A new and convenient method for qualitatively and quantitatively evaluating the physiological activity of therapeutic or beneficial products based on natural matrices in terms of benefit / risk in comparison with other products with pharmacological or nutraceutical activity based on synthetic molecules.

Abstract

Provided are new predictive in vitro or ex vivo methods for defining benefit / risk scores for therapeutic or beneficial products, methods based on data obtained from in vivo experiments for defining benefit / risk scores for therapeutic or beneficial products, and methods for selecting therapeutic or beneficial products for further development. [Solution] A computer-implemented method for providing a benefit / risk score for a therapeutic product of interest based on in vitro transcriptional data or ex vivo transcriptional data, comprising providing a sample of a biological substrate representing a disease state or pathological condition to be treated by the product, extracting RNA from the sample, performing pathway enrichment analysis and functional analysis of transcriptome data from the extracted RNA, calculating a benefit score, and calculating a risk score.

Description

[Technical Field]

[0001] The present invention addresses a conceptual gap in modern medicine and regulation that focuses on classical API-based products rather than new therapeutic or beneficial compositions made from 100% natural substances with physiological (as opposed to pharmacological) mechanisms of action.

[0002] Modern medicine is evolving toward increasing specialization, with therapeutic products centered around isolated or synthesized active pharmaceutical ingredients (APIs) and clinical approaches increasingly removed from the holistic view of the patient. The organism is now often interpreted through the lens of specialized subsystems (organs, functions, receptors), losing sight of the systemic interconnections. This specialization has shaped not only treatment strategies but also the entire regulatory framework governing the development and approval of therapeutic or beneficial products.

[0003] In particular, when exclusive, selective, and natural matrices, appropriately processed by specific processes and methods, are used to create final products intended for therapeutic or adjuvant purposes, such as restoring or adjuvanting organisms in restoring healthy physiological states, developers of such new products face several regulatory challenges. Indeed, developers of natural matrix-based therapeutic or beneficial products face significant regulatory obstacles, primarily because the current pharmaceutical framework is designed for single-compound drugs rather than complex, multi-component natural formulations. Regulatory bodies such as the U.S. Food and Drug Administration (FDA), the European Medicines Agency (EMA), and other global authorities require rigorous characterization, reproducibility, and clinical validation (e.g., standards that are difficult for plant-derived therapies to meet due to their natural variability, multi-target effects, and the impossibility of defining specific therapeutic active ingredients). Standard drug approval pathways, such as the U.S. New Drug Application (NDA) or European Marketing Authorization (MA), often require clearly defined active ingredients, precise dosing, and distinct pharmacokinetics, which are inconsistent with the synergistic and emergent properties of plant matrices.

[0004] The entire regulatory system, including methods for assessing the benefit / risk profile of new therapeutics, is built around this reductionist model. In current practice, benefit / risk assessments are generally qualitative in nature, based on clinical trial results, therapeutic efficacy, and side effects observed in patient populations, and are only subsequently refined through actual post-marketing data. In particular, there are no standardized methods for predictive early-stage benefit / risk assessments based on in vitro or ex vivo data. There are also no methods for interpreting this data through probabilistic and holistic models capable of generating measurable quantitative outputs, let alone diagnostic methods capable of identifying and intercepting such parameters (e.g., complete IVD procedures).

[0005] As a result, many plant-based products are classified as botanicals, dietary supplements, or traditional medicines, limiting their ability to make therapeutic claims or receive full drug approval, as well as the development of innovation in this field. Furthermore, one major gap in the current regulatory landscape is the lack of standardized benefit / risk assessment models tailored to multi-component natural matrices. Traditional risk assessments are based on dose-response relationships, toxicity thresholds, and drug-drug interaction studies, all of which are well-suited to synthetic molecules but fail to capture the overall synergistic effects of plant-based therapies. Conversely, the full range of benefits encompassing immune modulation, multi-pathway interactions, and adaptive physiological responses is also difficult to quantify and compare to conventional drugs under current guidelines.

[0006] This reinforces the need for a new assessment paradigm that reflects an integrated, systems-based view of human biology, one that has so far not been embraced by contemporary regulations shaped by increasingly fragmented medical thinking.

[0007] Another significant challenge is the requirement for Good Manufacturing Practices (GMP) to ensure batch-to-batch consistency, a major issue when dealing with natural (e.g., botanical) extracts, which are subject to geographic, seasonal, and genetic variations. Furthermore, natural matrix-based product developers also deal with intellectual property (IP) complexities. This makes it difficult to secure investment for research and development as companies struggle to establish exclusivity for their innovations. Furthermore, compliance with global safety regulations, such as the European REACH (Registration, Evaluation, Authorization, and Restriction of Chemicals) or the Chinese Traditional Herbal Medicine Registration, adds another layer of complexity, often requiring lengthy toxicology studies, clinical trials, and detailed pharmacovigilance programs. These regulatory barriers significantly delay downmarket entry, increase costs, and create uncertainty for innovators in the plant-based therapeutics field. On the other hand, the public is increasingly demanding a shift towards naturalness under the principles of One Health (which recognises the interconnectedness of human health, animal health and environmental health), and therefore demanding products whose every stage of production falls within said principles and which do not allow the use of artificial forces or substances.

[0008] Indeed, in the field of the present invention, any therapeutic or beneficial product of interest, i.e., a product that contains or consists of one or more natural matrices, maintains its natural intelligence, i.e., an imprint of the biological domain to which each component of the product belongs, thereby maintaining a network that can be interconnected and recognized with other networks, whether natural or artificial, i.e., original natural networks that have acquired a degree of artificiality through their interaction with artificial components. This interconnection is considered to be fundamental for rebalancing any disturbance in the network of events active in each interacting biological system.

[0009] Each network of each natural matrix contained in the product contributes to the formation of the network of the final matrix of the product of interest and can be defined as a UVCB substance (i.e., a substance of unknown or variable composition, a complex reaction product, or a biological material) according to the REACH (Registration, Evaluation, Authorization and Restriction of Chemicals) definition, since it is a processed product with respect to its self-assembly properties, which cannot be determined or verified based on small molecule chemistry protocols.

[0010] Each network is characterized by the establishment of connections within the matrix of the final product and within the physiological effects exerted by the product on the recipient organism. Product validation of this type of product can be performed and confirmed using probabilistic models based on the association between the preservation of physiological activity profiles and descriptors of the matrix itself, generated using multiple biophysical analysis systems, including spectroscopy (NIR and other techniques), mass spectrometry, and paper or X-ray crystallography (fractal measurements), as disclosed in Japanese Patent No. 7616724 and PCT / IB2024 / 054526. Indeed, while useful, traditional molecular chemical definitions of the individual substances contained in a material cannot be used to validate this type of product, as they do not represent its overall efficacy and quality.

[0011] The choice of matrices intended for administration must be verified according to the updated and specific current taxonomic criteria of the animal, plant and mineral kingdoms. When used in combination with natural physical phenomena, the relationship between action and efficacy may need to be verified, taking into account acoustic effects (musical or other forms) and effects in the wave-particle field, including those of a quantum nature.

[0012] At the current state of the art, it is not always possible to outline a fully explained mechanism of action; however, it is possible to examine the actions and reactions at the interconnections of the respective networks, which have already been verified at the biophysical level.

[0013] The present invention aims to enable the selection and provision to humans in need thereof of new therapeutic or beneficial entities or products, as well as systems that allow the rebalancing, activation or restriction of physiological functions in specific metabolic states of organisms, which are always in a state of continuous transformation.

[0014] The preparations thus conceived are capable of rebalancing the psychoneuroendocrine immune system, which is considered as the single system that governs and controls all other systems.

[0015] This invention contributes to a new cutting-edge technology that goes beyond alchemical techniques in the medical field, whose origins can be traced back to the early 16th century, and returns products and processes to the conceptual One Health goal already mentioned. It proposes a new deconvolution of artificial techniques and naturally self-assembling materials, recognizing existing rules or finding new ones to ensure the formation of verifiable entities, primarily based on the concept of verifying their effects and activity on other organisms. The latter are continuously changing organisms that require evaluation of their physiological state within defined intervals, a concept now included in personalized medicine. This invention fits the concept of science, understood as a set of knowledge that can demonstrably verify the effects of theoretical mechanisms of action. Today, these methods are applied to establishing interconnections between all forms of life, in a context where technological innovation advances at such a pace that it risks damaging the interconnections between human-generated (artificial) intelligence and nature.

[0016] The activity of therapeutic or beneficial natural matrix-based products is currently not covered by the state of the art, and therefore, under the One Health concept, the entire product cycle from the end user to the relevant societal context needs to be considered.

[0017] The operating paradigm within which the present invention is formulated is referred to herein as "Bios Physiological Health."

[0018] This paradigm aims to introduce innovative approaches to the medical technology field for the treatment and self-management of health using natural matrices alone or in combination, rebalancing the normal physiological state of various organisms, including humans, through endogenous physiological effects induced by the products. The problem is to identify, select, and assemble natural entities with newly emerging properties that can be verified by the physiological mechanism of action of the final product and other methods evolved in recent decades.

[0019] The reading of the context by both technoscientific and anthropological norms integrated in their transversality constitutes the basis of the proposed invention. Some of the properties of each matrix part of the product may already be known, but the newly emerged properties of the new composition are unexpected.

[0020] Of particular relevance is the role of determining the genetic and epigenetic aspects that determine networks representing natural matrices and their interpretation at their specific isotopic abundance levels.

[0021] To fulfill the paradigm of Bio-Physiological Health, each stage of processing, from the selection of regenerative materials to agricultural and industrial stages and methods of use, must preserve as much as possible the inherited integrity of the native programming inserted into each creative entity's natural intelligence, at least as far as is known on a global scale. It will be essential to validate matrices derived from similar epigenetic realities, recognized as reference standards for specific emerging metabolic characteristics of other organisms, including humans. For example, one of the factors negatively affecting epigenetic differentiation is represented by different soil conditions, along with their 24-hour, monthly, and yearly variations. To preserve the properties of natural systems, which are the only ones that can claim physiological interconnections with the whole product, it is impossible to use substances derived from alchemical processes such as distillation or other synthetic or semi-synthetic processes, or products derived from genetically modified or genetically altered organisms. A new interpretation of the mystery of natural programming responsible for the evolution of organic and inorganic life is needed. The recent establishment of scientific evolution makes it possible to reposition the understanding of the origins of progress based on reductionist determinism on the basis of the development of alchemical processes from the beginning of the 16th century, which, together with Paracelsus in medicine, marked the beginning of the current evolutionary process known as the Anthropocene.

[0022] The term Anthropocene refers to the current stage of human evolution and can be traced back to different eras. When considered in the context of this invention, the only significant date is 1492, which marks the end of the Humanistic / Neoplatonic period of the Early Renaissance. This period was politically represented by Cosimo the Elder and Lorenzo de'Medici, along with artists and scientists such as Piero della Francesca, Luca Pacioli, Leonardo da Vinci, and Dürer. In the 16th century, alchemical research, which was considered the possibility of humans controlling nature, continued to develop to this day under the aegis of artificial intelligence, as opposed to natural science, with the aim of improving the creation of natural phenomena, so that "man may have dominion over all creation."

[0023] The year 1492 is symbolic, marking the deaths of Lorenzo de' Medici and Piero della Francesca, while Columbus discovered America. The human species abandoned the Neoplatonic path of the 15th century and followed the Judeo-Catholic path, adapting the alchemical practices of Paracelsus to medicine, marking the transition to Renaissance Mannerism in the 1500s and leading, to this day, to the full-scale and irreversible extinction of its inventors for the sixth time.

[0024] The present invention has demonstrated the feasibility of industrial discoveries made in the medical field, but is in principle adaptable to any field of production and is intended to address changes in evolutionary paradigms. The inventors often speak of preserving biodiversity without addressing the real problem of billions of tons of exogenous, non-biodegradable, artificial materials released into the planetary system, a problem that is clearly obscured, while the "carpe diem" approach trumps the sense of species survival.

[0025] This invention is presented primarily in the context of a patent, with the hope of opening a new field of research exploring and sharing natural intelligence, rather than artificial intelligence, which is unlikely to halt or slow the sixth extinction or build the foundation for alternative advances to current ones. Inventor Valentino Mercati, along with his collaborator Jacopo Lucci, has chosen the path of natural research that may be useful for biological systems, developing knowledge in agricultural and industrial production systems for over 40 years and filing numerous patent applications following this operational strategy. Previously filed patents related to the methods of the invention are essentially based on instrumental and diagnostic readouts based on chemistry-related principles of linking physiological actions with newly emerged properties of natural matrices and the innate defenses of individual organisms to which they are interconnected.

[0026] The analyses that inspired the approach disclosed herein were unthinkable just a few decades ago, due to the technical inability to read the genetic and epigenetic information written into the cells of every living organism, as well as the role of atomic isotope differentiation in molecular self-assembly and the interconnection of every single / individual with the "universe." The conceptual difficulty of moving from the reassuringly controlled parameters of molecular artificiality (at least partially purified and linked by powerful thermodynamic forces that allow strong bonds, such as covalent bonds, that act to a reduced extent on the molecules of other organisms) to natural matrices that are by definition mysterious and still considered therapeutically unreliable today is enormous.

[0027] If, after five centuries of alchemical reductionism, a new interpretation of the invention is needed for the new medical status quo, this interpretation must unite the most distant concepts and processes in a single field of application. This, as already mentioned, is due to an ideological legacy that calls into question the human condition: was the human species, insofar as inventors can assume that it controlled creation, produced by an original vital intelligence, like all other species, for the purpose of life itself, or was it experimentally endowed with capacities distinct from those of other organisms already suitably inserted into creation in order to constitute a new ecological niche in the service of the universe?

[0028] The answer to this dilemma does not arise in this invention: humanity must return to the Neoplatonic thought of the early Renaissance, and the experimental duality of the human species must be liberated from the spirit of domination in order to share its unique capacities within the universe with all of creation. Humanity needs to reconsider Leonardo da Vinci's warning, "Man can only procreate himself..." and reflect on the depressing thoughts of sensible figures like Piero della Francesca, Luca Pacioli, and Dürer regarding the impossibility of understanding and expressing the beauty of creation and deciphering its mysteries.

[0029] The time has come to acquire new research centers in molecular and cell biology, with an essential focus on bioinformatics and the new physical sciences. Today, the inventors can base their research strategies and socio-economic applications on new therapeutic fields, especially in the field of complex and / or chronic degeneration, where the restoration of metabolic balance in organisms disturbed naturally or artificially is already an integral part of the future.

