Database for personalized microbiota correction by biologically active compounds and plants extracts

Підписуйтесь на наші соціальні мережі, щоб стежити за останніми новинами тут 💜:

Сайт: www.ediens.me

Facebook: www.facebook.com/profile.php?id=61561258499410  

LinkedIn: www.linkedin.com/ediens 

Instagram: www.instagram.com/ediens_official

TikTok: www.tiktok.com/@ediens_official 

Database for personalized microbiota correction by biologically active compounds and plants extracts Roman Rukavchuk1 , Volodymyr Drobnych2 , Oleksandra Pallah1,3, Nataliya Korol4 , Svitlana Burmei1,3 and Nadiya Boyko1,3* 1Research Development and Educational Centre of Molecular Microbiology and Mucosal Immunology, Uzhhorod National University, 88000, Uzhhorod, Narodna sq. 1, Ukraine 2 Department of Geodesy, Land Management and Geoinformatics, Uzhhorod National University, 88000, Uzhhorod, Narodna sq. 1, Ukraine 3 Department of Medical and Biological Science, Faculty of Dentistry, Uzhhorod National University, 88000, Uzhhorod Universytetska st. 16a, Ukraine 4 Department of Organic Chemistry, Educational and Research Institute of Chemistry and Ecology, Uzhhorod National University, 88000, Uzhhorod, Fedyntsa st. 53/1, Ukraine *corresponding author Nadiya Boyko, email: nadiya.boyko@uzhnu.edu.ua Abstract Biologically active compounds and plants extracts are promising alternatives for biological preparations in personalized medicine, especially for personalized microbiota correction. However, due to fact that microbiota is a large complex system, the composition of which varies from person to person, there are various issues that needs to be solved in order to use biologically active compounds and plants extracts for microbiota correction. In this study, we described all these issues as well as possible solutions to them and proposed the solution to the most fundamental issue for personalized microbiota correction i.e. data source (database). The developed database consists of 7,717 records on plants effects on microorganisms and 3,034 records on biologically active compounds effects on microorganisms. The database contains not only data of inhibitory effect but also stimulatory effect data as well as the unification of all data, which is a key feature of this database. We also performed various types of analysis on extracted data, i.e. correlation analysis, the results of which could be used for development different prognostics and machine-learning models. The database is a first step towards accurate personalized microbiome correction. Keywords This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed personalized medicine, polyphenols, anthocyanins, correlation analysis, data source 1. Introduction Microbiota is a determining factor of human health [1, 2]. Improper, inadequate nutrition, unhealthy lifestyle, harmful effects of the environment, etc. cause negative changes in the composition and in the functions of microbiota [3, 4], which are one of the key triggers of various diseases [5, 6]. Nowadays, various biological preparations are often used to correct the microbiota such as pro-, pre-, symbiotics, etc. One of the most promising alternative for them are biologically active compounds (BACs) isolated from plants, in particular polyphenols and anthocyanins [7-9]. BACs and their plant sources, primarily plant extracts (PEs), have a wide range of valuable properties [10], in particular, they are capable of stimulating or inhibiting growth of various microbiota components [11]. Due to these properties, BACs and PEs are often used as components of complex biological preparations and therapeutic personalized nutrition [12, 13]. However, the problem of their optimal or successful selection to ensure adequate correction of the disturbed microbiota of a specific person (hereinafter - the Problem) is still far from a final solution. 2. Theory In order to understand the Problem, let us determine it by considering a simplified problem of personalized microbiota correction, for example, of the intestinal microbiota. Suppose that: a) A list of those types of microorganisms that should be corrected has already been determined for the patient. b) For each of these species we know the initial (impaired) concentration and the range of acceptable final concentrations. c) It is necessary to determine the list and doses of those BACs and/or PEs, the use of which will return the concentrations of all problematic types of microorganisms to acceptable limits and at the same time will not unbalance other components of the intestinal microbiota. This problem can be solved in different ways (they are considered in the "Discussion" section), but in the presence of sufficient data on the effects of various BACs and PEs on problematic and other components of the intestinal microbiota. In our opinion, the most significant points of the Problem are precisely related to obtaining such information. The main one of these points is the need to obtain a very significant volume of the specified data due to the huge number of components of the intestinal microbiota [14]. In addition, the need to use a wide range of BACs and PEs, adequate to the variety of initial conditions of specific tasks of personalized correction of the microbiota. This information is currently being intensively accumulated. Although the process is still far from complete, the array of data obtained today This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed already allows us to move to the practical application of the idea of microbiota correction using BACs and PEs. However, the use a significant part of the array of information obtained is currently practically impossible due to its dispersion in numerous articles (for example [15] and local databases (such a database was created by our team based on the results of our own research on the impact of BACs