformed and analytical performance of the assay was validated. In total, 47 samples met the required quality criteria and were included in the analysis.

2.3.1. Evaluation of polymerase options to reduce 16S rRNA background amplification

During assay development, a positive amplification signal for the 16S rRNA target was observed even in the absence of template DNA. This phenomenon is well documented as the ubiquity of bacteria in the environment means that laboratory reagents -particularly commercially available PCR enzymes- may contain trace amounts of bacterial DNA or may be insufficiently purified, leading to false-positive amplification signals [39], [40].

To address this issue, an alternative enzyme, GoTaq® DNA Polymerase (Promega, USA), was evaluated and shown to be more suitable for bacterial DNA detection under the conditions of this protocol. According to CDC guidelines, amplification of the 16S rRNA target in no-template control after 30 cycles was considered acceptable, whereas amplification before 30 cycles indicated potential contamination and invalidated the run [41] (Fig. S1).

2.3.2. Analytical validation

Synthetic controls containing known copy numbers of each target gene were combined at equal concentrations (10⁵ copies/μl), and subsequently subjected to a series of 10-fold serial dilutions to generate four concentration levels (10⁴, 10³, 10² and 10 copies/μl). These dilutions series were employed to characterize the assay's linear dynamic range (LDR), limit of detection (LOD) and limit of quantification (LOQ). All reactions were performed in triplicate, with the lowest concentration tested ten independent replicates to robustly estimate the LOD and LOQ (Table S3).

2.3.3. Analytical sensitivity

The analytical sensitivity was evaluated by estimating the LOD using quantification calibrators containing a known number of copies/μl prepared as previously mentioned. For each gene target, a calibration curve was generated using serial dilutions of external standards in triplicate for each concentration, ranging from 10⁵ to 10 copies/ μL. The LOD and LOQ of the assay were determined at 5 copies/μl and 15 copies/μl for all the three genes. After back-calculation analysis, the LOD and LOQ were determined at 2500 copies/L and 8250 copies/L, respectively, based on calculations described elsewhere [42]. The linear regression plots (Cq values versus log₁₀ of dilution factor) demonstrated excellent linearity, with correlation coefficients (R²) of 0.9998, 0.9999, and 0.9995 for bla_KPC, bla_NDM and 16 s rRNA, respectively, confirming a precise linear relationship (Fig. S2).

The newly developed 3-plex assay was additionally compared with its corresponding singleplex assays to investigate potential differences in Cq values between the methods. A concentration of 10,000 copies/μl of each target was analyzed using the 3-plex assay in the presence of all primer and probe sets, as well as in the corresponding singleplex assays containing only the target-specific primer and probe set. The comparable amplification performance that was observed between the singleplex and multiplex protocols based on Cq values is presented in Table S4.

2.3.4. Specificity of the assay

The analytical specificity of the assay was evaluated by amplifying each target individually. Three synthetic controls, each corresponding to one of the target genes, were used at a concentration of 10⁴ copies/μl. No nonspecific amplification or cross-reactivity was observed among the three primer-probe sets. Each target yielded a positive signal exclusively in its designated fluorescence detection channel, with no signal was detectable signal in the remaining channels. These findings confirm that the assay demonstrates high analytical specificity and reliably discriminates among the individual ARG targets.

2.3.5. Statistical analysis

A Cross-Correlation Function (CCF) analysis was performed to evaluate the temporal relationships and identify potential time lags between SARS-CoV-2 loads and ARG concentrations in wastewater. Specifically, SARS-CoV-2 loads were cross-correlated against the abundance of blaKPC and blaNDM over a range of ±10 monthly lags. This time-series analysis was utilized to identify the specific lag periods at which the correlation coefficients (r) peaked and exceeded the upper and lower 95% confidence limits. The statistical analysis was conducted using IBM SPSS Statistics (Version 26.0; IBM Corp., Armonk, NY, USA), and the threshold for statistical significance was set at p < 0.05.