[0030] The present invention redefines the state of the art of medicine by introducing a holistic assessment model consistent with personalized medicine and systems biology. Unlike reductionist pharmacological approaches that isolate active ingredients, the present invention recognizes the self-organizing complexity of natural systems.

[0031] By establishing a rigorous scientific basis for the evaluation of natural matrix-based therapeutics, this invention fills a critical gap in current medical research. It addresses the limitations of medical and regulatory frameworks built around highly specialized single-compound pharmacology and lacks tools for systematically evaluating complex, multi-target products. It offers a new paradigm for benefit / risk assessment that integrates biophysical analysis, probabilistic modeling, and physiological impact assessment, introducing for the first time quantitative and predictive methods based on in vitro or ex vivo data that can support early-stage decisions and restore scientific validity to a holistic approach. This ensures that natural, rather than artificial, intelligence will guide future advances in medicine.

[0032] According to current procedures for assessing benefit / risk scores for medical products, regulatory agencies such as the U.S. Food and Drug Administration (FDA), the European Medicines Agency (EMA), and the World Health Organization (WHO) use rigorous benefit / risk assessment frameworks to evaluate medical products before granting market authorization. These assessments rely on quantitative and qualitative methodologies to weigh a product's therapeutic benefits against its potential risks, ensuring that the overall impact on public health is favorable and justified.

[0033] In addition to regulatory approval, many countries also require health technology assessment (HTA) as part of their decision-making process for pricing, reimbursement, and market access. HTA is a multidisciplinary evaluation framework that considers not only clinical effectiveness and safety, but also the economic, social, and ethical implications of medical products.

[0034] Standard benefit / risk assessment methods typically include:

[0035] Analysis of clinical trial data: Safety and efficacy are assessed through randomized controlled trials (RCTs) that provide statistical comparisons of investigational products against placebo or existing standard treatments.

[0036] Pharmacokinetic and pharmacodynamic (PK / PD) modeling: Determining how a drug is absorbed, distributed, metabolized, and excreted, along with its biological effects at different concentrations.

[0037] Toxicology and safety profiling: Identifying potential adverse effects, contraindications and drug interactions through preclinical (animal and in vitro) and clinical (human) studies.

[0038] Structured decision-making approaches: Use models such as multi-criteria decision analysis (MCDA) and quantitative benefit-risk models (QBRM) to systematically balance benefits and risks.

[0039] Standardized Risk Management Plans (RMPs) and post-marketing surveillance: Ensuring ongoing monitoring of safety signals through pharmacovigilance programs.

[0040] Health economics and cost-effectiveness analysis (HTA): Assessing whether a treatment provides good value for its costs by comparing quality-adjusted life years (QALYs) and other economic measures.

[0041] Risk Management Plans (RMPs) and Post-Marketing Surveillance: Ensuring ongoing monitoring of safety signals through pharmacovigilance programs.

[0042] These methodologies are highly structured and reproducible and suitable for synthetic drugs and well-defined biologics, where individual active pharmaceutical ingredients (APIs) can be precisely isolated, characterized, and administered.

[0043] These procedures cannot be directly applied to therapeutic or beneficial products based on natural matrices; indeed, despite their effectiveness in conventional pharmaceuticals, these existing benefit / risk and HTA assessment procedures face significant limitations when applied to therapeutic or beneficial products based on natural matrices, primarily due to the inherent complexity and variability of natural substances.

[0044] Unlike conventional drugs, which have a single, clearly defined active compound, natural matrices contain multiple bioactive compounds that interact with each other as a network system and synergistically contribute to their therapeutic effect; therefore, they lack a single active ingredient. While modern medicine, as well as modern regulatory science, is reductive in its focus on identifying single molecular entities responsible for efficacy, natural matrix-based products exert their effects through complex multi-target interactions. Current clinical and pharmacological models are not designed to capture or quantify these expression characteristics, and standard PK / PD modeling is inadequate. Furthermore, natural matrices are inherently subject to biological, environmental, and genetic variability, resulting in batch-to-batch compositional differences.

[0045] Pharmaceutical regulations and HTA models require highly standardized compositions that are precisely administered, which natural matrix-based products cannot easily provide without altering their inherent properties. Current analytical techniques struggle to measure efficacy without relying on a single molecular marker, complicating regulatory and HTA approval. Natural matrices often contain hundreds of chemical components, some of which may have nonlinear dose-response relationships or adaptive physiological effects that are difficult to predict using standard toxicology models. Traditional safety studies are designed for synthetic compounds with distinct pharmacokinetics, making it difficult to assess the risks of natural matrices, where multiple components dynamically interact with the body's metabolism.

[0046] Natural matrix-based products do not follow a linear dose-response curve, and although their benefits can be long-term and preventative, quantifying their effectiveness in economic terms is difficult due to the fact that cost-effectiveness models prioritize immediate, measurable results.

[0047] Essentially, the art needs new benefit / risk and HTA assessment paradigms to bridge the gap between regulatory and HTA requirements and the reality of natural matrix-based therapeutics that provide new benefit / risk and health technology assessment models.

[0048] Until such technologies are developed and accepted by regulatory agencies and HTA bodies, natural matrix-based therapeutics will remain at a disadvantage because they cannot be adequately evaluated using current pharmaceutical benefit / risk and HTA models.

[0049] Ultimately, the present invention forms the basis of a new paradigm for therapeutic development that is not only consistent with One Health and natural intelligence principles, but also introduces a prospective model for benefit / risk assessment.

[0050] By enabling quantitative, early-stage, and probabilistic assessment of both beneficial and potentially adverse physiological effects derived from in vitro or ex vivo transcriptional data, the present invention addresses fundamental limitations of current regulatory and scientific practice. It provides a specific, measurable alternative to the common qualitative methods designed around single-compound synthetic drugs, thus restoring scientific integrity and regulatory feasibility to complex natural products.

[0051] This approach does not simply complement the current system, but redefines it, providing a pathway to safer, more effective, and truly integrated care. Summary of the Invention

[0052] The present invention relates to a novel approach for quantifying the differential benefit / risk profile of therapeutic or beneficial products for a given pathological area of ​​interest. This approach takes into account transcriptional effects assessed in an in vitro or ex vivo context. The present invention introduces an analytical method designed as a discovery tool suitable for early research and development (R&D) stages.

[0053] The method is based on an integrated approach utilizing transcriptome data and advanced biological pathway analysis tools (e.g., Ingenuity Pathway Analysis, e.g., IPA) to identify gene expression changes and the resulting modifications in associated biological activity, addressing both therapeutic efficacy and potential adverse effects.

[0054] Specifically, the generated transcriptional profiles are analyzed to determine differentially expressed genes (DEGs) and key biological functions associated with the pathology of interest and actions that can counteract specific side effects caused by standard reference drugs for each treatment, thereby enabling the calculation of benefit / risk scores for different therapeutic or beneficial products based on a conceptual integration approach of transcriptional profiles obtained from in vitro or ex vivo studies. The method of the present invention is particularly suitable for therapeutic or beneficial products that include natural matrices.

[0055] The benefit / risk ratio is an important parameter that is very useful in understanding the differences between such important characteristics of therapeutic options or beneficial options. Nevertheless, such parameters are currently difficult to resummarize into a single objectively calculated numerical parameter that can facilitate a first approach to comparing therapeutic solutions. Here, the inventors provide a general method that can accurately predict and objectively compare the risk / benefit ratio of a new therapeutic or beneficial product of interest for the treatment of a given disease state or pathological condition with a known reference drug for the treatment of the same disease state or pathological condition, and the method of the present invention can be applied solely based on in vitro or ex vivo transcriptome data.

[0056] Therefore, the object of the present invention is to 1. A computer-implemented method for providing a benefit / risk score for a therapeutic product of interest based on in vitro transcriptional data or ex vivo transcriptional data, comprising the steps of: 1. Performing Transcriptomic Analysis by: 1.1 providing a sample of a biological substrate representative of the pathology or pathological state to be treated by said product; a. Treating one or more of said samples (hereinafter referred to as sample a) with said product; b. treating one or more of said samples (hereinafter referred to as sample b) with a reference agent for treating said disease state or pathological condition; and c. one or more samples of said biological substrate (hereinafter referred to as sample c) are used as relevant controls; 1.2 Extracting RNA from each of the samples a, b, and c; 1.3 Perform transcriptome raw data analysis from the RNA extracted in 1.2 to identify differentially expressed genes (DEGs) in each of samples a and b relative to sample c and their expression fold changes relative to sample c, thereby obtaining fold change values ​​for each of the DEGs; 1.4 Perform pathway enrichment analysis and functional analysis of the transcriptome data for the list obtained in 1.3, thereby obtaining a numerical value representing the variation in magnitude and direction of biological activity associated with the differential expression of said DEGs in each of samples a and b normalized to sample c; 2. Determine the benefit score by: 2.1 Selecting from among the biological activities in 1.4 a biological activity relevant to the therapeutic indication of said product; 2.2 Convert the relevant biological activity values ​​obtained in 1.4 into absolute values 2.3 summing said absolute values ​​of each of said biological activities, thereby obtaining a benefit score for the tested product; 3. Determine the risk score by: 3.1 Provide a list of known side effects associated with the reference drug for the treatment of said condition 3.2 Determine the biological activity associated with said side effects by pathway and functional analysis; 3.3 Construct an in silico model of risk using pathway and functional analysis by establishing the relationship between the gene expression patterns obtained in point 1.3 and the biological activity determined in point 3.2 for samples a and b; 3.4 The relevant fold change values ​​of each DEG in samples a and b obtained in 1.3 are input into the risk in silico model obtained in point 3.3, and pathway analysis and functional analysis are used to obtain data representing the direction and magnitude of regulation of each biological activity determined in point 3.2, and the data are converted into corresponding absolute numerical values; 3.5 The positive values ​​of the numerical values ​​obtained in 3.4 are summed to obtain a final value representing the risk score of the inspected product, and if the final value is less than a predetermined minimum positive value, the final value is automatically corrected to the minimum positive value. 4. Provide the value of the benefit / risk score of the product of interest as the ratio of the value of the benefit score obtained in point 2.3 to the value of the risk score obtained in point 3.5.

[0057] Another object of the present invention is a method for selecting one or more new therapeutic or beneficial products for clinical development, wherein benefit / risk scores for one or more therapeutic or beneficial products of interest for the treatment of a given disease state or pathological condition and for a reference drug for the treatment of said disease state or pathological condition are provided by carrying out the steps defined in the present specification and claims (see method above), and wherein each of said at least one therapeutic or beneficial product of interest is selected for clinical development if said benefit / risk score value provided for said therapeutic or beneficial product of interest is equal to or greater than said benefit / risk score value provided for the reference drug.

[0058] A further object of the present invention is a computer-implemented method for producing a therapeutic or beneficial product based on in vitro or ex vivo transcriptional data, comprising determining a benefit / risk score thereof, the method comprising the steps of: 1. Performing Transcriptomic Analysis by: 1.1 providing a sample of a biological substrate representative of the disease state or pathological condition to be treated by said product; a. Treating one or more of said samples (hereinafter referred to as sample a) with said product; b. treating one or more of said samples (hereinafter referred to as sample b) with a reference agent for treating said disease state or pathological condition; and c. one or more samples of said biological substrate (hereinafter referred to as sample c) are used as relevant controls; 1.2 Extracting RNA from each of the samples a, b, and c; 1.3 Perform transcriptome raw data analysis from the RNA extracted in 1.2 to identify differentially expressed genes (DEGs) in each of samples a and b relative to sample c and their expression fold changes relative to sample c, thereby obtaining fold change values ​​for each of the DEGs; 1.4 Perform pathway enrichment analysis and functional analysis of the transcriptome data for the list obtained in 1.3, thereby obtaining a numerical value representing the variation in magnitude and direction (i.e., up-regulation or down-regulation) of the biological activity associated with the differential expression of said DEGs in each of samples a and b normalized to sample c; 2. Determine the benefit score by: 2.1 Selecting from among the biological activities in 1.4 a biological activity relevant to the therapeutic indication of said product; 2.2 Convert the relevant biological activity values ​​obtained in 1.4 into absolute values 2.3 summing said absolute values ​​of each of said biological activities, thereby obtaining a benefit score for the tested product; 3. Determine the risk score by: 3.1 Provide a list of known side effects associated with the reference drug for the treatment of said condition 3.2 Determine the biological activity associated with said side effects by pathway and functional analysis; 3.3 Construct an in silico model of risk using pathway and functional analysis by establishing the relationship between the gene expression patterns obtained in point 1.3 and the biological activity determined in point 3.2 for samples a and b; 3.4 The relevant fold change values ​​of each DEG in samples a and b obtained in 1.3 are input into the risk in silico model obtained in point 3.3, and pathway analysis and functional analysis are used to obtain data representing the direction and magnitude of regulation of each biological activity determined in point 3.2, and the data are converted into corresponding absolute numerical values; 3.5 The positive values ​​of the numerical values ​​obtained in 3.4 are summed up to obtain a final value representing the risk score of the inspected product, and if the final value is less than a predetermined positive minimum value, the final value is automatically corrected to the predetermined positive minimum value.

[0059] The methods of the present invention are advantageously applicable to natural matrix-based therapeutic or beneficial products that cannot be defined in terms of a single active ingredient with a measurable dose-response relationship, as demonstrated herein and in the examples.

[0060] term Unless otherwise defined herein, scientific and technical terms used in connection with the present invention shall have the meanings commonly understood by those skilled in the art. Further, unless otherwise required by context, singular terms shall include pluralities and plural terms shall include the singular.

[0061] At any point in this specification or claims, the words "comprising" or "comprise(s)" may be replaced with "consisting of" or "consist(s) of."

[0062] Benefit / risk assessment, in the current state of the art, is merely a qualitative measure used to assess the overall value of a beneficial or medical product, intervention, or procedure by comparing its therapeutic or beneficial benefits with its potential risks or harms. This score is of great importance in regulatory decision-making, health technology assessment (HTA), and clinical practice, helping to determine whether a procedure provides a favorable balance between its positive effects and possible adverse effects. Therefore, this specification defines the benefit / risk assessment as a measure of the overall value of a beneficial or medical product, intervention, or procedure by comparing its therapeutic or beneficial benefits with its potential risks or harms. Score This term provides a qualitative and Concerning quantitative measures.

