Engineering an asymmetric cross-feeding consortium in Ogataea polymorpha driven by sulfur metabolism Authors: Yuxia Fan, Xiaoxin Zhai, Yongjin J Zhou, Jiaoqi Gao Published: 2026 ABSTRACT AbstractMicrobial cell factories offer a sustainable route for chemical production, yet reconstructing complex pathways in a single host often leads to metabolic burden and imbalanced flux. Here we establish an auxotrophy-based cross-feeding strategy in the non-conventional yeast Ogataea polymorpha. Through systematic screening of pairwise auxotrophic mutants, we identify several complementary combinations, among which the met10Δ-str3Δ pair exhibits robust and reproducible growth restoration under both glucose and methanol conditions. Mechanistically, this pair displays asymmetric metabolic dependency. met10Δ can utilize multiple sulfur-containing intermediates, whereas str3Δ relies exclusively on homocysteine. Population dynamics further reveal that the initial inoculation ratio critically determines community establishment and final biomass, with all co-cultures converging toward met10Δ enrichment over time. This work provides an experimental foundation for designing stable synthetic consortia in non-conventional yeasts.Keywords: Synthetic consortium, Cross-feeding, Ogataea polymorpha, Auxotrophy, Microbial cell factory Introduction Microbial cell factories have emerged as a promising platform for the sustainable production of chemicals, fuels, and pharmaceuticals, offering advantages over traditional chemical synthesis in terms of renewability and environmental friendliness [1,2]. Through metabolic engineering and synthetic biology, microorganisms can be reprogrammed to convert renewable feedstocks into value-added products [[3], [4], [5]]. However, when complex biosynthetic pathways, particularly those for natural products, are reconstructed in a single microbial host, several challenges commonly arise. These pathways are often lengthy, involve multiple enzymatic steps, require balanced precursor supply and cofactor regeneration, and are subject to intricate regulatory networks [[6], [7], [8], [9]]. Reconstituting such pathways in a single strain frequently leads to imbalanced metabolic flux, accumulation of toxic intermediates, and increased metabolic burden, which collectively impair cell growth and limit product titers [10]. To overcome these limitations, synthetic microbial consortia have gained increasing attention as an alternative strategy [[10], [11], [12], [13], [14], [15]]. By partitioning a complex biosynthetic pathway among multiple specialized strains, this division-of-labor approach reduces the metabolic burden on any single organism, alleviates pathway bottlenecks, and can improve overall productivity [11,16]. In such engineered communities, different members perform distinct tasks while interacting through metabolic exchange, signaling, or physical associations. Nevertheless, the practical utility of such consortia depends critically on their stability [17,18]. Differences in growth rates, competition for resources, and the emergence of non-cooperative variants can all lead to compositional drift and the eventual collapse of the intended division of labor [[19], [20], [21]]. Several strategies have been developed to stabilize engineered consortia, including spatial or nutritional niche segregation, signaling-based regulation, and auxotrophy-based cross-feeding [22,23]. Among the various strategies for stabilizing synthetic consortia, auxotrophy-based cross-feeding has proven particularly effective [21,24]. In this design, each strain is engineered to require an essential metabolite that its partner produces, creating reciprocal dependencies that can maintain community composition without continuous external selective pressure [25,26]. This approach has been successfully demonstrated in model organisms such as Saccharomyces cerevisiae and Escherichia coli, where it has enabled stable coexistence and, in some cases, enhanced production of target compounds [10,15,[27], [28], [29]]. Despite these advances, the application of synthetic consortia in non-conventional yeasts remains largely unexplored, although many of these organisms possess attractive traits for industrial biotechnology [30]. Ogataea polymorpha (formerly Hansenula polymorpha) is a thermotolerant methylotrophic yeast with several advantageous features [31]: it can utilize a wide range of carbon sources including methanol and xylose, exhibits strong protein expression capacity, and has a tightly regulated peroxisomal system for methanol metabolism [32]. In recent years, genome sequencing, omics resources, and