2.4. Quantification of ARGs concentration in wastewater

The population served by the WWTP was estimated in real time for each sampling day using concentrations of the physicochemical parameter NH₄-N, along with the corresponding daily flow rates. The number of inhabitants was estimated daily based on the population equivalent (PE) of 8.1 ± 0.37 g of NH₄-N/day/ individual [43], [44]. Genome copies/L of each virus were back-calculated, and subsequently normalized to the estimated population and daily flow rates, following established methodologies [21], [42].

3. Results & discussion

3.1. Temporal dynamics of ARGs during and after the COVID-19 pandemic in wastewater

The yearly distribution of bla_KPC, bla_NDM, and 16S rRNA gene copies revealed distinct temporal trends across the 2021–2025 monitoring period (Table 1). The bla_KPC gene was consistently detected throughout the study period, with concentrations increasing steadily and peaking in 2023, while blaNDM showed intermittent but pronounced surges, particularly in 2023. The 16S rRNA marker remained stable across all years, confirming uniform bacterial loads and extraction efficiency. Absolute copies/L and gene load normalized to 16 s rRNA (relative abundance) of the targets are provided in the Supplementary Material (Table S5).

Table 1.

Average copy numbers and detection rate of the three targets during the study period.

Year	blaKPC		blaNDM		16 s rRNA
	Average (copies/100 k)	Detection rate (%)	Average (copies/100 k)	Detection rate (%)	Average (copies/100 k)	Detection rate (%)
2021	3.1 × 1011	66.7	–	0	9.8 × 1015	100
2022	4.8 × 1011	37.5	1.5 × 1011	25	1.0 × 1016	100
2023	4.2 × 1012	100	8.0 × 1010	91.7	4.3 × 1015	100
2024	6.2 × 1011	58.3	4.1 × 1010	8.3	5.60 × 1015	100
2025	6.9 × 1011	88.9	6.5 × 1010	66.7	6.1 × 1015	100

A temporal comparison of SARS-CoV-2 concentrations with bla_KPC and bla_NDM abundances revealed a consistent pattern in which ARG increases followed major SARS-CoV-2 peaks throughout the 2021–2025 monitoring period (Fig. 1). Across multiple waves, the wastewater data displayed clear temporal alignment between SARS-CoV-2 surges and subsequent rises in carbapenemase genes. For bla_KPC, each substantial SARS-CoV-2 peak—most prominently in late 2021, early 2022, and winter 2022–2023—was followed by elevated bla_KPC concentrations in the ensuing months. This lagged pattern was especially pronounced in 2023, during which bla_KPC reached its highest concentrations yearly average shortly after one of the strongest viral peaks recorded in the study period. bla_KPC levels remained elevated into 2024 and increased further in 2025, although they did not reach the levels observed in 2023 (Fig. 1).

Fig. 1.

Comparison of SARS-CoV-2 and bla_KPC trends during the study period, expressed in copies/100 k inhabitants.

The bla_NDM gene showed a similar temporal pattern, though with more variability due to is absence in early years. After first appearing in 2022, bla_NDM levels rose markedly from late 2022 through mid-2023, again following periods of heightened SARS-CoV-2 viral load. This alignment was evident both in the increase in detection frequency (from 25% in 2022 to 91.7% in 2023) and in the rise in average concentrations during the same period. Although blaNDM declined in 2024, levels increased again in 2025, matching the broader post-peak elevation pattern also observed for blaKPC (Fig. 2).

Fig. 2.

Comparison of a) SARS-CoV-2 and bla_NDM trends during the study period, expressed in copies/100 k inhabitants.