and PEs on microorganisms [16]). This circumstance constitutes the second important point of the Problem, and its solution is the main goal of this work. The goal of the work was achieved by compiling the discussed array of available information into a widely available database, and which was designed to ensure the maximum convenience of using such data to solve the problem of microbiota correction using BACs and PEs. The construction of the specified database opened up a real possibility of applying modern data science to almost the entire array of available information on the impact of BACs and PEs on microorganisms. The proposed work is the first such study. Namely, in order to search for patterns hidden in the data (required, in particular, for the extrapolation of existing information about specific pairs "BACs - microorganism" and "PEs - microorganism" to pairs that have not yet been investigated), a correlation analysis of all data available in the database was carried out. 3. Materials and methods 3.1 Database development Development of the database as well as user interface for its usage included such steps: 1. Search, selection, extraction and analysis of literature data on the impact of BACs and PEs on microorganisms, 2. Development of a tabular structure of the database and the filling of its relational tables with selected and appropriately grouped information, 3. Construction of tools that enable convenient use of the information. PubMed was used to search for literature sources with the following query: ((plant[Title/Abstract]) OR (herb[Title/Abstract]) OR (vegetable[Title/Abstract])) AND ((compound[Title/Abstract]) OR (polyphenol[Title/Abstract]) OR (anthocyanins[Title/Abstract]) OR (flavonoids[Title/Abstract]) OR (bioflavonoids[Title/Abstract]) OR (phenolics[Title/Abstract])) AND ((microorganism[Title/Abstract]) OR (bacteria[Title/Abstract])) AND ((interaction) OR (influence) OR (activity) OR (inhibition)). 1,214 records were found, of which 107 publications were selected after preliminary analysis. From these publications 7,717 records on plants extracts effects on microorganisms and 3,034 records on BACs effects on microorganisms were selected, which included data on 191 plants (plants extracts), 643 microorganisms, and 167 BACs. The data This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed selected were first entered into an Excel spreadsheet and then, after analysis and grouping, into a PostgreSQL database. In order to provide online access to the database, the well-known Django 3.2 framework (https://www.djangoproject.com/) was used, which implements the MTV pattern (model template view, MVC analog — model view controller) and uses the Python programming language (https://www.python.org/). PostgreSQL database (https://www.postgresql.org/) was used to store, sort, and manage data. Ubuntu Linux was used to ensure online access as a web server nginx (https://nginx.org/ru/), which operates together with Gunicorn (https://gunicorn.org/). The Frontend part was implemented using HTML, CSS, JavaScript, and Bootstrap. 3.2 Statistical analysis Statistical analysis of extracted data (required for manipulation of the latter, in particular, for their aggregation) was carried out by well-known methods of descriptive statistics and statistical inference [17], using the libraries of the Anaconda – Python environment created for scientific research and the OriginPro software package from OriginLab for numerical data analysis and scientific graphics. The search for statistically significant pairwise Pearson and Spearman correlations between types of microorganisms (according to the effect on them of biologically active compounds and plant extracts presented in the database), and between types of BACs or PEs (according to their effect on the microorganisms presented in the database) was carried out by using the Stats library of the SciPy library module of the Anaconda - Python environment. 4. Results 4.1 Database The created database is available at https://pbaa.xyz/. After opening it, the user has access to six tabs of the built web interface of the database. The most important in terms of practical use of the database are the “Plants” and “Compounds” tabs (which open access to web tools designed to obtain information about the effect of the user-selected PEs or BACs on certain microorganisms), as well as the “Microorganisms” tab (which in a similar way, it provides data on the effect of certain PEs and BACs on the microorganism selected by the user). 4.1.1 Data for effects of PEs and BACs on microorganisms The "Plants" tab provides access to a general data table on the effects of plant extracts on microorganisms (information grouped by plants and microorganisms). Here, the user can conveniently and quickly find the data he needs using appropriate filters (fig. 1A). By choosing a specific pair "plant - microorganism" in the search results (see, for example, the last line in Fig. This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed 2A, which corresponds to the pair "African Olive - Candida albicans"), you can get the main information available in the database about this pair (fig. 1B). In particular, the values of the impact of PEs on the microorganism measured by the authors of the relevant articles (by one method or another, in one or another unit) ("Value" column), as well as the results of the reduction of these values to single units performed by DB tools ("Interpretation" column"). In addition, for the selected pair "PE - microorganism" it is possible to obtain extended, i.e., almost all-available information on the plant and microorganism in the database (fig. 1C). As for the data on the effects of BACs on microorganisms, the user receives them through the "Compounds" tab. Everything is organized exactly in the same way as the "Plants" case just considered