efficient gene-editing tools have been developed for O. polymorpha [[33], [34], [35], [36], [37]], establishing it as a promising next-generation chassis for green biomanufacturing. However, systematic studies on the design principles governing mutualistic consortia in O. polymorpha have yet to be conducted. In this study, we established an experimental framework for constructing auxotrophy-based synthetic consortia in the non-conventional yeast O. polymorpha (Figure 1). By generating a series of auxotrophic mutants and systematically screening pairwise combinations, we identified several pairs capable of robust cross-feeding, among which the met10Δ-str3Δ pair (both defective in sulfur assimilation) exhibited the most reproducible complementation under both glucose and methanol conditions. Mechanistic analysis revealed an asymmetric metabolic dependency, which refers to the unequal ability of the two mutants to utilize sulfur-containing intermediates. Population dynamics further demonstrated that all communities converging toward met10Δ enrichment over time regardless of the different initial inoculation ratios. This work provides a foundation for designing stable synthetic consortia in non-conventional yeasts and advances the potential of distributed biosynthesis in multi-strain systems. Asymmetric cross-feeding between met10Δ and str3Δ in Ogataea polymorpha. The met10Δ and str3Δ mutants harbor deletions in the sulfur assimilation pathway at MET10 and STR3, respectively. met10Δ can utilize cysteine (Cys), cystathionine (Cysta), and homocysteine (Hcy) for growth, whereas str3Δ depends solely on Hcy. Co-culture results in asymmetric metabolite exchange: met10Δ supplies Hcy to str3Δ, while str3Δ releases multiple intermediates. Four initial inoculation ratios (1:10, 1:5, 1:1, 5:1, and 10:1 met10Δ:str3Δ) were evaluated, and all communities converge to a similar final composition enriched in met10Δ (around 80%), illustrating ratio-dependent community assembly and asymmetric dynamics. Materials and methods Strains, plasmids, and culture conditions The parental strain used in this study was O. polymorpha JQcr03, a uracil and leucine auxotroph (ura3Δ leu2Δ) previously constructed in our laboratory [34]. JQcr03L was generated by restoring LEU2, resulting in a strain with only the ura3Δ defect. The prototrophic control strain FY00 was obtained by restoring URA3 in JQcr03L. Auxotrophic single mutants were constructed in the JQcr03L background using a CRISPR/Cas9-based method [34], followed by URA3 restoration. Strains used for initial screening and phenotypic analysis included FY09–FY16 and FY20–FY22. For population dynamics studies, FY31 (str3Δ-EGFP) and FY34 (met10Δ-mCherry) were used. Detailed strain and plasmid information is provided in the Supplementary Table 1 and Supplementary Table 2. All gene deletions and fluorescent marker integrations were verified by PCR. All primers used in this study were listed in Supplementary Table 3. E. coli DH5α was used for plasmid construction and propagation and was cultured in LB medium supplemented with ampicillin at 37 °C. O. polymorpha strains were routinely cultivated in Delft minimal medium at 37 °C with shaking at 220 rpm in 100 mL flasks containing 20 mL medium. Unless otherwise specified, glucose (20 g/L) was used as the carbon source. For methanol experiments, glucose was replaced with 10 or 20 g/L methanol. For supplementation assays, sodium sulfide (Na2S), cysteine, cystathionine, or homocysteine was added at a final concentration of 20 mg/L. Complete medium compositions are provided in the Supplementary Table 4. Growth assays Glycerol stocks stored at −80 °C were streaked onto YPD agar and incubated at 37 °C. Single colonies were inoculated into liquid YPD and grown overnight (15–18 h) at 37 °C and 220 rpm. Cells were harvested by centrifugation, washed twice with Delft medium, and the OD600 of the seed culture was measured. Inoculation volumes were calculated to achieve the desired initial OD600. For screening under glucose conditions, co-cultures were initiated in Delft medium containing 20 g/L glucose at an initial total OD600 of 0.2 with a 1:1 strain ratio. Single-strain controls were inoculated at the same OD600. Growth was monitored in real time using a multi-well plate reader. Shake-flask validation was conducted under identical conditions, with manual sampling and OD600 measurement after appropriate dilution. For methanol experiments, Delft medium containing 10 or 20 g/L methanol was used as the sole carbon source. The initial total OD600 was set to 0.6. No additional methanol was supplied during cultivation. For supplementation assays, single auxotrophic strains were inoculated at an initial OD600 of 0.2 into Delft medium supplemented with the indicated metabolites, and growth was monitored until stationary phase. Population ratio analysis by fluorescence microscopy FY31 (str3Δ-EGFP) and FY34 (met10Δ-mCherry) were co-cultured in Delft medium containing 20 g/L glucose at an initial total OD600 of 0.4. Initial ratios of met10Δ to str3Δ were set at 1:10, 1:5, 1:1, 5:1, and 10:1. Samples were collected every 24 h for OD600 measurement and population analysis. For microscopy, 300 μL of culture was washed three times with PBS and resuspended to an OD600 of approximately 3-4. 