[0063] In this application, a "natural matrix" refers to a material consisting of a network represented by a wide range of components / ingredients obtained (e.g., extracted) directly from a member of the natural world or its naturally occurring parts (i.e., from a natural source) without significant processing or synthetic alteration. "Without significant processing or synthetic alteration" means that no denaturing process is used to obtain the matrix from the source. In other words, the natural source is processed only by manual, mechanical, or gravitational means, such as dissolution in water or other naturally occurring solvents, such as water or water-alcohol solutions; by flotation; by extraction with water or other naturally occurring solvents; by steam distillation; or by heating only to remove water or any other naturally occurring solvent; or by any means and under conditions that exclude the member of the natural world "per se," i.e., extracted from air untreated. In particular, according to the present invention, a natural matrix is ​​a 100% natural and biodegradable material consisting of natural components that have not been modified by the process for producing the matrix from the starting material, without the intentional addition of synthetic products along the entire process. Herein, 100% biodegradability is considered "readily biodegradable" according to the OECD biodegradability test. These characteristics ensure the maintenance of the matrix effect imparted to the matrix by the structural interactions (material interactions) of its components and the presence of functional interactions (non-material interactions) that become apparent upon exposure of a biological system to the natural matrix. In other words, a natural matrix or a mixture of natural matrices is a material obtained from entities that naturally self-assemble and are processed to preserve their natural biophysical properties that determine their physiological interactions with other organisms, such as human organisms. These newly emerged properties may be expressed by contributing to the rebalancing of metabolic processes or states of the recipient organism and / or some organs or tissues, along with physiological effects activated in each specific situation. According to the present invention, the natural matrix may be derived from materials obtained from any source in the kingdoms of life, namely, Monera, Protista, Fungi, Plantae, and Animalia.Thus, the term encompasses plant natural matrices, animal natural matrices, fungal natural matrices, protist (archaeal or bacterial) natural matrices, and monera natural matrices. Natural matrices can also include natural inorganic materials, such as minerals obtained from natural raw materials. Synonyms for natural matrix or one or more natural matrices herein are "composite natural system" or "natural material," as defined below.

[0064] Examples of naturally occurring parts of an organism may be represented by, for example, roots, leaves, bark, fruits, flowers, plants or sections thereof, organs, tissues.

[0065] In any part of this specification the general term natural matrix may be replaced by: Plant natural matrices or natural matrices obtained from plants, Animal natural matrices or natural matrices obtained from animals or animal products such as eggs or milk, Fungal natural matrices or natural matrices obtained from fungi, Protist natural matrices or natural matrices obtained from protists, Monera natural matrix or natural matrix obtained from Monera, or plant material and / or extracts, extracts from animal tissues or organs, fungi and / or fungal extracts, or mixtures thereof, where the extraction method does not involve a denaturing step (e.g., temperature or the use of denaturing solvents).

[0066] Plants are synonymous with herbs.

[0067] The term "natural" matrix emphasizes that it retains the integrity and complexity of the component / component network as in the original natural source due to the absence of denaturing treatments to obtain it. Thus, natural matrix does not encompass naturally occurring compositions enriched in specific molecules that have been artificially synthesized or isolated from natural sources. Furthermore, natural matrices can only be obtained by processes that do not involve extensive treatment or chemical modification, isolation, purification, or molecular extraction.

[0068] Due to the supramolecular self-assembly of the components / ingredients of the natural matrix and the existence of functional interactions between them, the entire matrix behaves as a complex network that does not interact with a single target molecule, but with a network of recipients (also organized as a network) in the recipient organism. Thus, the interaction of the natural matrix recipient organism is not the result of point-to-point interactions, as with common pharmaceutical APIs, but rather the result of a (i.e., matrix)-"recipient" network (i.e., the organism to which the matrix is ​​administered) interaction.

[0069] The term natural matrix may be substituted for complex natural system in any part of the specification and claims.

[0070] Nowhere in this specification and claims can the term natural matrix be construed as a "natural product" per se; rather, a natural matrix is ​​a product obtained from a natural organism and processed therefrom (e.g., extracted) by techniques that do not substantially alter the biological structure and associated supramolecular and functional interconnections between components within said matrix, i.e., by techniques that do not employ denaturing techniques and do not involve additional isolated or synthetic molecules or classes of molecules.

[0071] As emerging properties according to the present specification and the art, this term defines a property of a natural matrix or material according to the present specification, i.e. a property that is not represented by the mere sum of the properties of each isolated component / ingredient of said matrix / material, but by both the functional and structural interactions between all components / ingredients of said matrix / material, which is also the result of supramolecular self-assembly of said components / ingredients within the matrix / material itself.

[0072] Thus, "emergent properties" refer to the technical effects, e.g., therapeutic or homeostatic-adjuvant properties (i.e., beneficial effects), that the interactions and relationships between components / components of a natural matrix have on a receiving living system. By definition, emergent properties are properties that are not immediately apparent or even predictable based solely on the individual properties of each component / component of the matrix. Instead, they "emerge" when all components / components of the matrix network interact with each other and with the receiving living system network in dynamic and complex ways. Emergent properties have been widely discussed in the art in various scientific and systems-oriented fields, including physics, chemistry, biology, and complex systems theory.

[0073] Thus, emergent properties are properties that cannot be predicted a priori by qualitative-quantitative knowledge of each component of a given composition or matrix, and therefore cannot be attributed to one or more specific APIs. Thus, while a multi-drug composition may exhibit unexpected synergistic effects, the properties of the composition are still attributable to the specific APIs and amounts thereof contained therein.

[0074] In the case of emerging properties characteristic of a natural matrix, the observed emerging properties cannot be replicated for a particular API and are maintained in different batches of a given matrix or in a given mixture of matrices, despite the different qualitative-quantitative composition of said batches (functional elasticity see below).

[0075] Synthetic herein has its conventionally accepted meaning in chemistry. Traditionally, in chemistry, the term "synthetic" refers to the origin or source of a material or substance. Synthetic substances or materials are produced by humans by artificial synthesis, i.e., laboratory chemical reactions that typically react simpler chemicals to produce more complex chemicals through processes that often employ different pathways, temperature conditions, pressure conditions, energy sources, and / or catalysts than those used by living organisms.

[0076] Examples: Synthetic substances or materials include plastics, pharmaceuticals, and many industrial chemicals. For example, nylon is a synthetic polymer made by chemical synthesis, and aspirin is a synthetic drug made by a specific chemical reaction.

[0077] The hallmarks of a disease or pathological or medical condition herein have the meaning conventionally used in the art. They can be easily identified by those skilled in the art in specialized databases, pathway and function databases, or disease-specific databases. Disease hallmarks are known to be indicators that can mark the progression or control of a given disease or pathological or pre-pathological condition, and together usually represent the general pathological state associated with a given condition. These hallmarks (also called "key indicators") are typically a set of features or patterns that physicians monitor over time to track the onset, progression, or regression of a particular disease. In summary, disease hallmarks are defining characteristics or properties whose alterations indicate a given pre-medical or medical condition and aid in its identification, diagnosis, monitoring, and understanding. For example, for neurodegenerative diseases (NDDs), at least eight hallmarks of NDDs are known in the art: (pathological protein) aggregation, synaptic and neuronal network (dysfunction), (abnormal) proteostasis, cytoskeleton (abnormal), (altered) energy homeostasis, DNA and RNA (defects), inflammation (increased), and neuronal cell death (increased). In cancer research, the hallmarks of cancer are a set of characteristic properties commonly found in cancer cells. These hallmarks include (sustained) proliferative signaling, (evasion of) growth suppressors, (resistance to) cell death, (enabling) replicative immortalization, (inducing) angiogenesis, and (activating) invasion and metastasis.

[0078] Disease hallmarks, parameters (e.g., biomarkers) associated with the hallmarks, one or more biological activities associated with the hallmarks, etc., provide a framework for studying a disease or pathological or medical condition using an integrated / holistic approach.

[0079] Hallmarks of an altered physiological state typically include observable changes in various aspects of bodily function, which may be manifested through symptoms, signs, or laboratory findings.

[0080] Altered physiological states typically reflect a disruption of the body's homeostatic mechanisms, resulting in deviations from normal physiological parameters. These imbalances may involve changes in temperature regulation, fluid and electrolyte balance, acid-base balance, glucose metabolism, or other regulatory processes.

[0081] Overall, the hallmarks of altered physiological states provide valuable clues for healthcare providers to identify underlying causes, assess severity, and guide appropriate interventions to restore normal function and promote recovery.

[0082] A reference drug is a drug commonly selected or chosen as the standard or preferred treatment for a particular medical condition or disease. It is often established based on factors such as its effectiveness, safety profile, cost, and clinical experience. A reference drug serves as a benchmark for comparison with other drugs, especially when evaluating generic versions, new treatments, or alternative therapies. It is typically the first-choice drug recommended by medical guidelines or healthcare providers to treat a particular condition. A reference drug according to the present invention is a drug listed in the official list of approved and reference drugs issued by regulatory agencies such as the FDA and EMA, which are used as comparators in clinical trials and for general drug approval.

[0083] As used herein, pathway function analysis or pathway enrichment and functional analysis (IPA-like analysis) refers to a bioinformatics approach that interprets large-scale omics data (e.g., transcriptomics, proteomics, metabolomics) by: Identifying differentially expressed genes (DEGs) or molecules from experimental data (e.g., RNA-seq, microarray, mass spectrometry) and mapping these genes / proteins / metabolites to known biological pathways (e.g., signaling cascades, metabolic pathways); performing enrichment analysis to determine which biological processes or activities, molecular functions, or cellular components are most significantly affected; optionally inferring upstream regulators (e.g., transcription factors, cytokines, microRNAs) that may explain the observed expression changes; predicting biological effects (e.g., activation or inhibition of pathways, disease associations, drug interactions).

[0084] The Z-score is a useful statistical tool for measuring relative position within a data set, detecting outliers, and standardizing comparisons across different distributions. As used herein, the term Z-score (also called standard score) has its commonly accepted meaning in the art, i.e., a statistical measure that describes how far a data point is from the mean of a data set, expressed in terms of standard deviation. It allows values ​​from different distributions to be compared by standardizing them to a common scale. The Z-score is calculated by the following formula: Z=X-μ / σ where: Z = Z score X = individual data point μ = mean (average) of the data set σ = standard deviation of the data set [Brief explanation of the drawings]

[0085] [Figure 1] 1 is a flow chart of a procedural process for calculating a benefit / risk score based on transcriptional profiles obtained from in vitro or ex vivo experiments. [Figure 2] Flowchart of the procedural process for calculating benefit / risk scores based on transcriptional profiles obtained from in vivo experiments. [Figure 3]Annotated side effects of cisplatin and IPA with corresponding biological activities. Column "Hallmarks": anatomical and functional areas potentially affected by cisplatin toxicity. Column "Annotated side effects of cisplatin": list of side effects of cisplatin that have been reported and classified as "common". Column "IPA biological activity": side effects of cisplatin outlined according to the biological activity documented in the IPA. Column "Favorable trend": healthy trend of the selected biological activity (up-modulation ↑ or down-modulation ↓). [Figure 4] Calculation of benefit scores for EpigenAU / 11 and cisplatin based on transcriptional data obtained from in vitro and ex vivo studies without using any numerical importance factor (a), and calculation of benefit scores for EpigenAU / 11 and cisplatin by multiplying the obtained Z-scores by their corresponding numerical importance factors (b). [Figure 5] Calculation of risk scores for EpigenAU / 11 and cisplatin based on transcriptional data obtained from ex vivo studies. [Figure 6] Benefit / risk scores obtained for ex vivo treatment with EpigenAU / 11 and cisplatin after 6 hours using only the mandatory steps according to the description (A). Benefit / risk scores obtained for EpigenAU / 11 and cisplatin treatment after 6 hours using the mandatory and optional steps according to the description (i.e., multiplying the obtained Z-score by the corresponding numerical coefficient of importance) were performed (C). Calculation of the fold change of the score obtained with EpigenAU / 11 relative to the score obtained with cisplatin: EpigenAU / 11 scores are twice as high as cisplatin, whether calculated using only the mandatory steps (B) or both the mandatory and optional steps (D). [Figure 7]Benefit / risk scores obtained by in vivo treatment with EpigenAU / 11 and cisplatin for an animal model (A). Risk (intended as side effects) was estimated by observing behavioral parameters (affecting the behavior itself, the locomotor system, muscle strength, and neuroreflexes) and potential weight loss of the animals (see table in Example 8.2). Alternatively, benefit (intended as therapeutic activity) was evaluated by taking into account the reduction in tumor mass size compared to untreated tumors (see table in Example 8.3). Calculation of the fold change of the score obtained with EpigenAU / 11 relative to the score obtained with cisplatin: the EpigenAU / 11 score is 3 times the score of cisplatin (D). [Figure 8] Annotated side effects of DIBASE and IPA with corresponding biological activity. Column "Hallmark": Anatomical and functional areas potentially affected by DIBASE toxicity. Column "DIBASE annotated side effects": List of DIBASE side effects reported and classified as "common". Column "IPA biological activity": Side effects of DIBASE outlined according to the biological activity documented in IPA. Column "Favorable trend": Healthy trend of selected biological activity (up-modulation ↑ or down-modulation ↓). [Figure 9] Calculation of benefit scores for Osteoredux and DIBASE based on transcriptional data obtained from in vitro and ex vivo studies without any numerical significance factor (a), and calculation of benefit scores for Osteoredux and DIBASE by multiplying the obtained Z-scores by their corresponding numerical significance factors (b). [Figure 10] Calculation of Osteoredux and DIBASE risk scores based on transcriptional data obtained from in vitro studies. [Figure 11]Benefit / risk scores obtained with in vitro treatment with Osteoredux and DIBASE, with only mandatory steps according to the description (A). Benefit / risk scores obtained with Osteoredux and DIBASE treatment after 6 hours using mandatory and optional steps according to the description (i.e., multiplying the obtained Z-score by the corresponding numerical coefficient of importance) were performed (C). Calculation of the fold change of the score obtained with Osteoredux relative to the score obtained with DIBASE: the Osteoredux score is 3-fold higher than DIBASE when only mandatory steps are used in the calculation (B), and 6-fold higher than DIBASE when both mandatory and optional steps are included (D). DETAILED DESCRIPTION OF THE INVENTION

[0086] As noted above, benefit / risk assessments are conceptually assessed based on qualitative data. The present invention instead relies on benefit / risk assessments. Score Risk is a quantitative and qualitative measure for assessing the overall value of a medical or beneficial product, intervention, or procedure by comparing its therapeutic or beneficial benefits with its potential risks or harms. It is therefore expressed as a ratio between the magnitude of benefit and the magnitude of risk, where benefit refers to the positive therapeutic or beneficial effects of a treatment, such as efficacy in disease management, symptom relief, improved survival, and improved quality of life. Risk, on the other hand, encompasses potential harms, including side effects, toxicity, long-term safety concerns, and contraindications.