4.1.2 “Microorganisms” tab The "Plants" and "Compounds" tabs should be used when solving problems in which the search for the necessary information in the database begins with the selection of PEs or BACs. The tab "Microorganisms" should be used in cases where it is more convenient to start the search for the necessary information with the selection of microorganisms. This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed The web tools activated by the "Microorganisms" tab are similar to those discussed in point 4.1.1. The main difference is that when starting the data selection filters corresponding to the selected microorganisms, both plants and BACs are selected. 4.1.3 Presentation of quantitative data on the impact of PEs and BACs on microorganisms As noted in section 3.1.1, quantitative data on the effect of PEs and BACs on microorganisms are presented in the created database as original values, determined by the authors of the corresponding articles by one or another method, in one or another units, and as unified, interpreted values, obtained by reducing the original values to single dimensionless units. The use of unified values makes it possible to solve the problems of personalized correction of microbiota using all available data on the effect of PEs and BACs on microorganisms, obtained by different methods and presented in different units. Recalculation of interpreted values is carried out by appropriate database tools that implement the following calculation algorithm: 1) All available original values (for plants and BACs separately) are grouped by methods of determination and within each method are reduced to one unit (µg · ml-1, mm, etc.) 2) Within each group, we determine the original value x0, which corresponds to the absence of influence, as well as the original value xmax of the maximum inhibitory influence, 3) The unified value of influence yi corresponding to the original value xi is determined separately for each group using the formula. ; (1) 4) Interpreted values for different groups were reduced to one group (in our case – the largest group of experimental data obtained by the minimum inhibitory concentration method), using the appropriate normalizing factors obtained by comparing the interpreted values for the same pairs of PEs-microorganisms or BACs-microorganisms, presented in different groups. This procedure is repeated every time when new information is introduced into the database, which ensures not only the expansion of the set of interpreted values, but also the constant refinement of unified values. Note that equation (1) is a variant of the standard formula for bringing a set of experimental data to a certain, predetermined range of values (https://scikitlearn.org/stable/modules/generated/sklearn.preprocessing.MinMaxScaler.html). Interpreted values of inhibitory effects are negative and range from -1 (maximum inhibitory effect) to 0 (absence of effect). As for the effects that correspond to the stimulation of the growth of 0 max 0 i i x x y x x     This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed microorganisms, their interpreted values according to formula (1) are positive. At the same time, the maximum stimulating effect, in contrast to the maximum inhibitory effect, does not correspond to a certain, preselected value. This asymmetry in taking into account inhibitory and stimulating effects in the construction of formula (1) is due to the fact that, in a significant majority of cases, the effects of PEs or BACs on microorganisms are either absent or inhibitory. 4.2 Database data analysis The creation of the discussed and similar databases opens up the possibility of applying modern data science approaches to most of the information obtained today regarding the effects of BACs and PEs on microorganisms. In order to identify hidden patterns in these data (required, in particular, for extrapolation of existing information about specific pairs "BACs - microorganism" and "PEs - microorganism" to pairs that have not yet been discovered), a correlation analysis of all data available in the database was carried out. The analysis consisted in finding statistically significant pairwise correlations between the types of a) BACs (according to their influence on the microorganisms presented in the database), b) PEs (according to their influence on the microorganisms presented in the database), and c) Microorganisms (according to their influence on them presented in the database biologically active compounds and plant extracts). The specified correlations (we will hereinafter abbreviate them as correlations of BACs, PEs and microorganisms, respectively) were determined by the unified values of the effects of BACs and PEs on microorganisms. The coefficients r of pairwise Pearson and Spearman correlations and the p levels of statistical significance of each of these coefficients were used as characteristics of these correlations (coefficients with p < 0.05 were considered significant). The peculiarity of the analysis was that for each pair of indicators studied for correlation, in addition to r and p, the set of initial information available in the database was determined. Data was checked for the "normality" of data distributions for its respective parts (the so-called samples, which, depending on the results of the test, were characterized by mean or median values of interpreted values for calculating Pearson's and Spearman's correlation coefficients, respectively). For many pairs of indicators, such a set of initial data was either missing or insufficient to determine the correlation coefficient. A specific pair of indicators in each of the three indicated pictures is presented either by a certain number (the value of the statistically significant coefficient r), or by indicating the fact that the obtained correlation coefficient is not statistically significant, or by indicating the fact that the initial data are insufficient to calculate the correlation coefficient (table 1).