A Cross Correlation Function (CCF) analysis was performed between SARS-CoV-2 viral loads and bla_KPC and bla_NDM abundance, to statistically evaluate the temporal trends and identify potential time lags. The analysis revealed a moderate to strong, statistically significant positive correlation at a 9-month lag (r = 0.621, p < 0.05) for bla_KPC and a 7-month lag (r = 0.681, p < 0.05) for bla_NDM, where the correlation coefficients clearly exceeded the upper 95% confidence limit (Fig. S3). No significant immediate (Lag 0) or short-term correlations were observed. This cross-correlation at lag 9 and lag 7 statistically substantiates an environmental response in ARGs concentrations, demonstrating that the major peaks in ARG abundance in wastewater (Fig. 1, Fig. 2) are coupled with the high SARS-CoV-2 viral loads quantified in wastewater approximately 9 and 7 months prior, for bla_KPC and bla_NDM, respectively. However, this observation should be further explored across different settings and regions.

These findings agree with previous reports showing higher AMR rates in the post-pandemic period compared to pre-pandemic levels [12], [45]. The extensive use of broad-spectrum antibiotics, including carbapenems, during the COVID-19 pandemic -often without confirmed bacterial co-infections [46] likely contributed to shifts in circulating carbapenemase types, particularly KPC and NDM.

Following the relaxation of restriction measures, AMR dissemination increased again and stabilized at elevated levels (Table 1). In Greece, a marked rise in carbapenemase-producing Enterobacterales was documented during the pandemic, with KPC remaining dominant and a transition from VIM to NDM observed in 2022 and 2023 [14]. After the temporary suppression of AMR transmission during lockdowns, a resurgence was likely driven by the resumption of routine hospital activities and the restoration of international travel networks [47]. The higher antimicrobial resistance rates observed in 2023 could be attributed to the overall burden on the local healthcare system, the misuse of antibiotics, and the inadequate implementation of infection-control measures during the unprecedented pressure experienced by healthcare facilities throughout the peak phase of the pandemic in 2021 and 2022. These epidemiological trends align closely with the present wastewater surveillance data, which show a clear rise in bla_KPC levels beginning in 2022, further supporting the post-pandemic amplification of AMR in the community.

3.2. Year-specific seasonal signatures and post-COVID-19 shifts in ARG levels in wastewater

Seasonal variation in ARG abundance did not follow a uniform pattern across the study period, indicating that seasonal effects were modulated by year-specific epidemiological and environmental conditions. In 2023 and 2024, both bla_KPC and bla_NDM exhibited their highest concentrations during the spring and summer months (Fig. 3a and b), representing an atypical warm-season elevation. In contrast, 2025 showed a shift toward autumn–winter peaks, a pattern more consistent with classical seasonal trends associated with colder periods [48], [49]. These contrasting patterns highlight that while seasonality influences ARG dynamics, its expression varies substantially across years and does not always align with traditional expectations.

Fig. 3.

a) bla_KPC and b) bla_NDM trends per year in the post-pandemic period.

Published studies similarly report inconsistent seasonal signature. For example, Keer et al. observed highest bacterial diversity in winter but greater 16S rRNA gene abundance in summer [50], while Honda et al. reported seasonal differences in both ARG abundance and composition, with notable increase during summer months [51]. Other studies documented elevated ARG levels either in warmer seasons [50], [52] or in colder periods [48], [49], [53], [54] whereas some reported only minor seasonal variation [55]. Overall, the literature confirms that the seasonality of ARG occurrence remains ambiguous, with diverse and sometimes contradictory trends.

The present study contributes to this body of evidence by showing a post-pandemic shift in the seasonal distribution of bla_KPC and bla_NDM in Attica wastewater with pronounced peaks during the warmer months of 2023–2024. By 2025, the seasonal pattern resembled pre-pandemic observations, suggesting a gradual re-establishment of typical seasonal behavior following the disruptions of COVID-19 period. For the year-specific seasonal analysis, only years with sufficient monthly/seasonal representation and quantifiable ARG measurements were considered. Samples from 2021 and 2022 were included in the overall longitudinal and annual analyses but were not used for formal seasonal comparison because of incomplete seasonal coverage and the presence of values below the LOD, which could bias seasonal grouping.