[0087] In the art, a 4R approach is often applied to ensure comprehensive and adaptive assessment: review clinical data to assess benefits and risks, refine methodologies as new evidence emerges, reassess the benefit-risk balance over time, and transparently report findings to stakeholders. This iterative process increases the accuracy, reliability, and applicability of benefit / risk assessments and supports evidence-based decision-making in healthcare. The methods provided herein enable in vitro or ex vivo prediction of benefit / risk scores for therapeutic or beneficial products, which is highly advantageous because it allows for early, controlled, and cost-effective evaluation of product efficacy and safety before proceeding to animal models or clinical trials. By providing quantitative and merely qualitative insights into therapeutic effects and potential toxicities at the cellular level, such methods help refine candidate selection, reduce reliance on in vivo studies, and accelerate regulatory decision-making. This approach may result in more efficient drug development, minimize late-stage adverse events, and optimize patient safety while streamlining the path to clinical application and approval.

[0088] Indeed, the method of the present invention is particularly advantageous because it allows for a direct comparison between the tested product and a well-established reference drug. By applying the same standardized evaluation criteria to both (the product under test and the reference drug), the method provides a controlled, quantitative evaluation of efficacy and safety, allowing researchers to measure the benefits and risks of the tested product compared to the known profile of the reference drug.

[0089] This method applies to both synthetic and natural matrix-based therapeutic or beneficial products, but is particularly advantageous for the latter. Indeed, the benefits and risks of synthetic drugs are easier to assess even with classical methods, since they contain a single active ingredient with a measurable dose-response relationship (known API, known receptor). On the other hand, therapeutic or beneficial products based on natural matrices involve multiple components, synergistic effects, and compositional variability, making it difficult with classical methods to easily identify the specific toxicological regions of interest involved.

[0090] The present invention provides a new method for calculating benefit / risk scores for therapeutic or beneficial products, which is also applicable to natural matrix-based products. The method is based on an integrated approach that utilizes transcriptome data and advanced biological pathway analysis tools (e.g., Ingenuity Pathway Analysis (IPA)) to identify gene expression changes and associated alterations in biological activity, addressing both the therapeutic efficacy and potential adverse effects of the product.

[0091] The benefit / risk ratio is an important parameter that is extremely useful in understanding the differences between such important characteristics of treatment options. Nevertheless, such parameters are currently difficult to resummarize into a single, objectively calculated numerical parameter that can facilitate a first-order approach to comparison between treatment solutions.

[0092] The present invention provides methods that can accurately predict and objectify comparisons between the benefit / risk ratios of two therapeutic solutions based solely on in vitro and ex vivo transcriptome data. Furthermore, the present invention provides additional methods for streamlining and summarizing benefit / risk ratios into a single numerical parameter from data obtained by in vivo animal models (see tables in Examples 8.2 and 8.3).

[0093] State-of-the-art benefit / risk assessment of a drug involves comparing its observed positive clinical effects (benefits) with its observed negative clinical effects (risks). This process usually begins with clinical trials to evaluate efficacy and safety. Benefits are assessed based on the drug's ability to effectively treat or prevent symptoms. Risks are considered by identifying side effects, toxicity, and long-term effects. Data from preclinical studies, clinical trials, and post-marketing surveillance aid in this assessment. The drug's safety profile, the severity of side effects, and the severity of the condition being treated are all considered. Regulatory agencies such as the FDA evaluate the evidence before approval, considering whether the benefits outweigh the risks. Ongoing monitoring ensures continued safety after approval. If the risks outweigh the benefits, the drug can be withdrawn or its use restricted.

[0094] The present invention utilizes a known reference drug with known side effects as a product control to assess the benefit / risk profile of a new potentially useful drug or product. Score This provides a new method for determining

[0095] As described in the glossary, a reference drug for a disease (also called a standard of care or gold standard therapy) is an established medication that serves as a benchmark for efficacy and safety in treating a particular condition. It is typically the most widely accepted and prescribed treatment for a disease, supported by strong clinical evidence from large-scale trials and real-world data, and is usually used as a comparator in clinical trials of new drugs. For example, regulatory agencies such as the FDA and EMA use reference drugs to evaluate whether new drugs offer added value in terms of efficacy, safety, or tolerability. Therefore, those skilled in the art can easily identify reference drugs because regulatory agencies maintain official lists of approved and reference drugs used as comparators for clinical trials and general drug approval. For example, reference drugs are officially listed in the FDA's Orange and Purple Books, the EMA's EPARs, the WHO's Essential Medicines List, and various Health Technology Assessment (HTA) organizations and clinical guidelines, ensuring consistency, safety, and effectiveness in treatment recommendations and drug approvals worldwide.

[0096] The method of the present invention is particularly suitable for use in the predictive assessment of the benefit / risk ratio of a new product under evaluation when compared with the benefit / risk ratio obtained by the same method for a reference drug for the treatment of the same condition treated by the product under evaluation.

[0097] Accordingly, the present invention provides a computer-implemented method for providing a benefit / risk score for a therapeutic product of interest based on in vitro or ex vivo transcriptional data, the method comprising the steps of: 1. Performing Transcriptomic Analysis by: 1.1 providing a sample of a biological substrate representative of the disease state or pathological condition to be treated by said product; a. Treating one or more of said samples (hereinafter referred to as sample a) with said product; b. treating one or more of said samples (hereinafter referred to as sample b) with a reference agent for treating said disease state or pathological condition; and c. one or more samples of said biological substrate (hereinafter referred to as sample c) are used as relevant controls; 1.2 Extracting RNA from each of the samples a, b, and c; 1.3 Perform transcriptome raw data analysis from the RNA extracted in 1.2 to identify differentially expressed genes (DEGs) in each of samples a and b relative to sample c and their expression fold changes relative to sample c, thereby obtaining fold change values ​​for each of the DEGs; 1.4 Perform pathway enrichment analysis and functional analysis of the transcriptome data for the list obtained in 1.3, thereby obtaining a numerical value representing the variation in magnitude and direction (i.e., up-regulation or down-regulation) of the biological activity associated with the differential expression of said DEGs in each of samples a and b normalized to sample c; 2. Determine the benefit score by: 2.1 Selecting from among the biological activities in 1.4 a biological activity relevant to the therapeutic indication of said product; 2.2 Convert the relevant biological activity values ​​obtained in 1.4 into absolute values 2.3 summing said absolute values ​​of each of said biological activities, thereby obtaining a benefit score for the tested product; 3. Determine the risk score by: 3.1 Provide a list of known side effects associated with the reference drug for the treatment of said condition 3.2 Determine the biological activity associated with said side effects by pathway and functional analysis; 3.3 Construct an in silico model of risk using pathway and functional analysis by establishing the relationship between the gene expression patterns obtained in point 1.3 and the biological activity determined in point 3.2 for samples a and b; 3.4 The relevant fold change values ​​of each DEG in samples a and b obtained in 1.3 are input into the risk in silico model obtained in point 3.3, and pathway analysis and functional analysis are used to obtain data representing the direction and magnitude of regulation of each biological activity determined in point 3.2, and the data are converted into corresponding absolute numerical values; 3.5 Add up the positive values ​​of the numerical values ​​obtained in 3.4 to obtain a final value representing the risk score of the inspected product, and if the final value is less than a predetermined minimum positive value, the final value is automatically corrected to the minimum positive value; 4. Provide the value of the benefit / risk score of the product of interest as the ratio of the value of the benefit score obtained in point 2.3 to the value of the risk score obtained in point 3.5.

[0098] Thus, in the methods of the present invention, transcriptional data is generated in vitro or ex vivo, for example, using a suitable biological substrate that is a representative model of the disease state or pathological condition of interest.

[0099] Those skilled in the art can easily select suitable biological substrates to be treated with the product of interest or its relative reference drug from known in vitro or ex vivo models to generate the necessary transcriptional data according to the methods of the present invention. Non-limiting examples include cultured cells, primary cells, human or animal-derived cells from relevant tissues (e.g., lung epithelial cells for respiratory pathologies), primary cells from patients (e.g., cancer cell lines), immortalized cell lines, stem cell-derived models (e.g., iPSC-derived neurons for neurodegenerative treatment), organ-specific cell lines (e.g., HepG2 for liver metabolism research), organoids (e.g., organoids such as intestinal, brain, and kidney that mimic in vivo conditions), spheroids (e.g., 3D cancer or stem cell models), and microfluidic "organ-on-a-chip" systems that simulate tissue-level responses. By way of further non-limiting example, the ex vivo model may be a tissue-based system such as a tissue explant (organotypic slice), a human or animal tumor biopsy (e.g., for testing anti-cancer products), a liver or kidney slice, a brain slice, a patient-derived xenograft (PDX model), etc. Relevant controls according to the present invention are selected by those skilled in the art to represent a pathological condition without treatment with the product of interest or a reference agent.

[0100] The methods of the present invention are applicable to therapeutic or beneficial products generally, and the appropriate biological sample will vary mutatis mutandis depending on the therapeutic or beneficial purpose of the product being tested, as will be apparent from the present specification and examples. The present specification and examples demonstrate the applicability of the methods of the present invention to products with very different therapeutic purposes, and therefore, the methods should not be limited to specific pathological settings, as they can be easily applied to numerous pathological settings by those skilled in the art. Indeed, the methods of the present invention were developed to provide a general procedural protocol that allows developers to evaluate benefit / risk scores for different therapeutics or products and compare said scores to one of a select reference drug for said product. Furthermore, the methods disclosed herein are based on in vitro or ex vivo systems and provide rapid and reliable benefit / risk scores, thereby enabling developers to select the most promising candidates among new therapeutic or beneficial products under development. Reference drugs according to the present methods are as defined in the glossary, and additional studies and evaluations can be performed using additional drugs used as references, if necessary.

[0101] Transcriptome analysis can be performed according to any conventional technique available to the skilled artisan suitable for assessing whole transcriptome expression profiles, such as high-throughput RNA sequencing, qRT-PCR, microarrays, etc. According to the present invention, pathway and function analysis tools can be performed using commercially available or open source tools, one of the most well-known tools is QIAGEN Ingenuity® Pathway Analysis (IPA®), other known commercially available alternatives to IPA are MetaCore™ (Clarivate Analytics), GeneGo™ (by MetaCore™), QIAGEN OmicSoft®, Isevier® Pathway Studio®, Ingenuity® Variant Analysis (IVA®) (QIAGEN), and open source tools include DAVID (Database for Annotation, Visualization, and Integrated Discovery), GSEA (Gene Set Enrichment Analysis), STRING™ (Search Tool for the Retrieval of Interacting Genes / Proteins), KEGG Pathway Analysis, Metascape®, Cytoscape® with enrichment plugins (e.g., ClueGO®, ReactomeFI). Unless otherwise specified, this specification refers to the latest version of each of the above tools available at the time of filing of the present application.

[0102] In a preferred embodiment of the present invention, IPA can be used when pathway and functional analysis is mentioned in the steps of the methods disclosed herein, and in particular, when pathway enrichment and functional analysis of the transcriptome is mentioned (e.g., step 1.4), core analysis by IPA can be used.

[0103] According to the present invention, the value of 1.4 indicates the magnitude (e.g., fold change relative to the control) and direction (e.g., upregulation with a positive value and downregulation with a negative value). The value can be expressed as a Z-score. For example, IPA, MetaCore, GSEA, Enrichr provide results for Z-score-based gene regulation analysis. The fold change value obtained in step 1.4 can be expressed as a Z-score value to normalize gene expression changes across different datasets or experimental conditions.

[0104] When the DEG analysis of 1.4 is performed using tools that provide results in terms of visual representations such as color-coded heat maps, network diagrams, and enrichment plots, said representations are derived from quantitative data and can therefore be easily converted into numerical values ​​by those skilled in the art using commonly available statistical and computational methods.

[0105] For example, when referring to variation in magnitude and direction relative to a control (e.g., sample c), e.g., in 1.4, variation is assessed in terms of folding (magnitude) and up-regulation (direction), with up-regulation represented by a positive value, the number of said values ​​representing a fold change, and down-regulation represented by a negative value, the number of said values ​​representing a fold change.

[0106] Whole transcriptome expression profiles can be generated according to any method commonly used by those skilled in the art; a non-limiting example includes using the Human Clariom™ S Pico Assay HT (Applied Biosystems, ThermoFisher Scientific) on a GeneTitan MC Instrument (Applied Biosystems, ThermoFisher Scientific) according to the manufacturer's instructions. CEL intensity files can be generated using the Affymetrix GeneChip Command Console Software (AGCC, ThermoFisher Scientific). Data analysis can be performed using the Transcriptomic Analysis Console Software (TAC, ThermoFisher Scientific), which provides quality control analysis, performs normalization and summarization based on the Signal Space Transformation-Robust Multi-Chip Analysis (SST-RMA) analysis algorithm, and provides a list of differentially expressed genes (Limma Bioconductor package). Alternative commercially available and / or open-source tools known to those skilled in the art can be used in any part of generating the transcriptome expression profiles exemplified above. This step is performed to obtain a list of differentially expressed genes (DEGs) identified based on their expression fold change relative to relevant control experimental conditions (i.e., tumor burden without treatment, cell lines not treated with the product of interest, etc.).

[0107] According to one aspect of the present invention, analysis of transcriptional profiles can be performed using IPA, as described above. The use of IPA makes it possible to easily predict how and to what extent modulation of gene expression in a biological system affects biological activities related to a pathology of interest.

[0108] The resulting transcriptional modification profile is then subjected to functional pathway enrichment analysis. One commercially available tool that can be used, and was used in the examples provided herein, is Ingenuity Pathway Analysis (IPA version 94302991, Qiagen) [Kramer et al. (2014)]. When IPA is used, a list of differentially expressed genes and corresponding data measurements identified in different experimental conditions is uploaded to the application, and then available identifiers are mapped to their corresponding entities in Qiagen's Knowledge Base.

[0109] By initiating a "core analysis," significantly perturbed genes, called Network Eligible Molecules, are overlaid onto a global molecular network developed from information contained in the QIAGEN Knowledge Base. Networks of Network Eligible Molecules are then algorithmically generated based on their connectivity.

[0110] The "Core" analysis provides a comprehensive list of approximately the top 500 biological activities derived from the generated network. Associations between biological activities (also referred to herein as biological activity) and genes are always supported by corresponding annotations in scientific peer-reviewed publications, which are validated through automatic association to a Z-score [Kramer et al. (2014)], which represents the directionality and magnitude of modulation of the calculated biological activity. Essentially, this value represents a statistical metric that assesses the similarity between the observed pattern of differentially expressed genes (DEGs) and the expected pattern based on existing literature for a given annotation.