As can be seen from the fig. 2, most of the found statistically significant correlations belong (according to the absolute values of the coefficients r) to the range of 0.9-1, i.e., according to the generally accepted classification, they belong to “very high” correlations. At the same time, 30.9% of all statistically significant correlation for BACs, 40% - for plants and 37.5% - for microorganisms belong to the range 0.98 - 1, which indicates practically identical properties of both components of the correlated pair of indicators. For statistically significant correlations for BACs, they were found for 366 pairs of biologically active compounds (see Table 1), in which 95 BACs appear, i.e., 57.6% of the total number of biologically active compounds presented in the database. 92.6% of these 366 statistically significant correlations are characterized by positive correlations (corresponding to positive values of r coefficients) and 7.4% by negative correlations (for them, r values are negative). Quercetin (27 correlations), Catechin (26), Hesperidin (25), Ferulic acid (24), Naringenin (22), Apigenin (20) and Caffeic Acid (20) are distinguished by the number of statistically significant correlations with other BAСs. In terms of statistically significant correlations, o-coumaric and p-coumaric can also be distinguished. They form a correlated pair with the maximum possible r = 1, and they are also equally correlated (with r = 1) to the same BACs mentioned before. This circumstance defines existence and the possibility of identifying groups of biologically active compounds with an identical effect on certain sets of microorganisms. For extracted data, we calculated correlations matrix, which contains comprehensive information about the found correlations of the biologically active compounds presented in the database (fig. 3). Fig 3. presents 39 BACs, each of which has 8 or more statistically significant correlations with other biologically active compounds. The selected BACs are located on the diagram in order of decreasing number of their statistically significant correlations (from 27 for Quercetin to 8 for ocoumaric and p-coumaric). The diagram allows identifying new groups of biologically active compounds characterized by a high similarity of effects on certain sets of microorganisms (analogous to the case of o-coumaric and p-coumaric considered above). As can be seen from the figure, the symmetry of the correlation This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed pattern corresponding to such groups occurs for the pairs 1) Pelargonidin and Resorcinol (it can be seen that this pair is characterized by 10 identical statistically significant correlations), 2) Kaempferol and Luteolin (15 statistically significant correlations) and 3) 3-Dihydroxyphenylacetic acid and 3 -Methoxy-4-hydroxyphenylacetic acid (18). The above-mentioned pair of o-coumaric and p-coumaric can be supplemented with BAC Ginestein. It is not difficult to identify other, slightly less ideal, but more numerous groups of BACs, which are characterized by a high similarity of effects on microorganisms. Finally, all 39 selected BACs form a certain group according to correlation properties, because the vast majority of their statistically significant correlations are implemented within this group, (exceptions are only some BACs, which implement part of their statistically significant correlations outside this group). For plant extracts grouped by plant families statistically significant correlations were found for 557 pairs of plants families (see Table 1), in which 120 of the 186 plant families that represented in the database appear. 75.0% and 25.0% of these 557 correlations are positive and negative; the percentage of negative statistically significant correlations for plants is almost 4 times higher than in the case of BACs. Cissus trifoliata (22 correlations), Rubus idaeus (22), Chrysanthemum morifolium Ramat (21), Cinnamomum burmannii B. (21), Fagopyrum cymosum (Trev.) Meisn (21), Mentha x piperita (20), Sanguisorba officinalis L. (20), Cinnamomum cassia Presl (19), Fragaria ananassa (19), Mentha canadensis L. (19), Sambucus nigra L. (19). At the same time, as can be seen from fig. 2, the percentage of the highest absolute values of statistically significant correlations (belonging to the range 0.98 – 1) here is significantly higher than in the case of BACs. Comparing the correlations of plants families and BACs in terms of the percentage of all detected statistically significant correlations the cases of plant families and BACs differ insignificantly (3.2% and 2.7%, respectively); in terms of percentages of non-statistically significant correlations, plants families significantly prevails over BACs (48.2% and 31.2%, respectively), and the percentage of correlations not calculated due to lack of initial data is significantly lower for plants families than for BACs (48.6% and 66.1%). The resulting matrix of correlations is represented in Fig. 4 in a part of 40 plants families, each of which has 14 or more statistically significant correlations with other plant genera. The selected plants families are located on the diagram in the order of decreasing numbers of their statistically significant correlations connections (from 22 for Cissus trifoliata to 14 for Zanthoxylum bungeanum Maxim.). This diagram is similar to the diagram in fig. 3 and allows identifying groups of indicators by correlation, which are characterized by high similarity of influence on certain sets of microorganisms. As can be seen from fig. 4, the symmetry of the correlation pattern required for the presence of such a group is present for the trinity of Satureja montana ЅL., Teucrium This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed montanum ЅL., and Urtica dioica ЅL., which should be supplemented with Polygonum aviculare ЅL., Veronica officinalis, and Galium verum ЅL. All six plants families from this group have 14 statistically significant correlations each, of which 13 are within the set of 40 plants families selected for this diagram. The absence of green color on this diagram (in contrast to Fig. 3) is related to the relatively satisfactory amount of data from the database with information on the effects of plant extracts from the 40 plant families presented on the diagram on various types of microorganisms