4. Limitations of the study

Although the present study provides useful longitudinal information on the occurrence and temporal dynamics of blaKPC and blaNDM in wastewater from the Attica region, several limitations should be acknowledged. First, although the study is positioned within the One Health framework, it should be interpreted as a One Health–relevant environmental surveillance study rather than a comprehensive One Health investigation, since animal, food-chain, agricultural, downstream environmental, antibiotic consumption, hospital activity, and clinical isolate data were not directly included. Second, the monthly sampling frequency and the total number of analyzed samples may have limited the detection of short-term fluctuations, transient ARG peaks, and high-resolution seasonal patterns; therefore, the observed seasonal trends and temporal associations should be interpreted cautiously as indicative and hypothesis-generating. Future studies should include higher-frequency and spatially expanded wastewater sampling, additional upstream or sentinel sites, integration with clinical AMR surveillance and antibiotic-use data, hospital activity indicators, animal/agricultural sources, and downstream environmental monitoring to better characterize the drivers and transmission pathways of ARGs across the One Health continuum.

5. Conclusions

This study investigated the impact of COVID-19 on the occurrence and seasonal patterns of two highly prevalent ARGs in Greece -bla_KPC and bla_NDM- though a five-year wastewater monitoring campaign (2021–2025). Our findings revealed remarkably low levels of both genes in 2021 and 2022, the peak years of the COVID-19 pandemic, when strict non-pharmaceutical interventions were implemented to limit the disease transmission. In 2023, following the relaxation of these measures, both genes showed a pronounced increase, characterized by high concentrations and high annual detection frequencies, followed by a gradual decline in subsequent years. bla_NDM generally appeared at lower levels and with more variable detection rates, while bla_KPC remained consistently prevalent through the study period.

A notable shift in seasonality was observed in 2023 and 2024, the years immediately following the pandemic peaks, during which both genes reached their highest levels in spring and summer, contrasting with patterns documented in pre-pandemic years. By 2025, concentrations had decreased, and seasonal trends resembled those of the pre-pandemic period, with higher levels detected during winter. Overall, our results indicate that the COVID-19 pandemic influenced both the dissemination and seasonal dynamics of these two major ARGs, further contributing to the spread of AMR in Greece. Statistical analysis confirmed a distinct time-lagged relationship, revealing a significant positive correlation at a 9-month lag (r = 0.621, p < 0.05) and a 7-month lag (r = 0 0.681, p < 0.05) for bla_KPC and bla_NDM, respectively. These findings confirm that the delayed increase of ARG concentrations in wastewater is significantly coupled with the preceding SARS-CoV-2 peaks. However, due to the complex factors affecting the dissemination of ARGs in communities, future studies are needed to elucidate the role of the COVID-19 pandemic on AMR shifts and to confirm the time lag periods between high SARS-CoV-2 loads and ARG elevated levels.

This study highlights the value of wastewater-based epidemiology not only for pathogen surveillance, but also for monitoring antibiotic-resistant agents, providing timely insights into the post-pandemic shifts in ARG behavior. From a “One Health” perspective, urban wastewater serves as a critical interface between human clinical pressures and the environment, acting as a repository for resistance genes shed by human populations. Our findings highlight that major public health threats, such as the COVID-19 pandemic, leave a distinct signature on the environmental resistome in Greece. Integrating long-term WBE data into global AMR surveillance frameworks aligns with the primary objectives of the “One Health” approach, demonstrating how environmental monitoring can reflect community-level epidemiological trends and inform mitigation strategies to limit the dissemination of critical threats like carbapenemase-producing organisms.

CRediT authorship contribution statement

Anastasia Zafeiridou: Writing – original draft, Validation, Methodology. Foteini Pitaouli: Methodology. Georgia Georgakopoulou: Methodology. Nikolaos Thomaidis: Visualization, Resources, Project administration, Conceptualization. Athina Markou: Writing – original draft, Visualization, Validation, Supervision, Project administration, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.