5 μL of the suspension was placed on a glass slide and observed under a 40× objective. Three biological replicates were used, and three random fields were imaged per replicate. Red and green fluorescence containing cells were counted, and the fraction of FY34 was calculated as Nred/(Nred + Ngreen), where Nred and Ngreen represent the total numbers of red and green fluorescent cells, respectively. Statistical analysis Unless otherwise stated, all experiments were performed with at least three biological replicates. Data are presented as mean ± SD. Statistical analysis was performed using unpaired two-tailed Welch's t-test for comparisons between two groups and one-way ANOVA followed by Tukey's multiple-comparison test for comparisons among multiple groups. Differences were considered statistically significant at p < 0.05. Results Construction and validation of auxotrophic strains A set of auxotrophic mutants was generated in O. polymorpha by deleting genes involved in amino acid or nucleotide biosynthesis. All deletions were confirmed by PCR. As expected, the prototrophic strain FY00 grew readily in minimal glucose medium, reaching stationary phase at OD600 ≈ 9 after 13 h (Fig. S1a). In contrast, most auxotrophic strains failed to grow without exogenous supplementation, with OD600 remaining below 1.2 throughout cultivation (Fig. S1b and c). Screening for complementary auxotrophic pairs To identify pairs capable of reciprocal cross-feeding, all mutants were co-cultured in minimal glucose medium at a 1:1 ratio. Growth complementation was quantified using a “Recovery Index” (RI), defined as the ratio of the maximum OD600 of the co-culture to that of FY00 (≈9). RI ≥ 0.8 was considered strong complementation, 0.4 ≤ RI < 0.8 moderate, and RI < 0.4 weak or absent. The complete screening results are presented as a heatmap in Fig. 2a and Fig. S2, where each cell represents the RI value for a specific pairwise combination. Most pairs showed little to no complementation (RI < 0.4). However, several combinations exhibited strong or moderate complementation (Fig. S2). Notably, met10Δ-str3Δ, met14Δ-str3Δ, and met10Δ-met3Δ all achieved RI > 0.8, suggesting that these combinations are capable of robust reciprocal cross-feeding. Pairs such as ura3Δ-met10Δ and met10Δ-ade2Δ showed moderate complementation (RI 0.4–0.7). Screening and validation of auxotrophic pairs capable of cross-feeding. (a) Heatmap showing the Recovery Index (RI) for all pairwise combinations of auxotrophic strains. The RI is defined as the ratio of the maximum OD600 achieved by the co-culture to that of the prototrophic control strain FY00 (OD600 ≈ 9). Pairs with RI ≥ 0.8 were considered strongly complementary, 0.4 ≤ RI < 0.8 moderately complementary, and RI < 0.4 weakly or non-complementary. (b) Scatter plot of maximum OD600 and lag phase for selected strongly complementary pairs. Colors correspond to the same pairs as in (c) and (d). (c–d) Growth curves of strongly complementary pairs in shake-flask cultures. All co-cultures were initiated at a 1:1 ratio with total initial OD600 = 0.2 in Delft minimal medium containing 20 g/L glucose at 37 °C. Data are mean ± SD from four biological replicates. To further quantify the growth characteristics of the most promising pairs, we conducted the shake-flask validation, which confirmed the reproducibility of most top-performing pairs. Detailed growth curves for the selected complementary pairs are presented in Fig. 2c–d. Consistent with the initial screening, most pairs reached OD600 values of over 24 in a relatively stable manner, comparable to the FY00 control. In contrast, pairs such as met14Δ-str3Δ, met3Δ-str3Δ and ura3Δ-met14Δ showed high variability across replicates, and were therefore excluded from subsequent analyses. On this basis, two key parameters, maximum OD600 and lag phase, were compared among these promising pairs (Fig. 2b). The strongly complementary pairs (met10Δ-str3Δ and str3Δ-leu2Δ) reached maximum