Statistically significant correlations for microorganisms species according to the effect on them of biologically active compounds and plant extracts presented in the database were found for 424 pairs of (see Table 1). These correlations include 108 of the 236 types of microorganisms represented in the database, i.e., 45.8%. Of the identified statistically significant correlations, 81.4% are characterized by positive correlations and, accordingly, 7.4% by negative correlations.

The correlation matrix obtained for microorganisms species presented in Fig. 5 contains 38 bacterial species, each of which has 8 or more statistically significant correlations with other types of microorganisms. The selected microorganisms species are located on the diagram in the order of decreasing numbers of their statistically significant correlation (from 45 for Escherichia coli to 8 for Shigella dysenteriae).

This diagram allows identifying groups of microorganisms species characterized by a high similarity of the response to the action of certain BACs and plant extracts. As can be seen from fig. 5, the symmetry of the correlation pattern is required for the presence of such a group, for example, for the pair of Listeria innocua, Sarcina lutea, which should be supplemented with Micrococcus flavus. These three microorganisms species have 10, 10 and 11 statistically significant correlations, of which only one statistically significant correlation (referring to Micrococcus flavus) exists outside the set of 38 microorganisms species selected for this diagram. Table 1 demonstrates that the pattern of microorganisms species correlations differ from the two correlation patterns discussed above by a much higher percentage of correlations, which cannot be calculated at the moment due to the lack of relevant initial data. On the diagram, this should be manifested in an increase in the relative number of green squares. But as can be seen from fig. 5, such percentage on the diagram naturally increases when moving from microorganisms species This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed with the largest amounts of statistically significant correlations to microorganisms species with decreasing amounts of statistically significant correlations (that is, when moving along the yellow diagonal from the bottom left to the top right). This indicates that the smaller numbers of statistically significant correlations for microorganisms species, located in the second half of the list of 38 selected bacterial species, will most likely significantly increase when the database is supplemented with new raw data. Accordingly, information will appear about new groups of microorganisms species, characterized by high similarity in the effect of biologically active compounds and plant extracts on them. 5. Discussion Accurate personalized selection of BACs or plants for microbiota correction is a certain kind of abstraction, because adjusting all available microorganisms of the microbiota to specific values is an extremely time-consuming task due to a number of reasons. The first is the determination of the appropriate values of the quantities of microorganisms for adjustment, in other words, reference ranges. Microbiota and microbiome are complex systems and contains a significant number of non-trivial relationships between its components, which include not only microorganisms, but also various metabolites, viruses, etc. [18]. In addition, the number of factors that can affect the microbiome and microbiota is significant, and their researches continues [19, 4]. Another feature that follows from the mentioned is the specific features of each specific organism for which the appropriate adjustment is needed. For accurate adjustment, it is necessary to have an exhaustive list of its parameters, which is currently an intractable task. Accordingly, reference ranges usually mean certain ranges of values that are characteristic of a certain group of healthy organisms [20]. Determining such reference ranges of microbiota is an extremely nontrivial task due to the high variability of microbiota in healthy organisms [18, 14]. This problem became especially relevant with the rapid development of sequencing methods and the discovery of uncultivated microorganisms [22, 21]. Nowadays, there are many results in the direction of determining microbiota reference ranges [23, 24], however, they do not yet provide a clear understanding of this concept and the corresponding meanings. Therefore, much simplified models or time-tested postulates of microbiology are often used in clinical practice [25]. Another reason for the abstractness of the concept of precision selection for microbiota adjustment is the limited data on the effect of desired components, plants or BACs, on microorganisms. For accurate adjustment, it is necessary to have data on the effect of all selected components on each microorganism. Considering the large number of microorganisms of the microbiota [13], it is necessary to have an extremely large amount of such data, which is currently an unsolvable problem. Moreover, such data should be obtained as a result of experiments on the This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed influence of plants or BACs in vivo. The mentioned studies are increasingly being conducted recently [26, 27], however, compared to in vitro studies, their number is insufficient. Therefore, data obtained from the results of in vitro experiments are used more often. Such data are also insufficient to accurately determine the components for microbiota correction, including due to the limitations of appropriate experimental methods [28, 29]. The mentioned limitation of data can be solved by using data on the already available effects of plants and BACs on microorganisms and the corresponding relationships between the effects of these components or microorganisms. The possibility of such a solution to the mentioned problem is clearly demonstrated by the results obtained by us on the determination of correlation coefficients (Fig. 3-5). Correlation coefficients for microorganisms are of particular value in this direction, according to which it is possible to assert about one or another effect of plants or BACs on a microorganism, based on known data on the effect of these components on another microorganism, species, family, genus, class and relationship of the corresponding impact with impact on the desired microorganism. The obtained data also make it possible to solve the inverse problem with known data on the relationship between the influence of plants or BACs on microorganisms. Obtaining the results described above regarding correlations, and in general, the selection of plants or BACs is a non-trivial task without quick access and the ability to manipulate data on the effects of components on microorganisms, as well as their unification. The mentioned problem is solved by the database created by us. Unlike existing similar databases [31-33], it is open and contains appropriate data unification. The