[0111] Biological activities relevant to the particular condition under investigation can be selected by examining appropriate AI, or selection can be based on the relevant state of the art. Selection of relevant biological activities is based on identified hallmarks of the condition of interest and applies a set Z-score threshold with statistical significance indicated by a p-value of 0.05 or less.

[0112] Hallmarks of a given pathology can be readily identified by one of skill in the art in specialized databases, through pathway and function databases, or through disease-specific databases.

[0113] The associated Z-score values ​​are then used to indicate the direction and magnitude of modulation of each biological function (Figures 4 and 9).

[0114] If the IPA "core analysis" (or its equivalent using a different open source or commercial tool) does not yield sufficient relevant information, an alternative available QIAGEN IPA approach called "overlay analysis" (or its equivalent using a different open source or commercial tool) can be used. This analysis focuses on biological activities identified by an "in silico model of the pathophysiological state." The selection of biological activities is built on identified hallmarks of the pathology of interest and can be done by examining appropriate AI or can be based on the latest technology.

[0115] "Overlay analysis" establishes relationships between patterns of differentially expressed genes and selected biological activities (always supported by corresponding annotations in scientific peer-reviewed publications demonstrating the directionality and magnitude of modulation of biological activity).

[0116] As an example, this can be done using the following steps:

[0117] Import the set of biological activities selected from the in silico model of the pathophysiological state into a new sheet called "my pathways".

[0118] Use the "Build Tool" and "Grow Tool" to identify differentially expressed genes (DEGs) that belong to the transcriptome profile under investigation and are associated with the regulation of the biological activity selected in the previous step.

[0119] The "Overlay" and "Molecule Activity Predictor" tools (MAP) are used to determine the predicted calculated impact of such experimentally observed gene expression modulations on biological activity. The "Prediction" function is activated within the MAP tool to calculate the predicted modulation that results from biological activity (note that QIAGEN IPA defines biological activity as "diseases and biological functions").

[0120] In step 2.1, apparent biological activities that, if present, contradict the observed therapeutic or biological effect of the selected reference drug are not selected, e.g., increased tumor formation by an antitumor drug.

[0121] According to the present description and claims, step 2.2, which converts the numerical values ​​obtained in 1.4 into absolute values, means that for each value obtained in 1.4, the modulus of said value is taken into account, i.e. the distance of the number (value) from zero on the number line, regardless of direction. Absolute values ​​are always non-negative.

[0122] Summing up each value means summing up each value of 2.2.

[0123] According to one aspect of the invention, the method comprises: Step 2.1 further comprises step 2.1.a of clustering said relevant biological activities into hallmarks of disease states or pathological conditions treated by the product of interest and providing a numerical coefficient of importance for each of said hallmarks.

[0124] Indeed, to refine the benefit score, one aspect of the present invention allows for the identification of hallmarks for a condition of interest and the assignment of a "numerical importance coefficient" based on their importance in disease progression and the impact of treatment. Clustering the relevant biological activities selected in 2.1 into the hallmarks and assigning a numerical importance coefficient to each hallmark allows for a more refined calculation of the relevance of each of the biological activities. A higher coefficient indicates that targeting the hallmark is important for effective treatment, while a lower coefficient suggests an indirect or emerging role.

[0125] Preferably, established hallmarks are used, i.e., characteristics officially recognized in the art. Hallmarks of disease states or pathological conditions can be easily retrieved from the state of the art, or they can be retrieved by computational methods or by querying appropriate databases, AI (e.g., ChatGPT, etc.) or dedicated computer programs.

[0126] As an example, hallmarks of pathologies can be searched for using computational methods by combining bioinformatics tools, machine learning, network analysis, and natural language processing (NLP).

[0127] Non-limiting examples of importance factors according to the present invention are provided below: Generally, hallmarks can be divided into two main categories, with the importance factor indicated in parentheses: Highly relevant hallmarks (2): A therapeutic or beneficial product must target said hallmark for meaningful therapeutic benefit.

[0128] Low association hallmark (1): Useful for complementary effects but not for primary efficacy markers.

[0129] The following provides non-limiting examples of hallmarks of different pathologies, along with importance factors suitable for the method of the present invention:

[0130] [Table 1] [Table 2] [Table 3] [Table 4] [Table 5]

[0131] The hallmarks and their importance coefficients for a given condition can be easily retrieved by one skilled in the art by appropriately consulting available AIs, including ChatGPT.

[0132] Once the hallmarks have been identified, dedicated programs or statistical or machine learning techniques can be used to group the biological activities into hallmarks using clustering algorithms.

[0133] Clustering related biological activities into hallmarks of a given pathology can also be readily performed by one skilled in the art by integrating biological data, applying computational methods, and leveraging existing biological knowledge (such as pathway or gene ontologies).

[0134] As discussed above, clustering biological activities into hallmarks of disease states or pathological conditions of interest can be achieved by applying computational methods and integrating biological data with existing biological knowledge (e.g., pathway or gene ontologies). Numerical importance factors are typically determined based on biological and / or clinical relevance or machine learning-based feature importance.

[0135] Furthermore, in step 2.2, the method may further include multiplying each of the numerical values ​​of the relevant biological activities obtained in 1.4 by the numerical coefficient of importance of the clustered hallmarks provided in 2.1.a before converting the numerical values ​​to absolute values. The numerical coefficient of importance of a hallmark allows for a quantitative assessment of its relevance to a given pathology, enabling a more accurate and objective calculation of the benefit score. This approach aids in prioritizing treatments or interventions by weighting the effects according to the biological significance of the hallmarks, ensuring that treatments targeting the most important disease mechanisms receive greater emphasis. It also facilitates personalized treatment strategies, standardized comparisons across studies, and efficient resource allocation, ultimately enhancing decision-making in research, clinical practice, and drug development.

[0136] According to the present invention, the side effects considered in 3.1 are those that are indicated in the reference drug leaflet as very common (affecting more than 1 in 10 patients) and common / frequent side effects (affecting more than 1 in 100 patients) or adverse drug reactions.

[0137] It is worth noting that side effects of drugs are reported according to global pharmacovigilance systems and are therefore officially registered, so that the skilled person, when carrying out the present invention, can simply rely on the leaflet accompanying the reference drug for the official list of the most common and common side effects.

[0138] In step 3.2, a pathway and functional analysis biological activity associated with the side effect can be determined by inputting the side effect into a pathway and functional analysis program and searching for a biological activity associated therewith.

[0139] In step 3.3, for samples a and b, an in silico model of risk using pathway and functional analysis by establishing the relationship between the gene expression patterns obtained in point 1.3 and the biological activity determined in point 3.2 can be determined using the tools of the selected pathway enrichment analysis software; as an example, if IPA is used, the Grow tool is suitable to build an in silico model of risk.

[0140] In 3.4, the associated fold-change value may be a Z-score value, as described above. The indications regarding Z-score values ​​provided above apply mutatis mutandis to step 3.4 of the method.

[0141] For 2.2, the data is also converted to absolute values ​​in 3.4.

[0142] In the event that the achieved modulation of biological activity does not indicate the activation of a side effect, the corresponding numerical value is set to 0. Conversely, if the achieved modulation of biological activity indicates the activation of a side effect, the numerical value is converted to its absolute value. In the proposed computer-implemented method for determining the benefit / risk score of a therapeutic or beneficial product, step 3.5 ensures that the risk score has a predetermined positive minimum value, even if the calculated risk score is initially zero or very low. The reason for this adjustment is that the benefit score must always be higher than the lowest biological activity value, which serves several important purposes.

[0143] To prevent a mathematically undefined benefit / risk ratio when the risk score is zero (i.e., no risk-relevant biological activity is detected), a predetermined positive minimum value of 3.5 is introduced, which makes the formula mathematically undefined because it divides by 0. This prevents meaningful interpretation of the results. By setting a minimum positive risk score, the system allows for differentiation between products with different safety profiles while avoiding splitting errors.

[0144] Furthermore, if the risk score is very low but not zero, the ratio may be artificially inflated, giving the unrealistic impression of a very high benefit / risk score. Normalizing the risk score prevents inflated values.

[0145] In practical applications, comparing multiple drugs or test compounds requires a standardized benefit / risk calculation. If the risk score approaches 0 for some products, their benefit / risk scores will be disproportionately higher than others, even if their actual clinical performance is similar. By enforcing a minimum positive risk value, the method ensures fair and standardized comparisons between different therapeutic candidates.

[0146] Furthermore, the methods of the present invention take into account that therapeutic or beneficial products are generally not completely risk-free, and therefore, even if in vitro or ex vivo transcriptional analysis does not detect strong activation of pathways associated with known side effects, this does not guarantee that the product will be completely risk-free in real-world use.

[0147] By setting a minimum positive risk value, the system recognizes the inherent uncertainty and ensures that even highly favored drugs have a finite, quantifiable risk component.

[0148] Finally, the method relies on an in silico risk model (step 3.3) that predicts risk based on pathway and functional analysis. While this model is powerful, it cannot capture all possible adverse effects due to potential toxicities not detected by transcriptional events, possible limitations in current pathway databases, and a potential lack of knowledge about all molecular interactions and differences between in vitro / in vivo conditions and actual patient responses.

[0149] Enforcing a minimum positive risk score helps correct potential underestimation of risk due to model limitations and prevents overconfidence in drug safety.

[0150] Therefore, according to the method of the present invention, the predetermined minimum positive value should depend on the scale of measurement and the type of numerical transformation applied to the risk-related biological activity. The value selected should balance preventing partitioning errors (avoiding infinite or artificially high benefit / risk ratios) while still allowing meaningful differentiation between drugs.

[0151] According to one aspect of the present invention, the predetermined positive minimum value may be a value between 0.001 and 1.0.

[0152] As an example, if the risk score is calculated as a sum of positive values ​​(e.g., pathway activation score, Z-score, or weighted functional annotation), a reasonable predetermined minimum positive value should be small but not negligible.

[0153] According to one embodiment, for example, if the sum of the pathway activity scores is generally in the range of 0 to 10, the predetermined minimum positive value may be 0.1, while if the sum is in the range of 1 to 100, the predetermined minimum positive value may be 1.0.

[0154] Generally, one skilled in the art can select said predetermined minimum positive value to be between 1 and 10%, for example 5% of the median observed risk score.

[0155] When an IPA-like pathway analysis is used, the predetermined minimum positive value is preferably selected from the range of 0.5 to 1.0.

[0156] The predetermined positive minimum value may also be selected relative to the benefit score, for example, between 1 and 10%, such as 5%, of the obtained benefit score.

[0157] Those skilled in the art who practice the method of the present invention know that the benefit score in a quantitative benefit / risk (B / R) assessment represents the total impact of therapeutically relevant biological activity induced by a drug. To ensure that a drug provides a meaningful therapeutic effect, its benefit score must always be higher than the lowest measurable biological activity value. If this condition is not met, proceeding with the B / R assessment is not justified. This also applies to the method of the present invention.

[0158] Furthermore, in one aspect of the present invention, an in silico model of risk may also be integrated to assign each adverse reaction associated with the reference drug provided in 3.1 a numerical factor of importance based on the severity of each adverse reaction, based on its clinical relevance, severity, frequency, and impact on the patient's quality of life.

[0159] For example, FDA EMA and WHO standards classify adverse reactions according to their severity, e.g. Classification by Severity (based on FDA, EMA, and WHO standards). Mild - No serious health effects; most recover without treatment. Non-limiting examples: nausea, dry mouth, mild headache, drowsiness Moderate: May interfere with daily activities and require medical intervention. Non-limiting examples: dizziness, muscle pain, skin rash, gastrointestinal discomfort Severe → Serious or life-threatening; medical attention required. Non-limiting examples: severe allergic reactions (anaphylaxis), liver failure, cardiac arrhythmias Lethal → Causing death or contributing significantly to a fatal condition. Non-limiting examples: severe drug-induced liver injury (DILI), toxic epidermal necrolysis (TEN)

[0160] Numerical coefficients of severity of adverse events associated with the reference drug can be provided, for example, based on standard clinical classification systems (e.g., the Common Terminology Criteria for Adverse Events from CTCAE-NCI).

[0161] For example, by assigning a severity score based on grade: Grade 1 (mild): Factor = 1 Grade 2 (moderate): Factor = 2 Grade 3 (severe): Factor = 3 Grade 4 (life-threatening): Factor = 4 Grade 5 (death): coefficient = 5, or can be calculated based on the frequency of each adverse event in clinical trials or real-world data, e.g., common (incidence > 10%): coefficient = 1.0 Infrequent (1-10% incidence): coefficient = 0.75 Rare (incidence less than 1%): coefficient = 0.5

[0162] By way of example, risk score adjustment can be easily performed with the aid of the following computer software: Assign a numerical importance factor to each side effect based on the severity, frequency, quality of life impact, and duration described above; Calculate risk contribution using gene expression fold changes, Calculate an overall risk score and apply a predetermined minimum; A benefit / risk score is generated by dividing the benefit score by the risk score.

[0163] Displays a summary table containing all relevant data.

[0164] That is, the method of the present invention further comprises: Assign a numerical importance factor to each side effect determined in step 3.1, and multiply the fold-change value obtained in step 1.3 associated with each side effect by the corresponding numerical importance factor.

[0165] As mentioned above, the numerical coefficient of importance is calculated, preferably in a computer-aided manner as disclosed above, based on one or more of the following parameters: Severity of side effects according to clinical classification systems, frequency of occurrence in clinical data, impact on patient-reported quality of life, and duration of side effects.

[0166] According to the present invention, step 2.1.a and / or step 3.2 may be performed using a machine learning model, such as a neural network or a support vector machine, trained on a known transcriptome dataset.

[0167] It should be noted that in mathematics and herein the value 0 is not considered a positive or negative number, and therefore a correction to a minimum positive value is necessarily applied when a value of 0 is obtained in the risk score values ​​provided by the methods disclosed herein.

[0168] Preferably, the method of the present invention is also performed on a reference drug, thereby providing a benefit / risk score for the product of interest and the reference drug, thereby allowing a direct comparison of the benefit / risk scores of the product of interest and the reference drug. Advantageously, the benefit / risk score of the reference drug can be used as a benchmark for assessing the therapeutic effectiveness of the product of interest.

[0169] Therefore, the present invention also refers to a method for selecting one or more new therapeutic or beneficial products for clinical development, wherein benefit / risk scores for one or more therapeutic or beneficial products of interest for the treatment of a given disease state or pathological condition and for a reference drug for the treatment of said disease state or pathological condition are provided by carrying out the steps defined above and in the claims, and wherein each of said at least one therapeutic or beneficial product of interest is selected for clinical development if the benefit / risk score value provided for at least one of said therapeutic or beneficial products of interest is equal to or greater than the benefit / risk score value provided for the reference drug.