OD600 values of 27.3–28.5, slightly higher than that of the FY00 control (OD600 = 26.8, p > 0.1). However, their lag phases were obviously longer than that of FY00 (p < 10−4), reflecting the time required for metabolic exchange and population establishment. In comparison, moderately complementary pairs (met10Δ-leu2Δ, met10Δ-met14Δ, ura3Δ-leu2Δ, and met10Δ-met3Δ) exhibited lower maximum OD600 (24-26) and extremely extended lag phases (70-144 h). Among all tested pairs, met10Δ-str3Δ and str3Δ-leu2Δ exhibited the most reproducible and robust complementation. Community adaptation to methanol as carbon source To assess whether complementation persisted under alternative carbon sources, selected pairs were cultivated in minimal medium containing 10 or 20 g/L methanol (Fig. 3). Methanol metabolism is energetically less favorable than glucose metabolism and involves distinct one-carbon and peroxisomal pathways, potentially altering the dynamics of cross-feeding. Growth of strongly complementary auxotrophic pairs on methanol as the sole carbon source. (a) Growth curves with the specific growth rates (μ) and maximum OD600 in minimal medium containing 10 g/L methanol. (b) Growth curves with the specific growth rates (μ) and maximum OD600 in minimal medium containing 20 g/L methanol. The prototrophic strain FY00 is included as a reference. Co-cultures were initiated at a 1:1 ratio with total initial OD600 = 0.6 and cultivated at 37 °C with shaking at 220 rpm. Data are mean ± SD from four biological replicates. Under 10 g/L methanol, FY00 entered exponential phase at ∼12 h and reached OD600 of 14 by ∼60 h, after which a decline was observed. Most complementary pairs showed reduced overall biomass, but several sulfur-metabolism-related pairs (met10Δ-met14Δ, met10Δ-str3Δ, met10Δ-met3Δ) initiated growth at ∼12 h, close to FY00, albeit with slower rates and lower final densities (OD600 ≈ 5 at 72 h, Welch's t-test, both p < 0.001) (Fig. 3a). Under 20 g/L methanol, FY00 exhibited a longer lag phase (∼24 h) and reached OD600 of 22-25 by 72 h. Among the tested pairs, met10Δ-str3Δ showed the strongest recovery, with rapid growth between 24 and 48 h and a peak OD600 of ∼13-14 at 96 h. Even so, compared with FY00, met10Δ-str3Δ showed significantly lower maximum OD600 (Welch's t-test, p < 0.001) and μmax (Welch's t-test, p < 0.01). Other pairs, including ura3Δ-leu2Δ, met10Δ-leu2Δ, and met14Δ-leu2Δ, exhibited slower but sustained growth, reaching final OD600 values of 7-9. met10Δ-met14Δ and met10Δ-met3Δ showed biphasic growth patterns, with an initial rise followed by a plateau and a later increase (Fig. 3b). Overall, met10Δ-str3Δ consistently performed well under both methanol concentrations and was selected for further characterization. Population dynamics as a function of initial inoculation ratio To investigate how initial composition affects community establishment, FY31 (str3Δ-EGFP) and FY34 (met10Δ-mCherry) were co-cultured at initial ratios ranging from 1:10 to 10:1 (met10Δ:str3Δ). Growth and population ratios were monitored over time (Fig. 4a). Population dynamics of the met10Δ-str3Δ consortium under different initial inoculation ratios. (a) Schematic illustration of the experimental setup and representative fluorescence microscopy images showing the distribution of met10Δ-mCherry (red) and str3Δ-EGFP (green) cells in co-culture. (b) Growth curves of co-cultures initiated at different met10Δ:str3Δ ratios (1:10, 1:5, 1:1, 5:1, and 10:1). (c) Dynamics of the met10Δ fraction over time for the same initial ratios. The fraction was determined by counting red and green fluorescent cells in at least three random microscopy fields per sample (≥300 cells total per replicate). (d) Endpoint quantification of the met10Δ fraction at 168 h in co-cultures with different initial inoculation ratios. Cells were cultivated in Delft minimal medium containing 20 g/L glucose at 37 °C with shaking at 220 rpm. Total initial OD600 was set to 0.4. Data points represent mean ± SD from three biological replicates. Statistical analysis was performed for the 168 h time point using one-way ANOVA followed by Tukey's multiple comparison test. Groups that do not share a lowercase letter are significantly different (p < 0.05). The initial ratio strongly influenced overall biomass accumulation (Fig. 4b). Communities with initially higher frequencies of str3Δ (1:10, 1:5, and 1:1 groups) achieved higher final OD600 values. In contrast, groups initially dominated by met10Δ (5:1 and 10:1) showed weak growth throughout. Population tracking revealed substantial ratio reorganization during cultivation (Fig. 4c). In the 1:10 