resulting unification of data is also an abstraction due to significant differences in research methods and conditions of conducting experiments. It greatly simplifies the comparison of relevant results and allows you not to go into the specific details of each experiment, but does not allow you to perform the selection accurately. It is also obvious that only data on the effect of plants and BACs on microorganisms is often not enough for a full selection of the appropriate plant or BAC. The task arises for the selected plant to determine the BACs composition and the effect of each individual component on microorganisms, or vice versa, for the selected BAC to determine their content in the plant and the corresponding effect in the complex. The mentioned problems cannot be solved without the availability of data on the BACs content in plants. Such data are presented in relevant databases, in particular Phenol-Explorer and ePlantLibra [34, 35], the latter has no open access policity. You can find the necessary BACs in Phenol-Explorer using the PubChem ID, or the corresponding BAC name, and plants using their names. To facilitate the search for relevant BACs in the created database, the PubChem ID field is provided for each added BAC, and therefore the problem described above can be solved quite simply. This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed It is also worth noting the possibility of varying the effect of plants on microorganisms, which are connected with the variety of their growing conditions [35], which obviously affect the BACs content in them. Therefore, it is promising to consider this fact for the selection and when researching the effect of plants on microorganisms. The solution to this problem can be the creation of a certain GIS, which would allow a wide range of scientists for each study to note the data on the origin of the plant, as well as to conduct various kinds of analyzes on the BACs content depending on the climate, origin, etc. Summarizing the above, when we talk about the exact personalized selection of BACs or plants for microbiota correction, we mean selection with a certain probability, which follows from the features described above. The more at one or another stage of the development of science we can move from abstractions to concrete models, the more accurate the appropriate selection will become. An important prerequisite for any accurate selection for microbiota correction is the presence of a mathematical basis. The best-known approach to targeted personalized microbiota and microbiome correction is described in the work of Zeevi et al. [36], which is based on predicting the glycemic response to the use of certain products. It is based on machine learning methods, in particular Gradient Boosting, and therefore, in order to obtain the corresponding results, it is necessary to carry out a significant number of calculations and use the corresponding source data. Other known algorithms use similar methods, usually machine learning, and have the same shortcomings. Therefore, our idea is to use linear programming methods for personalization, which have proven their effectiveness in solving similar problems, in particular, the well-known problem of calculating the cheapest food diets (Stigler's diet problem) [37] and its exact solution by George Danzig [38]. The proposed ideas are simple to implement, and the complexity of their implementation does not increase with an increase in the number of microorganisms. They can be implemented as tools in most known programming languages, the sufficient conditions are the use of appropriate programs or libraries capable of solving linear programming problems. It is obvious that the corresponding ideas, because of their simplicity, give a less accurate result, and require large-scale clinical studies before putting them into practice. However, the first successful results in this direction have already been made earlier by us [39]. An important advantage of the developed approach is the lack of dependence on the chosen method of microbiota or microbiome research, and therefore it is possible to use either classical methods of microbiology and sequencing, or their combination, which is impossible when using the above-mentioned machine learning methods. The advantage of the developed approach over analogues is also that, with the availability of data, it is possible to select plants or BACs with its help to adjust not only intestinal This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed microbiota, but also, for example, vaginal, oral, etc., which has recently become extremely relevant [40-42]. 5. Conclusion The described idea of using linear methods for personalized adjustment of microbiota using BACs or plants can be used for the development of appropriate biotechnology for personalized microbiota correction, as well as the development of biological preparations. The approach can also be used to select other components, such as microorganisms, and a necessary condition for this is the availability of relevant data on the effect of these components on specific representatives of the microbiota. An important result is the developed database on the effects of plants and BACs on microorganisms, which solves the actual problem of initial data for selection and allows a wide range of users to use the relevant data also for solving other problems, in particular, preliminary evaluation of components during scientific research. The database is mainly intended for use by microbiologists, because it allows you to learn about the possible influence of this or that plant, or a separate component from it, on microorganisms. All database data can be downloaded and used for further analysis and development of certain models. The results obtained by us from the correlation analysis of the created database are unique and allow us to reveal certain regularities in the influence of certain BACs, plants or the sensitivity of microorganisms to them. They can serve as a basis both for the selection and for replacing the corresponding components with others in the composition of biological preparations. These data can be used in the future to create a new, more accurate, personalized approach for microbiota correction, as well as to build various models, including machine learning. Funding: This work was supported by Ministry of Education and Science, Topic: "Personalized approaches to the diagnosis, prevention and treatment of heart diseases with prognostic modeling of the individual development of atherosclerosis", No 0120U102244 Author Contributions: Roman Rukavchuk: Analyzed and interpreted the data; Wrote the manuscript. Oleksandra Pallah: Analyzed and interpreted the data; Wrote the manuscript. Volodymyr Drobnych: Analyzed and interpreted the data; Wrote manuscript. Nataliya Korol: Analyzed and interpreted the data; Edited the manuscript. Svitlana Burmei: Analyzed and interpreted the data; Edited the manuscript. Nadiya Boyko:Conceived and designed the experiments.