[0170] The present invention further discloses a method for providing a benefit / risk score for a therapeutic product of interest for the treatment of a given condition based on data obtained from an in vivo animal model that can be performed in addition to the in vitro or ex vivo methods of the present invention to validate the results obtained with the in vitro or ex vivo methods disclosed herein based on data obtained in vivo.

[0171] Thef further method of providing a benefit / risk score for the therapeutic product of interest for the treatment of a given condition based on data previously obtained from in vivo animal models includes: I. To provide values ​​obtained from the following animal models that represent the pathology: a. Groups treated with said products and b. A group treated with a reference drug for the treatment of said condition; and c. Related control group The values ​​represent each of the following parameters: therapeutic efficacy observed in each of groups a and b normalized to the control group c; behavioral changes indicative of animal distress; and weight loss indicative of animal distress. II. determining a benefit score by summing the therapeutic efficacy values, thereby obtaining a benefit score value for the tested product; III. Determine a risk score; III.1 Identify the changes in animal distress and weight loss, as quantified in 1., as risk parameters. III.2. Add up the values ​​(for group a) from III.1 to obtain a final value representing the risk score of the inspected product; if the final value is less than a predetermined positive minimum value, it is automatically corrected to the positive minimum value. IV. Providing the value of the benefit / risk score of the product of interest as the ratio of the value of the benefit score obtained in point II to the value of the risk score obtained in point III.2.

[0172] The above methods can be practiced in addition to in vitro or ex vivo based methods of the present invention.

[0173] Preferably, the above method is based on existing data obtained from in vivo animal models, and the method uses only ethically approved, preferably existing animal study data or data obtained from publicly available scientific databases.

[0174] The method according to the invention can therefore be read as follows: A computer-implemented method for providing a benefit / risk score for a therapeutic or beneficial product of interest for treating a given condition based on existing data obtained from in vivo animal models.

[0175] I. Providing values ​​derived from previously obtained data on an in vivo animal model representative of said pathology, a. Group treated with said product; b. A group treated with a reference drug for the treatment of said condition; and c.Relevant control group Includes; where the numbers are: i. Treatment efficacy; ii. Behavioral changes indicative of treatment effects; and iii. Weight change observed in groups a and b normalized to control group c Representing, providing, II. Determining a benefit score by summing said therapeutic efficacy values, thereby obtaining a risk score for the tested product; III. Determining a risk score, comprising: III.1. Using the values ​​to identify risk parameters in relation to behavioral changes and weight changes; and III.2. Summing up the said values ​​related to the adverse effects of group a (identified in III.1) to obtain a final value representing the risk score of the tested product. If the final value is below a predetermined minimum positive threshold, it is adjusted to said minimum value. IV. Providing a benefit / risk score by calculating the ratio between the benefit score (obtained in II) and the risk score (obtained in III.2).

[0176] Where animal data are used, only data obtained from animals treated according to ethically approved protocols will be used.

[0177] As noted above, methods based on data obtained from in vivo models can be advantageously combined with methods based on in vitro or ex vivo data to further validate the predictive results obtained therefrom.

[0178] Data obtained from the in vivo models in the following examples, which validate results obtained in vitro or ex vivo using the methods of the present invention, demonstrate that the in vitro or ex vivo methods of the present invention do indeed provide reliable benefit / risk scores, thereby resulting in unnecessary in vivo experimentation.

[0179] Additionally, the methods disclosed herein for selecting one or more new therapeutic or beneficial products for clinical development may further comprise calculating a benefit / risk score for each test product and reference drug using a method based on data obtained from an in vivo model as disclosed above, wherein each of said at least one therapeutic or beneficial product of interest is selected for clinical development if the benefit / risk score value provided for at least one of said therapeutic or beneficial products of interest in both methods (based on data obtained from the in vitro or ex vivo and in vivo model) is equal to or greater than both of the relative benefit / risk score values ​​provided for the reference drug (based on data obtained from the in vitro or ex vivo and in vivo model).

[0180] Equal to or higher, as used herein, means that the value is numerically higher, i.e., the benefit / risk score value of the tested product is numerically equal to or higher than the benefit / risk score of the reference drug, preferably it is even statistically significantly superior.

[0181] All further steps disclosed for the in vitro or ex vivo-based methods above apply mutatis mutandis to the methods based on data obtained from the in vivo models disclosed herein. The therapeutic efficacy parameter depends on the pathological condition of interest; for example, in the case of an anti-cancer drug, the parameter is primarily represented by tumor burden reduction, optionally represented by the presence or absence of metastasis; in the case of an antidepressant, the parameter may include behavioral changes, such as an improvement in the severity of depression. Thus, the parameter is one that generally correlates with the assessment of therapeutic efficacy in treating a given pathological condition.

[0182] Parameters indicative of behavioral changes indicative of animal distress are those coded by standard tests commonly used in the art, such as the Irwin test, used in formal protocols for assessing animal distress in preclinical studies or in housing.

[0183] Even in this case, a benefit / risk score can be provided for the reference drug as well.

[0184] As mentioned above, preferably, where applicable, both methods are performed to verify the results obtained with either one of them.

[0185] In a preferred embodiment of the present invention, in the methods described herein, said therapeutic or beneficial product of interest is a product comprising one or more natural matrices, such as natural matrices obtained from eukaryotic sources, prokaryotic sources, such as plant sources, marine sources, bacterial sources, fungal sources, yeast sources, animal sources, natural sources; for example, a product comprising one or more of the following: chopped or crushed plant parts, plant extracts, fractions of said extracts, microorganisms, honey, propolis, silk, wax, plant resins, plant gums, plant exudates, plant oils, plant essential oils, animal tissue lysates, or plant or animal body fluids.

[0186] In any part of this specification and claims, the word comprising may be replaced with the word consisting of.

[0187] In any part of the specification or claims, where calculations are performed, this may be performed in a computer-implemented mode.

[0188] Wherever in this specification an internet address or URL is provided, it relates to information retrieved from said internet address available on the filing date of this application, i.e., from the most recent version of the content at said internet address on the filing date of this application.

[0189] Wherever in this specification or claims reference is made to commercially available products, the symbols TM or (registered trademark) are considered implicit and may be added to each of said products at any time.

[0190] Anywhere in this specification or claims, where a reference is made to a "computer-implemented method for providing a benefit / risk score," the sentence may be replaced with "a computer-implemented method for producing a therapeutic or beneficial product, the method comprising determining a benefit / risk score thereof."

[0191] Examples are reported below with the purpose of better illustrating the embodiments disclosed herein; such examples should in no way be construed as limiting the scope of the foregoing specification and subsequent claims; furthermore, the following examples report all work carried out on the products of the present invention and support all claimed subject matter.

[0192] example 1. Composition of the test product Product A, also referred to herein as Osteoredux, for the treatment of bone fragility Coral skeleton powder 32%w / w Bird eggshell powder 30.2%w / w Dried coral skeleton + lemon juice powder (contains calcium citrate) 13% w / w Agaricus bisporus powder 4.65% w / w Horsetail (Equisetum arvene) flower apex dry extract 2% w / w Acerola (Malpighia punicifolia) loaded with dry extract of inulin 2.0% w / w Island moss (Cetraria islandica) powder 2% w / w Sisal (Agave sisalana) leaf powder 12% w / w Acacia senegal powder 2.15% w / w

[0193] Product B herein is also EpigenAU / 11 for the treatment of cancer. 36.05% by weight of lyophilized ingredient 1 63.06% by weight of freeze-dried component 2 0.89% lyophilized component 3 by weight. Ingredient 1 Laurus nobilis leaves 25% w / w Ashwagandha (Whitania somnifera) root 25% w / w Filipendula vulgaris leaves and flowers 25% w / w Broccoli (Brassica oleracea L.botrytis cymosa) seeds 25% w / w Co-extraction in water Ingredient 2 Artichoke (Cynara scolymus L.) leaves 14.30% w / w Turmeric (Curcuma longa L.) root 42.85% w / w Feverfew (Tanacetum parthenium L.) flowers 42.85% w / w Co-extraction in water Component 3 Freeze-dried extract of sisal (Agave sisalana) leaves.

[0194] 2. Overview of the Protocol Used [Table 6-1] [Table 6-2] [Table 6-3] [Table 6-4]

[0195] 3. Techniques and settings for calculating benefit / risk scores for products A and B. 3.1 In vitro or ex vivo sample treatment Bone fragility: Product A - Osteoredux Human adipose tissue-derived mesenchymal stem cell lines (hADMSCs), capable of differentiating into osteoblasts and mineralizing extracellular matrix (ECM), were used in this study. These cells were obtained during general surgery from three different patients (PA42, PA59, and PA69) (Romagnoli et al., "In Vitro Behavior of Human Adipose Tissue-Derived Stem Cells on Poly(ε-caprolactone) Film for Bone Tissue Engineering Applications," BioMed Research International, Vol. 2015, Article ID 323571, p. 12, 2015. https: / / doi.org / 10.1155 / 2015 / 323571). These cell lines have been characterized for key stemness markers of mesenchymal stem cells (CD44, CD105 and STRO1) and by studying their multipotentiality towards an osteogenic phenotype at the Department of Surgery and Translational Medicine of the University of Florence.

[0196] hADMSCs were cultured in growth medium (GM) and grown to 70-80% confluence. Cells were then seeded into 24-well plates at a concentration of 1 x 10 cells / well. After 1 week, GM was replaced with osteogenic medium (OM) containing 1 µg / mL of the fluorophore calcein and incubated with or without Osteoredux or the reference agent DiBase for 28 days. Medium with or without Product C was refreshed twice weekly.

[0197] Cancer area: Product B-EpigenAU / 11 To conduct the experiment, ex vivo tumor masses were generated from the FaDu head and neck squamous cell carcinoma cell line and implanted into immunocompromised mice. Once tumors reached the appropriate size, they were excised, divided into 40 mg portions, and treated in triplicate with EpigenAU / 11 or the reference drug Cisplatin for 6 hours. The masses were then lysed, and RNA was extracted for transcriptional analysis.

[0198] 3.2 Transcriptome raw data analysis Whole-transcriptome expression profiles were assessed using the Human Clariom™ S Pico Assay HT (Applied Biosystems, ThermoFisher Scientific) on a GeneTitan MC Instrument (Applied Biosystems, ThermoFisher Scientific) according to the manufacturer's instructions. CEL intensity files were generated using the Affymetrix GeneChip Command Console Software (AGCC, ThermoFisher Scientific). Data analysis was performed using the Transcriptomic Analysis Console Software (TAC, ThermoFisher Scientific), which provides quality control analysis, normalization, and summarization based on the Signal Space Transformation-Robust Multi-Chip Analysis (SST-RMA) analysis algorithm, and provides a list of differentially expressed genes (Limma Bioconductor package). This step allows for obtaining a list of differentially expressed genes (DEGs) identified based on their expression fold changes relative to relevant control experimental conditions (i.e., tumor burden without treatment, cell lines not treated with the product of interest, etc.). A detailed protocol is provided in the table above.

[0199] 3.3 Ingenuity Pathway Analysis IPA (QIAGEN IPA) Analysis of Transcription Profiles The use of IPA allows one to predict how and to what extent modulation of gene expression in a biological system affects biological activity associated with a pathology of interest.

[0200] The resulting transcriptional modification profiles were subjected to functional pathway enrichment analysis using Ingenuity Pathway Analysis (IPA version 94302991, Qiagen) [Kramer et al. (2014)]. For each benefit / risk assessment (e.g., Product A, Product B), a list of differentially expressed genes and corresponding data measurements identified across different experimental conditions were uploaded to the application.

[0201] Available identifiers were mapped to their corresponding entities in the QIAGEN Knowledge Base.

[0202] By initiating the IPA "core analysis," significantly perturbed genes, called Network Eligible Molecules, are overlaid onto a global molecular network developed from information contained in the QIAGEN Knowledge Base. Networks of Network Eligible Molecules are then algorithmically generated based on their connectivity.

[0203] The "core" analysis provides a comprehensive list of approximately the top 500 biological activities derived from the generated network. The IPA association of biological activity and genes is always supported by corresponding annotations in scientific peer-reviewed publications that demonstrate the directionality and magnitude of the calculated biological activity modulation through automated association of z-scores [Kramer et al. (2014)] (when referring to biological activity in this invention, the IPA designation is "biological activity"). Essentially, the resulting value represents a statistical metric that assesses the similarity between the observed pattern of differentially expressed genes (DEGs) and the expected pattern based on existing literature for a given annotation.

[0204] The biological activities relevant to the specific pathology under investigation were selected based on the relevant state-of-the-art, but the same can be done by examining appropriate AI. The selection of biological activities is based on identified hallmarks of the pathology of interest and applies a set of Z-score thresholds with statistical significance indicated by a p-value of 0.05 or less.

[0205] The associated Z-score values ​​were then used to indicate the direction and magnitude of modulation of each biological function (Figures 4 and 9).

[0206] When the IPA core analysis did not yield sufficient relevant information, an alternative approach called "overlay analysis" was used. This analysis focused on biological activities identified by "in silico models of the pathophysiological state." The selection of biological activities was based on identified hallmarks of the pathology of interest. This could be done by examining appropriate AI or could be based on the latest technology.

[0207] "Overlay analysis" is constructed by establishing relationships between patterns of differentially expressed genes and selected biological activities (always supported by corresponding annotations in scientific peer-reviewed publications that demonstrate the directionality and magnitude of modulation of biological activity).

[0208] This was done using the following procedure:

[0209] A set of biological activities (called functions in IPA) selected from the in silico model of the pathophysiological state was imported into a new sheet called “my pathways”.

[0210] The "build tool" and "grow tool" were used to identify differentially expressed genes (DEGs) belonging to the transcriptome profile under investigation and associated with the regulation of the biological activity selected in the previous step.

[0211] To determine the predicted calculated impact of such experimentally observed gene expression modulation on biological activity, the "Overlay" and "Molecule Activity Predictor" tools (MAP) were used. The "Prediction" function was activated within the MAP tool to calculate the resulting predicted modulation of biological activity.

[0212] 3.4 Image analysis of the resulting color intensity for biological activity Because "overlay analysis" does not directly calculate Z scores for each biological activity (a function in IPA), the intensity of the modulation signal was quantified. This was achieved by converting the data using the dedicated app "IPAmmap_Parser" (version 2.1-1).