and 1:5 groups, the fraction of met10Δ increased rapidly during early growth. The 1:1 group also shifted toward met10Δ enrichment over time. In the 5:1 and 10:1 groups, the initial dominance of met10Δ was maintained, but this did not translate into high biomass. Endpoint quantification at 168 h showed a similar trend that the final community composition generally showed a bias toward met10Δ (Fig. 4d). Metabolite basis of complementation in the met10Δ-str3Δ pair In yeast, the sulfur assimilation pathway converts inorganic sulfate into organic sulfur-containing amino acids (Fig. 5a). Sulfate is first reduced to sulfide, which is then incorporated into homocysteine (Hcy) via O-acetylserine (OAS) and O-acetylhomoserine (OAH). Hcy serves as a central precursor for the biosynthesis of l-cysteine (L-Cys), cystathionine (Cysta), and l-methionine (L-Met). In this pathway, MET10 encodes a sulfite reductase that catalyzes the reduction of sulfite to sulfide, while STR3 encodes a cystathionine β-lyase that cleaves cystathionine to homocysteine [38,39]. Disruption of these genes leads to auxotrophies that can potentially be complemented by cross-feeding of specific intermediates. Metabolic basis of complementation in the met10Δ-str3Δ pair. (a) Simplified schematic of asymmetric metabolite exchange between met10Δ and str3Δ. The met10Δ and str3Δ mutants harbor deletions in the sulfur assimilation pathway at MET10 and STR3, respectively. met10Δ supplies Hcy to str3Δ, while str3Δ releases Cys, Cysta, and Hcy for met10Δ growth. Abbreviations: L-Cys, l-cysteine; Hcy, homocysteine; Cysta, Cystathionine; OAS, O-acetylserine; OAH, O-acetylhomoserine; L-Met, l-methionine. (b) Growth of the met10Δ single mutant in minimal glucose medium (20 g/L) supplemented with various sulfur-containing compounds at 20 mg/L. (c) Growth of the str3Δ single mutant under the same supplementation conditions as in (b). All cultures were inoculated at an initial OD600 = 0.2 and incubated at 37 °C with shaking at 220 rpm. To explore the metabolic basis of the observed interaction, met10Δ and str3Δ were cultivated individually in minimal medium supplemented with various sulfur-containing intermediates: sodium sulfide (Na2S), L-Cys, Cysta, or Hcy (Fig. 5b and c). met10Δ recovered growth in the presence of L-Cys, Cysta, or Hcy, but not with Na2S. str3Δ, by contrast, showed substantial growth recovery only with Hcy. These results are consistent with the known positions of the metabolic blocks except Na2S (Fig. 5a), which might be attributed to impaired sulfur assimilation pathway, lack of efficient uptake systems, or cytotoxicity. This supplement experiment may suggest a plausible model in which met10Δ can utilize multiple downstream sulfur-containing organic intermediates to restore growth, while str3Δ shows a higher dependence on Hcy. Discussion In this study, we systematically evaluated an auxotrophy-based cross-feeding strategy in the non-conventional yeast O. polymorpha and identified the met10Δ-str3Δ pair as a robust cooperative consortium with asymmetric metabolic dependency. These findings provide both experimental foundations and mechanistic insights for designing stable synthetic consortia in non-conventional yeasts. The observation that only a few auxotrophic pairs showed effective complementation highlights a key principle in consortium design: successful cross-feeding depends not simply on combining any two auxotrophs, but on the specific metabolic nodes disrupted and the exchange potential between partners [26,29]. The prevalence of sulfur-metabolism-related pairs among the best-performing combinations suggests that certain pathways are more amenable to cross-feeding, likely due to the availability and diffusibility of key intermediates. The sulfur assimilation pathway, with its multiple branch points and exchangeable organic intermediates (L-Cys, Cysta, Hcy), appears particularly suitable for mutualistic exchange [40]. The met10Δ-str3Δ pair showed strong complementation under both glucose and methanol conditions. These two mutants are disrupted at different steps in the sulfur assimilation pathway, creating an asymmetric relationship that may explain their superior performance. Supplementation assays revealed that met10Δ could utilize L-Cys, Cysta, or Hcy, whereas str3Δ responded only to Hcy. This asymmetry, inferred from single strain supplementation profiles, may enable a directional flow of intermediates that stabilizes the interaction. str3Δ provides early precursors allowing met10Δ to initiate growth, after which met10Δ expands more rapidly due to its broader metabolic capacity. The isotopic labeling