Conflicts of Interest: The authors declare no conflicts of interest in relation to this article.

References

1. Altveş, S., Yildiz, H. K., Vural, H. C., 2020. Interaction of the microbiota with the human body This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed in health and diseases. Biosci. Microbiota, Food Heal, Volume 39, P. 23–32. https://doi.org/10.12938/bmfh.19-023 2. Fan, Y., Pedersen, O., 2021. Gut microbiota in human metabolic health and disease. Nat. Rev. Microbiol.Volume 19, P.55–71. https://doi.org/10.1038/s41579-020-0433-9 3. Ramirez, J., Guarner, F., Bustos Fernandez, L., Maruy, A., Sdepanian, V.L., Cohen, H., 2020. Antibiotics as Major Disruptors of Gut Microbiota. Front. Cell. Infect. Microbiol.Volume 10. https://doi.org/10.3389/fcimb.2020.572912 4. Wen, L., Duffy, A., 2017. Factors Influencing the Gut Microbiota, Inflammation, and Type 2 Diabetes. J. Nutr. Volume 147, P. 1468S-1475S. https://doi.org/10.3945/jn.116.240754 5. Carding, S., Verbeke, K., Vipond, D.T., Corfe, B.M., Owen, L.J., 2015. Dysbiosis of the gut microbiota in disease. Microb. Ecol. Heal. Dis. Volume 26. https://doi.org/10.3402/mehd.v26.26191 6. Wang, B., Yao, M., Lv, L., Ling, Z., Li, L., 2017. The Human Microbiota in Health and Disease. Engineering Volume 3, P. 71–82. https://doi.org/10.1016/J.ENG.2017.01.008 7. Cowan, M.M., 1999. Plant Products as Antimicrobial Agents. Clin. Microbiol. Rev. Volume 12, P. 564–582. https://doi.org/10.1128/CMR.12.4.564 8. Chen, C., Mohamad Razali, U. H., Saikim, F. H., Mahyudin, A., & Mohd Noor, N. Q. I. 2021. Morus alba L. plant: Bioactive compounds and potential as a functional food ingredient. Foods Volume 10, No. 3, P. 689. https://doi.org/10.3390/foods10030689 9. Makarewicz, M., Drożdż, I., Tarko, T., Duda-Chodak, A., 2021. The Interactions between Polyphenols and Microorganisms, Especially Gut Microbiota. Antioxidants Volume 10, P.188. https://doi.org/10.3390/antiox10020188 10. Tajner-Czopek, A., Gertchen, M., Rytel, E., Kita, A., Kucharska, A.Z., Sokół-Łętowska, A., 2020. Study of Antioxidant Activity of Some Medicinal Plants Having High Content of Caffeic Acid Derivatives. Antioxidants Volume 9, P. 412. https://doi.org/10.3390/antiox9050412 11. Pallah, O. V., Meleshko, T. V., Bati, V. V., & Boyko, N. V., 2019. Extracts of edible plants as beneficial microorganisms growth stimulators. Biotechnol. Acta Volume 12, P. 67–74. https://doi.org/10.15407/biotech12.03.067 12. Blasco, T., Pérez-Burillo, S., Balzerani, F., Hinojosa-Nogueira, D., Lerma-Aguilera, A., Pastoriza, S., Cendoya, X., Rubio, Á., Gosalbes, M.J., Jiménez-Hernández, N., Pilar Francino, M., Apaolaza, I., Rufián-Henares, J.Á., Planes, F.J., 2021. An extended reconstruction of human gut microbiota metabolism of dietary compounds. Nat. Commun. Volume 12, P. 4728. https://doi.org/10.1038/s41467-021-25056-x 13. Mithul Aravind, S., Wichienchot, S., Tsao, R., Ramakrishnan, S., Chakkaravarthi, S., 2021. This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed Role of dietary polyphenols on gut microbiota, their metabolites and health benefits. Food Res. Int. Volume 142, P. 110189. https://doi.org/10.1016/j.foodres.2021.110189 14. Turnbaugh, P. J., Ley, R. E., Hamady, M., Fraser-Liggett, C. M., Knight, R., Gordon, J. I., 2007. The Human Microbiome Project. Nature 449, 804–810. https://doi.org/10.1038/nature06244 15. Ozdal, T., Sela, D.A., Xiao, J., Boyacioglu, D., Chen, F., Capanoglu, E., 2016. The Reciprocal Interactions between Polyphenols and Gut Microbiota and Effects on Bioaccessibility. Nutrients Volume 8, P. 78. https://doi.org/10.3390/nu8020078 16. Meleshko, T., Rukavchuk, R., Buhyna, L., Pallah, O., Sukharev, S., Drobnych, V., Boyko, N., 2021. Biologically Active Substance Content in Edible Plants of Zakarpattia and Their Elemental Composition Model. Biol. Trace Elem. Res. Volume 199, P. 2387–2398. https://doi.org/10.1007/s12011-020-02345-y 17. Glantz, S. A., 2005. Primer of biostatistics. McGraw-Hill Medical Pub. Division. 18. Jiang, H., Cai, M., Shen, B., Wang, Q., Zhang, T., & Zhou, X., 2022. Synbiotics and Gut Microbiota: New Perspectives in the Treatment of Type 2 Diabetes Mellitus. Foods, Volume 11, No. 16, P. 2438. https://doi.org/10.3390/foods11162438 19. Hasan, N., Yang, H., 2019. Factors affecting the composition of the gut microbiota, and its modulation. PeerJ Volume 7, e7502. https://doi.org/10.7717/peerj.7502 20. Royston, P., 1991. Constructing time-specific reference ranges. Stat. Med. Volume 10, P. 675– 690. https://doi.org/10.1002/sim.4780100502 21. Nayfach, S., Shi, Z. J., Seshadri, R., Pollard, K. S., Kyrpides, N.C., 2019. New insights from uncultivated genomes of the global human gut microbiome. Nature, Volume 568, P. 505–510. https://doi.org/10.1038/s41586-019-1058-x 22. Zhu, B., Wang, X., Li, L., 2010. Human gut microbiome: the second genome of human body. Protein Cell Volume 1, P. 718–725. https://doi.org/10.1007/s13238-010-0093-z 23. Almonacid, D. E., Kraal, L., Ossandon, F. J., Budovskaya, Y. V., Cardenas, J. P., Bik, E. M., Goddard, A. D., Richman, J., Apte, Z. S., 2017. 