[0213] The app is a web port of Pipeline Pilot, designed to assign scores, called z-scores, to genes and biological functions based on coloring within biological pathways generated by QIAGEN's Ingenuity Pathway Analysis software. A key step in the algorithm is the conversion from the RGB color model to the LAB [https: / / www.xrite.com / it-it / blog / lab-color-space] model, a colorimetric encoding that allows for recording of color intensity as well as RGB composition. This conversion is performed within Pipeline Pilot's "components," which utilize procedures written in R software, relying on specific functionality from the colorspace package [https: / / cran.r-project.org / web / packages / colorspace / index.html].

[0214] 4. Calculating the risk score 4.1 Definition of Side Effects Officially recorded side effects that were very common or common (more than 1 in 10 patients or more than 1 in 100 patients) were identified by examining specific sources: -https: / / www.torrinomedica.it -https: / / www.drugs.com / -Reference product specific leaflet Area of ​​interest specific sources Osteoarthritis: -https: / / www.torrinomedica.it / schede-farmaci / kenacort / -https: / / www.drugs.com / sfx / triamcinolone-side-effects.html Bone fragility: -https: / / www.torrinomedica.it / schede-farmaci / dibase / -https: / / www.drugs.com / sfx / cholecalciferol-side-effects.html Cancer area: -https: / / www.torrinomedica.it / schede-farmaci / cisplatino-3 / -https: / / www.drugs.com / sfx / cisplatin-side-effects.html

[0215] 4.2 Identifying IPA biological functions by browsing terms used to annotate side effects The information found in the aforementioned resources is used to identify the BF (i.e., biological activity) associated with the identified adverse reaction and to investigate the IPA by the following steps: Each side effect term was entered into the "Disease and Function" query box, and then the search was initiated.

[0216] The resulting analysis of the recursive tables was filtered to identify diseases / functions with a high level of evidence. The relevant sources were the Ingenuity Knowledge Base, which includes journal articles, OMIM, JAX, and curation from ClinicalTrials.gov.

[0217] The in silico model was limited to genes and mRNAs, and thus the tool could associate each biological activity with a defined number of genes whose modulation can affect the modulation of the biological activity itself (which in this particular case can represent side effects).

[0218] 4.3 Overlay Analysis The overlay analysis focuses on biological activities identified by the "in silico model of side effects." The selection of biological activities is constructed based on the identified side effects of the reference treatment.

[0219] An "overlay analysis" is constructed by establishing a relationship between patterns of differentially expressed genes and selected biological activities using the following procedure (always supported by corresponding annotations in scientific peer-reviewed publications demonstrating the directionality and magnitude of modulation of biological activity):

[0220] The set of biological activities selected from the in silico model of SIDE EFFECT PROFILE was imported into a new sheet called "my pathway".

[0221] The "build tool" and "grow tool" were used to identify differentially expressed genes (DEGs) belonging to the transcriptome profile under investigation and associated with the regulation of the biological activity selected in the previous step.

[0222] To determine the predicted calculated impact of such experimentally observed gene expression modulations on biological activity, the "Overlay" and "Molecule Activity Predictor" tools (MAP) are used. The "Prediction" function is activated within the MAP tool to calculate the predicted resulting modulation of biological activity.

[0223] 4.4 Image analysis of the color intensity obtained for biological activity As mentioned above, since the "overlay analysis" does not directly calculate the Z-score of each biological function, the intensity and direction of modulation were converted from the color heatmap to numerical values ​​using the dedicated app "IPAtmap_Parser" to obtain Z-scores representing the intensity and directionality of each modulation.

[0224] 4.5 Calculating the Global Risk Score A risk score was then calculated by summing each positive Z-score value obtained from the overlay analysis (to indicate that a given side effect was treatment-induced).

[0225] 5. Calculating the Benefit Score 5.1 Core analysis, biological activity selection and export Core analysis was used to select appropriate biological activities (biological functions to use the IPA language) considering trends consistent with the therapeutic indications of the product under investigation. Biological activities relevant to specific treatments can be identified by pathway enrichment analysis, using AI suitable for literature searches, or using disease state-specific databases.

[0226] The following biological activities were selected for each pathological area examined and used in conjunction with IPA: Cancer area Known alterations characteristic of healthy physiological states associated with cancer were examined with particular attention to the following areas of involvement: cell damage Energy metabolism and insulin sensitivity Epithelial-mesenchymal transition growth factors inflammation Mitosis / proliferation regulation pace The above was used to select and classify the output of the IPA core analysis. Bone fragility The known changes characteristic of a healthy physiological state of "bone fragility" were examined with particular attention to the following areas of involvement: -Bone remodeling -osteoporosis -Osteoblast differentiation -Mineralization -Reduces inflammation -Reduction of bone and fatty tissue

[0227] The above was used to investigate IPA with the "IPA Bioprofiler" tool using the following keywords: osteoporosis, postmenopausal osteoporosis, bone mineralization, osteoblast and osteoclast differentiation, bone mineral density.

[0228] BF was grouped by activity hallmarks (optionally, each hallmark is given a weighting factor based on its importance in the pathogenesis of the subject of interest). Therefore, the BF Z-score values ​​were converted into absolute values, and the list thus obtained was exported to a table reporting the specifications of biological functions and their relative adjustment values. The sum of each value for each treatment was considered as the benefit score.

[0229] 6. Calculating the Benefit / Risk Score Once the risk and benefit scores were calculated, they were applied to the following formula: (Benefit score) / (Risk score)

[0230] Therefore, the ratio value must be interpreted such that the higher the value obtained, the greater the safety of the administered treatment.

[0231] 7. New Benefit / Risk Assessment Product A (Computer-Implemented / Supported) 7.1 Data Assembly 7.1.1 Ex vivo sample treatment and RNA extraction (steps 1.1 and 1.2) To perform the experiment, ex vivo tumor masses were generated from the FaDu head and neck squamous cell carcinoma cell line and implanted into immunocompromised mice. Once tumors reached the appropriate size, they were excised, divided into 40 mg portions, and treated in triplicate with EpigenAU / 11 (sample a) or cisplatin (sample b), or left untreated (sample c) for 6 hours. The masses were then lysed, and RNA was extracted according to standard protocols for transcriptional analysis.

[0232] 7.1.2 Transcriptome raw data analysis (step 1.3) Whole-transcriptome expression profiles were assessed using the Human Clariom™ S Pico Assay HT (Applied Biosystems, ThermoFisher Scientific) on a GeneTitan MC Instrument (Applied Biosystems, ThermoFisher Scientific) according to the manufacturer's instructions. CEL intensity files were generated using the Affymetrix GeneChip Command Console Software (AGCC, ThermoFisher Scientific). Data analysis was performed using the Transcriptomic Analysis Console Software (TAC, ThermoFisher Scientific), which provides quality control analysis, normalization, and summarization based on the Signal Space Transformation-Robust Multi-Chip Analysis (SST-RMA) analysis algorithm, providing a list of differentially expressed genes (Limma Bioconductor package). This step allows for the generation of a list of differentially expressed genes (DEGs) identified based on their expression fold-change relative to the relevant control experimental condition (i.e., tumor burden without treatment).

[0233] 7.1.3 IPA analysis of transcriptional profiles (implemented computer) (step 1.4) The use of IPA allows us to predict how and to what extent modulation of gene expression in a biological system affects biological activity associated with a pathology of interest.

[0234] The transcriptional modification profiles thus obtained were subjected to functional pathway enrichment analysis. One commercially available tool that can be used is Ingenuity Pathway Analysis (IPA version 94302991, Qiagen) [Kramer A et al., A Causal Approach in Ingenuity Pathway Analysis 2014]. The list of differentially expressed genes and the corresponding data measurements (fold changes relative to the "untreated tumor mass") identified in different experimental conditions were uploaded to the application.

[0235] The differentially expressed genes and corresponding fold changes are subjected to a filtering process to select only genes that are significantly perturbed, as indicated by their fold changes compared to the "untreated tumor mass."

[0236] The fold change threshold was set to encompass values ​​less than or equal to -2 and greater than or equal to +2, with statistical significance indicated by a p-value of less than or equal to 0.05.

[0237] Available identifiers were mapped to their corresponding entities in the QIAGEN Knowledge Base.

[0238] By starting with a "core analysis," significantly perturbed DEGs, called Network Eligible Molecules, were overlaid onto a global molecular network developed from information contained in the QIAGEN Knowledge Base. Networks of Network Eligible Molecules were then algorithmically generated based on their connectivity.

[0239] The "Core" analysis provides a comprehensive list of approximately the top 500 biological activities derived from the generated network. Biological activity-gene associations are always supported by corresponding annotations in scientific peer-reviewed publications that demonstrate the directionality and magnitude of the calculated biological activity modulation through automated association z-scores [Kramer et al. (2014)]. Essentially, this value represents a statistical metric that assesses the similarity between the observed pattern of differentially expressed genes (DEGs) and the expected pattern based on existing literature for a given annotation.

[0240] Step 2.1: Selection of biological activities relevant to the specific pathology under investigation. The selection of biological activities is based on the identified hallmarks of the pathology of interest and applies a Z-score threshold set to encompass values ​​below -2 and above +2, with statistical significance indicated by a p-value of 0.05 or less.

[0241] Step 2.2: The associated Z-score values ​​(Figure 4 (values ​​in column cluster a)) were used to indicate the directionality and magnitude of modulation of each biological function, and the values ​​were converted to absolute values ​​and summed to obtain the benefit score (step 2.3).

[0242] 7.2 Risk Score Calculation 7.2.1 Defining cisplatin side effects (step 3.1) Officially recorded side effects of cisplatin were identified by examining specific sources: -https: / / www.torrinomedica.it / schede-farmaci / cisplatino-3 / (document made available by AIFA on 07 / 13 / 21) -https: / / www.drugs.com / sfx / cisplatin-side-effects.html as last updated on March 17,2024

[0243] 7.2.2 Identifying IPA biological functions from terms used to annotate side effects (step 3.2) Using information found in the aforementioned resources, identify the biological activity associated with the identified adverse reaction (the IPA source used is the Ingenuity Knowledge Base, which includes curation from journal articles, OMIM, JAX, and ClinicalTrials.gov) and investigate the IPA using the following procedure.

[0244] Side effect terms are written one by one into the "Disease and Function" query box, and then the search is initiated.

[0245] The resulting resumption table allowed filtering diseases / functions arising from a large body of evidence (it is possible to understand the correspondence between IPA biological activities and annotated side effects, as shown in Figure 3). In silico modeling was performed to reduce each biological function to a defined number of genes whose modulation can affect the modulation of the biological function itself (which in this particular case can represent side effects).

[0246] 7.2.3 Overlay analysis and construction of in silico models (step 3.3) The overlay analysis focuses on biological activities identified by the "in silico side effect model." The selection of biological activities is based on the identified side effects of cisplatin.

[0247] An "overlay analysis" is constructed by establishing a relationship between patterns of differentially expressed genes and selected biological activities using the following procedure (always supported by corresponding annotations in scientific peer-reviewed publications demonstrating the directionality and magnitude of modulation of biological activity):

[0248] Import the set of selected biological activities from the in silico model of the SIDE EFFECT PROFILE of cisplatin into a new sheet called "my pathway".

[0249] Utilizing the "Build Tool" and "Grow Tool", differentially expressed genes (DEGs) belonging to the transcriptome profile under investigation (cisplatin and EpigenAU / 11) are identified and associated with the regulation of the biological activity selected in the previous step.

[0250] 7.2.4 Image analysis of the color intensity of the obtained biological activity (3.4) Modulation of identified DEGs is depicted using green (indicating down-modulators) and red (indicating up-modulation) colors.

[0251] To determine the predicted calculated impact of such experimentally observed gene expression modulations on biological activity, the "Overlay" and "Molecule Activity Predictor" tools (MAP) are used. The "Prediction" function is activated within the MAP tool to calculate the predicted resulting modulation of biological activity. Thus, a color coding is established: -Orange: Increased activity - Blue: decreased activity -White: unattainable / unpredictable

[0252] "Overlay analysis" does not directly calculate the Z-score of each biological function. Therefore, it is necessary to convert the color intensity of the modulation signal into a numerical value. This is achieved by using the dedicated app: "IPAmmap_Parser" (version 2.1-1).

[0253] The app is a web port of Pipeline Pilot, designed to assign scores, called z-scores, to genes and biological functions based on coloring within biological pathways generated by QIAGEN's Ingenuity Pathway Analysis software. A key step in the algorithm is the conversion from the RGB color model to the LAB [https: / / www.xrite.com / it-it / blog / lab-color-space] model, a colorimetric encoding that allows for recording of color intensity as well as RGB composition. This conversion is performed within Pipeline Pilot's "components," which utilize procedures written in R software, relying on specific functionality from the colorspace package [https: / / cran.r-project.org / web / packages / colorspace / index.html].

[0254] Figures 33A-B show the final list of biological activities and corresponding values ​​obtained for each treatment under investigation.

[0255] 7.2.5 Calculating the Global Risk Score (Step 3.5) Thus, the risk score is calculated by the sum of each positive value obtained from the overlay analysis (to indicate that a given side effect is induced by the treatment).

[0256] Calculation of the benefit score for EpigenAU / 11 and cisplatin treatment after 6 hours of treatment (Figure 33C shows the risk score obtained by summing each annotated biological activity value for each of the treatments under investigation).

[0257] Calculation of the benefit / risk score was performed by dividing the resulting benefit score by the resulting risk score (see Figures 6A and B).

[0258] The benefit / risk score for Product A was calculated using an additional step: the introduction of a numerical factor for the importance of the hallmark.

[0259] 7.3 Benefit Score Calculation 7.3.1 Core Analysis Physiological Function Selection and Export Using the core analysis obtained in 7.1.3, biological activities were selected based on their tendency to match the therapeutic indication of the product under investigation (i.e., antitumor activity). The biological activities were grouped into activity hallmarks, and each hallmark was assigned a weighting factor based on its importance in the pathogenesis of the target disease. Therefore, the biological activity Z-score values ​​were converted to absolute values, and the resulting list was exported to a table reporting the biological function specifications and their relative regulatory values. In this case, the biological activity values ​​were multiplied by the corresponding numerical importance factor according to step 2.1a, thereby obtaining the final list values ​​used for the final calculation of the benefit score (Figure 4 (values ​​in column b of cluster b)).

[0260] The sum of each value for each treatment provided the benefit score (Figure 6).

[0261] Risk scores were assessed as previously indicated.

[0262] 7.4 Calculating the Benefit / Risk Score Once the risk and benefit scores are calculated, they are applied to the following formula: JPEG0007758403000011.jpg12165 (Benefit score / Risk score) (See Figures 6C and D).

[0263] The value of the ratio must therefore be interpreted as follows: the higher the value obtained, the safer the treatment applied, since the benefits outweigh the risks.