or targeted metabolomics would help to directly identify the specific metabolites exchanged in co-culture [41,42]. Studies on yeast sulfur metabolism have shown that different pathway disruptions impose distinct organic sulfur requirements [40,43], providing a metabolic basis for the combination-specific differences observed here [44]. Fluorescence-based population tracking revealed substantial compositional reorganization during cultivation. When str3Δ was initially more abundant, the community established readily and achieved higher final biomass, while the fraction of met10Δ increased rapidly. Conversely, when met10Δ was initially overrepresented, its ratio advantage was maintained but overall growth was severely limited. These findings highlight the critical role of initial inoculation ratio in community assembly [14,45]. Our results also show that “stability” in such consortia should not mean constant composition, but rather the capacity to maintain growth and function within a range of compositional fluctuations [17]. The met10Δ-str3Δ system thus represents a dynamically stable consortium with pronounced initial ratio dependence. Another important finding in this study manifested the ability of met10Δ-str3Δ to maintain strong complementation under methanol, which extends its potential utility beyond glucose-based fermentations. As a methylotrophic yeast, O. polymorpha is increasingly used as an industrial chassis for methanol-based biomanufacturing. Our results show that this host supports not only single-strain engineering but also synthetic consortia based on nutritional complementation. The biphasic growth was observed for certain pairs (e.g., met10Δ-met14Δ and met10Δ-met3Δ) under 20 g/L methanol, which may reflect a two-stage metabolic adaptation resulting from the gradual release of accumulated sulfur intermediates into the medium. An initial slow growth phase possibly due to limited methanol assimilation or intermediate accumulation, followed by a second growth phase after the establishment of effective cross-feeding. In the future, long-term continuous culture would further assess the inherent stability of this consortium. In practical biomanufacturing scenarios, the presence of a biosynthetic pathway and associated metabolic burden may exert stronger selective pressure. Evaluating stability under product-synthesis conditions will expand the applicability of this cross-feeding design toward industrial application. Conclusion This study demonstrates that auxotrophy-based cross-feeding can be successfully implemented in the non-conventional yeast O. polymorpha, but that successful community assembly depends critically on the choice of metabolic nodes and the initial inoculation ratio. The met10Δ-str3Δ pair exhibits robust, asymmetric complementation that is stable across carbon sources and amenable to population-level control. These findings offer a reference framework for developing similar cross-feeding strategies and open new avenues for distributed biosynthesis strategies in non-conventional yeasts. CRediT authorship contribution statement Yuxia Fan: Writing – original draft, Visualization, Investigation. Xiaoxin Zhai: Validation, Investigation. Yongjin J. Zhou: Writing – review & editing, Funding acquisition, Conceptualization. Jiaoqi Gao: Writing – review & editing, Visualization, Supervision, Funding acquisition. Data availability statements The data that support the findings of this study are available within the article, Supplementary Information files. Declaration of competing interests The author Yongjin J. Zhou is an Editor-in-Chief for Synthetic and Systems Biotechnology and was not involved in the editorial review or the decision to publish this article. All authors declare that there are no competing interests. Acknowledgments This work was supported by National Key Research and Development Program of China (2024YFB4206300), the National Natural Science Foundation of China (22478382), and Key project at central government level: The ability establishment of sustainable use for valuable Chinese medicine resources (2060302). Footnotes Peer review under the responsibility of Editorial Board of Synthetic and Systems Biotechnology. Supplementary data to this article can be found online at https://doi.org/10.1016/j.synbio.2026.05.008. Supplementary data The following is the Supplementary data to this article: Multimedia component 1 References Associated Data Supplementary Materials Multimedia component 1 Data Availability Statement The data that support the findings of this study are available within the article, Supplementary Information files.