16S rRNA gene sequencing and healthy reference ranges for 28 clinically relevant microbial taxa from the human gut microbiome. PLoS One, Volume 12, e0176555. https://doi.org/10.1371/journal.pone.0176555 24. Carroll, K. C., Hobden, J. A., Miller, S., Morse, S. A., Mietzner, T. A., Detrick, B., Mitchell, T. G., McKerrow, J. H., Sakanari, J. A., 2019. Normal Human Microbiota, in: Jawetz, Melnick, & Adelberg’s Medical Microbiology, 27e. McGraw-Hill Education, New York, NY. 25. Davis, C. P., 1996. Normal flora. Med. Microbiol. 4th Ed. 26. Rodríguez-Daza, M. C., Pulido-Mateos, E. C., Lupien-Meilleur, J., Guyonnet, D., Desjardins, This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed Y., Roy, D., 2021. Polyphenol-Mediated Gut Microbiota Modulation: Toward Prebiotics and Further. Front. Nutr. Volume 8. https://doi.org/10.3389/fnut.2021.689456 27. Yamakoshi, J., Tokutake, S., Kikuchi, M., Kubota, Y., Konishi, H., Mitsuoka, T., 2001. Effect of Proanthocyanidin-Rich Extract from Grape Seeds on Human Fecal Flora and Fecal Odor. Microb. Ecol. Health Dis. Volume 13, P. 25–31. https://doi.org/10.1080/089106001750071672 28. Dickert, H., Machka, K., Braveny, I., 1981. The uses and limitations of disc diffusion in the antibiotic sensitivity testing of bacteria. Infection, Volume 9, P. 18–24. https://doi.org/10.1007/BF01640803 29. Wen, X., Gehring, R., Stallbaumer, A., Riviere, J. E., Volkova, V. V., 2016. Limitations of MIC as sole metric of pharmacodynamic response across the range of antimicrobial susceptibilities within a single bacterial species. Sci. Rep. Volume 6, 37907. https://doi.org/10.1038/srep37907 30. Amirkia, V. D., Qiubao, P., 2011. The Antimicrobial Index: A comprehensive literature-based antimicrobial database and reference work. Bioinformation, Volume 5, P. 365–366. https://doi.org/10.6026/97320630005365 31. Babu, P. A., Kumar, P. S., Padmaja, P., Rao, T. K., Chitti, S., 2009. MIC database: A collection of antimicrobial compounds from literature. Bioinformation Volume 4, P. 75–77. https://doi.org/10.6026/97320630004075 32. Thomas, J., Navre, M., Rubio, A., Coukell, A., 2018. Shared Platform for Antibiotic Research and Knowledge: A Collaborative Tool to SPARK Antibiotic Discovery. ACS Infect. Dis. Volume 4, P. 1536–1539. https://doi.org/10.1021/acsinfecdis.8b00193 33. Bucchini, L., Rodarte, A., Restani, P., 2011. The PlantLIBRA Project: how we intend to innovate the science of botanicals. Food Funct. Volume 2, P. 769. https://doi.org/10.1039/c1fo10150a 34. Neveu, V., Perez-Jimenez, J., Vos, F., Crespy, V., du Chaffaut, L., Mennen, L., Knox, C., Eisner, R., Cruz, J., Wishart, D., Scalbert, A., 2010. Phenol-Explorer: an online comprehensive database on polyphenol contents in foods. Database 2010, bap024–bap024. https://doi.org/10.1093/database/bap024 35. Mikhailov, A. L., Timofeeva, O. A., Ogorodnova, U. A., Stepanov, N. S., others, 2019. Comparative analysis of biologically active substances in Trifolium pratense and Trifolium repens depending on the growing conditions. J. Environ. Treat. Tech. Volume 7, P. 874–877. 36. Zeevi, D., Korem, T., Zmora, N., Israeli, D., Rothschild, D., Weinberger, A., Ben-Yacov, O., Lador, D., Avnit-Sagi, T., Lotan-Pompan, M., Suez, J., Mahdi, J.A., Matot, E., Malka, G., Kosower, N., Rein, M., Zilberman-Schapira, G., Dohnalová, L., Pevsner-Fischer, M., This preprint research paper has not been peer reviewed. Electronic copy available at: https://ssrn.com/abstract=4956609 Preprint not peer reviewed Bikovsky, R., Halpern, Z., Elinav, E., Segal, E., 2015. Personalized Nutrition by Prediction of Glycemic Responses. Cell Volume 163, P. 1079–1094. https://doi.org/10.1016/j.cell.2015.11.001 37. Garille, S. G., Gass, S. I., 2001. Stigler’s diet problem revisited. Oper. Res.Volume 49, P. 1– 13. 38. Dantzig, G. B., 1990. The diet problem. Interfaces (Providence). Volume 20, P. 43–47. 39. Meleshko, T., Rukavchuk, R., Levchuk, O., & Boyko, N., 2021. Personalized Nutrition for Microbiota Correction and Metabolism Restore in Type 2 Diabetes Mellitus Patients. In Advances in Microbiology, Infectious Diseases and Public Health: Volume 16, P. 1-16. Cham: Springer International Publishing. 40. Barrientos-Durán, A., Fuentes-López, A., de Salazar, A., Plaza-Díaz, J., García, F., 2020. Reviewing the Composition of Vaginal Microbiota: Inclusion of Nutrition and Probiotic Factors in the Maintenance of Eubiosis. Nutrients Volume 12, P. 419. https://doi.org/10.3390/nu12020419 41. Sedghi, L., DiMassa, V., Harrington, A., Lynch, S. V., Kapila, Y.L., 2021. The oral microbiome: Role of key organisms and complex networks in oral health and disease. Periodontol. 2000 Volume 87, P. 107–131. https://doi.org/10.1111/prd.12393 42. Milo, I., Bedora-Faure, M., Garcia, Z., Thibaut, R., Périé, L., Shakhar, G., Deriano, L., Bousso, P., 2018. The immune system profoundly restricts intratumor genetic heterogeneity. Sci. Immunol. Volume 3. https://doi.org/10.1126/sciimmunol.aat1435