[0264] Figure 6 summarizes the results of the benefit / risk scores obtained with EpigenAU / 11 and cisplatin treatment after 6 hours (A). Calculation of the fold change in the score obtained with EpigenAU / 11 compared to the score obtained with cisplatin: EpigenAU / 11 score is double the score for cisplatin (B).

[0265] When calculating the benefit / risk score taking into account the optional steps described above (numerical coefficients of importance), the results obtained are shown in Figure 6C, showing the calculation of the fold change in the score obtained with EpigenAU / 1 compared to the score obtained with cisplatin: the EpigenAU / 11 score is still double that of cisplatin (Figure 6D).

[0266] Therefore, benefit / risk scores calculated with and without the optional step are consistent with each other.

[0267] 8. Validation of the proposed transcriptome-based "Method A" in comparison with "Method B," intended as a general assessment of the benefit / risk profile associated with in vivo administration of EpigenAU / 11 and cisplatin. 8.1. In vivo assay (step 1) EpigenAU / 11 (DoE2) and cisplatin were administered in vivo in a mouse model. Risks (intended as side effects) were estimated by observing behavioral parameters (affecting behavior itself, the locomotor system, muscle strength and neuroreflexes) and potential weight loss in the animals.

[0268] 8.2 Comparison of the Benefit / Risk Ratios of EpigenAU / 11 and Cisplatin Based on Data Obtained from In Vivo Experiments 8.2.1 Risk Score Calculation Data obtained from animal studies that are considered in relation to risk are behavioral parameters and weight loss at the end of the experiment.

[0269] Behavioral parameters monitored are listed below: 1. Loss of spontaneous activity 2. Loss of cleaning 3. Loss of curiosity 4. Loss of responsiveness 5. Loss of corrective reflex 6. Loss of physical strength 7. Eyelid opening disorder 8. Lid reflex 9. Tremors 10. Pallor 11. Stereotypes 12. Passivity

[0270] The values ​​of parameters 1 to 8 are normalized as "untreated animal value" (always equal to 0) + "treated animal value".

[0271] The values ​​of parameters 9-12 are normalized as "untreated animal value" + "treated animal value". Potential weight loss in treated animals was also assessed. The following formula was used to process the data: 1-[[Weight of treated animals (observation day: 27) / Weight of untreated animals (observation day: 19)]]

[0272] The table below summarizes the risk score calculations for EpigenAU / 11 and cisplatin treatment using the sum of the data obtained from the in vivo experiments.

[0273] The table below shows experimental values ​​representing in vivo side effects that have not yet been reprocessed for the purposes of risk score calculation. [Table 7]

[0274] 8.3 Benefit Score Calculation (See Method B, Step 2) The data obtained from the animal studies that are considered in relation to benefit is the size of the tumor mass reached at the end of the in vivo testing of each treatment.

[0275] The data was calculated as "reduction in tumor mass size" according to the following formula: 1 - [[Tumor mass size collected from treated animals (observation day: 27) / Tumor mass size collected from untreated animals (observation day: 19)]]

[0276] The calculation of the benefit scores for EpigenAU / 11 and cisplatin treatment using data obtained from in vivo experiments is reported below.

[0277] Instead, benefit was assessed by taking into account the reduction in size of the tumor mass compared to untreated tumors (see table in Example 8.3).

[0278] For this purpose, tumors 100 mm 3 Immunocompromised mice bearing xenografted FaDu head and neck squamous cell carcinoma cells were enrolled in the study once tumor volume reached 100 μg / mL. Mice received daily intratumoral injections of EpigenAU / 11 and intraperitoneal cisplatin every other day.

[0279] The table below shows experimental values ​​that represent in vivo benefit scores that have not yet been reprocessed for the purposes of benefit score calculation. [Table 8]

[0280] The obtained values ​​indicate that the standard-assessed benefit / risk profile associated with EpigenAU / 11 is more favorable than that associated with cisplatin. Thus, data generated using standard in vivo observations confirm the predictive potential and demonstrate the validation of the novel method reported herein based on ex vivo transcriptomics.

[0281] Alternatively, previously obtained and independently obtained data from in vivo experiments can be used in the methods of the present invention.

[0282] 8.4. Benefit / Risk Score Calculation Once the risk and benefit scores are calculated, they are applied to the following formula: JPEG0007758403000014.jpg12165

[0283] Therefore, the ratio values ​​should be interpreted as follows: The higher the value obtained, the safer the treatment applied, since the benefits outweigh the risks. .

[0284] Figure 7 shows the benefit / risk scores obtained with EpigenAU / 11 and cisplatin treatment in vivo (A). Calculation of the fold change in the score obtained with EpigenAU / 11 compared to the score obtained with cisplatin: EpigenAU / 11 score is 3 times higher than the score for cisplatin (B).

[0285] Two methods (with or without optional steps) disclosed in Examples 7 and 8 have proven useful for objectively comparing the benefit / risk profiles of two anticancer treatments. The first method, based on ex vivo transcriptome data, has proven capable of accurately predicting the favorability of the EpigenAU / 11 profile relative to that of cisplatin and has undergone validation based on comparison with results generated using standard qualitative approaches based on in vivo data. The second method allows for the summarization of such in vivo data with a single parameter, making the comparison between the benefit / risk profiles of treatment options easier and more accessible. Both methods yield similar results demonstrating the superiority of the EpigenAU / 11 benefit / risk score relative to cisplatin (Figure 36), thus demonstrating the consistency of the ex vivo method with those generated using standard in vivo approaches, thereby returning results that represent a favorable profile of EpigenAU / 11 relative to that of cisplatin and are consistent with those generated using standard in vivo approaches.

[0286] 9. New Benefit / Risk Assessment Product B (Computer-Implemented / Supported) 9.1 Data Assembly 9.1.1 In vitro sample treatment and RNA extraction (steps 1.1 and 1.2) Human adipose tissue-derived mesenchymal stem cell lines (hADMSCs), capable of differentiating into osteoblasts and mineralizing extracellular matrix (ECM), were used in this study. These cells were obtained during total body surgery from three different patients (PA42, PA59, and PA69) (Romagnoli et al., "In Vitro Behavior of Human Adipose Tissue-Derived Stem Cells on Poly(ε-caprolactone) Film for Bone Tissue Engineering Applications," BioMed Research International, Vol. 2015, Article ID 323571, p. 12, 2015. https: / / doi.org / 10.1155 / 2015 / 323571). These cell lines have been characterized for key stemness markers of mesenchymal stem cells (CD44, CD105 and STRO1) and by studying their multipotentiality towards an osteogenic phenotype at the Department of Surgery and Translational Medicine of the University of Florence.

[0287] hADMSCs were cultured in growth medium (GM) and grown to 70-80% confluence. Cells were then seeded into 24-well plates at a concentration of 1 x 10 cells / well. After 1 week, GM was replaced with osteogenic medium (OM) containing 1 µg / mL of the fluorophore calcein and incubated with or without Osteoredux or DiBase for 28 days. Medium with or without Product C was refreshed twice weekly.

[0288] 9.1.2 Transcriptome raw data analysis (step 1.3) Whole-transcriptome expression profiles were assessed using the Human Clariom™ S Pico Assay HT (Applied Biosystems, ThermoFisher Scientific) on a GeneTitan MC Instrument (Applied Biosystems, ThermoFisher Scientific) according to the manufacturer's instructions. CEL intensity files were generated using the Affymetrix GeneChip Command Console Software (AGCC, ThermoFisher Scientific). Data analysis was performed using the Transcriptomic Analysis Console Software (TAC, ThermoFisher Scientific), which provides quality control analysis, normalization, and summarization based on the Signal Space Transformation-Robust Multi-Chip Analysis (SST-RMA) analysis algorithm, providing a list of differentially expressed genes (DGEs) (Limma Bioconductor package). This step allowed us to obtain a list of differentially expressed genes (DEGs).

[0289] 9.1.3 IPA analysis of transcriptional profiles (implemented in silico) (step 1.4) The use of IPA allows us to predict how and to what extent modulation of gene expression in a biological system affects biological activity associated with a pathology of interest.

[0290] The transcriptional modification profiles obtained in this way were subjected to functional pathway enrichment analysis. One commercially available tool that can be used is Ingenuity Pathway Analysis (IPA version 94302991, Qiagen) [Kramer A et al., A Causal Approach in Ingenuity Pathway Analysis 2014]. The list of differentially expressed genes and the corresponding data measurements identified under different experimental conditions were uploaded to the application.

[0291] The differentially expressed genes and corresponding fold changes are subjected to a filtering process to select only genes that are significantly perturbed, as indicated by their fold changes compared to the positive control.

[0292] The fold change threshold was set to encompass values ​​below -1.5 and above +1.5, with statistical significance indicated by a p-value of 0.05 or less.

[0293] Available identifiers were mapped to their corresponding entities in the QIAGEN Knowledge Base.

[0294] By starting with a "core analysis," significantly perturbed DEGs, called Network Eligible Molecules, were overlaid onto a global molecular network developed from information contained in the QIAGEN Knowledge Base. Networks of Network Eligible Molecules were then algorithmically generated based on their connectivity.

[0295] The "Core" analysis provides a comprehensive list of approximately the top 500 biological activities derived from the generated network. Biological activity-gene associations are always supported by corresponding annotations in scientific peer-reviewed publications that demonstrate the directionality and magnitude of the calculated biological activity modulation through automated association z-scores [Kramer et al. (2014)]. Essentially, this value represents a statistical metric that assesses the similarity between the observed pattern of differentially expressed genes (DEGs) and the expected pattern based on existing literature for a given annotation.

[0296] Step 2.1: Selection of biological activities relevant to the specific pathology under investigation. Selection of biological activities was based on identified hallmarks of the pathology of interest, applying a fold-change threshold set to encompass values ​​below -1.5 and above +1.5, with statistical significance indicated by a p-value of 0.05 or less.

[0297] Step 2.2 The associated Z-score values ​​(Figure 9 (values ​​in column cluster a)) were used to indicate the direction and magnitude of modulation of each biological function, and the values ​​were converted to absolute values ​​and summed to obtain the benefit score (step 2.3).

[0298] 9.2 Risk Score Calculation 9.2.1 Defining DIBASE Side Effects (Step 3.1) Officially recorded adverse events in DIBASE were identified by examining specific sources: -https: / / www.torrinomedica.it / schede-farmaci / dibase / (document made available by AIFA on 07 / 13 / 21) -https: / / www.drugs.com / sfx / cholecalciferol-side-effects.html as last updated on March 17,2024

[0299] 9.2.2 Identifying IPA biological functions from terms used to annotate side effects (step 3.2) Using information found in the aforementioned resources, identify the biological activity associated with the identified adverse reaction (the IPA source used is the Ingenuity Knowledge Base, which includes curation from journal articles, OMIM, JAX, and ClinicalTrials.gov) and investigate the IPA using the following procedure. Side effect terms are written one by one into the "Disease and Function" query box, and then the search is initiated.

[0300] The resulting resumption table allowed filtering diseases / functions arising from a large body of evidence (it is possible to understand the correspondence between IPA biological activities and annotated side effects, as shown in Figure 8). The creation of an in silico model was performed to reduce each biological function to a defined number of genes whose modulation can affect the modulation of the biological function itself (which in this particular case can represent side effects).

[0301] 9.2.3 Overlay analysis and construction of in silico models (step 3.3) The overlay analysis focuses on biological activities identified by the "in silico side effect model." The selection of biological activities is based on the identified side effects of DIBASE.

[0302] An "overlay analysis" is constructed by establishing a relationship between patterns of differentially expressed genes and selected biological activities using the following procedure (always supported by corresponding annotations in scientific peer-reviewed publications demonstrating the directionality and magnitude of modulation of biological activity): Import the set of biological activities selected from the in silico models of the SIDE EFFECT PROFILE in DIBASE into a new sheet called "my pathway".

[0303] Utilizing the "Build Tool" and "Grow Tool", differentially expressed genes (DEGs) belonging to the transcriptome profile under investigation (DIBASE and Osteoredux) are identified and associated with the regulation of the biological activity selected in the previous step.

[0304] 9.2.4 Image analysis of the color intensity of the obtained biological activity (3.4) Modulation of identified DEGs is depicted using green (indicating down-modulators) and red (indicating up-modulation) colors.

[0305] To determine the predicted calculated impact of such experimentally observed gene expression modulations on biological activity, the "Overlay" and "Molecule Activity Predictor" tools (MAP) are used. The "Prediction" function is activated within the MAP tool to calculate the predicted resulting modulation of biological activity. Thus, a color coding is established: -Orange: Increased activity - Blue: decreased activity -White: unattainable / unpredictable

[0306] "Overlay analysis" does not directly calculate the Z-score of each biological function. Therefore, it is necessary to convert the color intensity of the modulation signal into a numerical value. This is achieved by using the dedicated app: "IPAmmap_Parser" (version 2.1-1).

[0307] The app is a web port of Pipeline Pilot, designed to assign scores, called z-scores, to genes and biological functions based on coloring within biological pathways generated by QIAGEN's Ingenuity Pathway Analysis software. A key step in the algorithm is the conversion from the RGB color model to the LAB [https: / / www.xrite.com / it-it / blog / lab-color-space] model, a colorimetric encoding that allows for recording of color intensity as well as RGB composition. This conversion is performed within Pipeline Pilot's "components," which utilize procedures written in R software, relying on specific functionality from the colorspace package [https: / / cran.r-project.org / web / packages / colorspace / index.html].

[0308] Figure 9 shows the final list of biological activities and corresponding values ​​obtained for each treatment under investigation.

[0309] 9.2.5 Calculating the Global Risk Score (Step 3.5) Thus, the risk score is calculated by the sum of each positive value obtained from the overlay analysis (to indicate that a given side effect is induced by the treatment).

[0310] Calculation of the benefit scores for Osteoredux and DIBASE treatments (Figure 10 shows the risk scores obtained by summing each annotated biological activity value for each of the treatments under investigation).

[0311] Calculation of the benefit / risk score was performed by dividing the resulting benefit score by the resulting risk score (see Figures 11A and B).

[0312] The benefit / risk score for product B was calculated using an additional step: the introduction of a numerical factor for the importance of the hallmark.

[0313] 9.3 Benefit Score Calculation 9.3.1 Core Analysis Physiological Function Selection and Export Using the core analysis obtained in 7.1.3, biological activities were selected based on their tendency to correspond to the therapeutic indication of the product under investigation (i.e., bone remodeling). The biological activities were grouped into activity hallmarks, and each hallmark was assigned a weighting factor based on its importance in the pathogenesis of the target disease. Therefore, the biological activity Z-score values ​​were converted to absolute values, and the resulting list was exported to a table reporting the biological function specifications and their relative regulatory values. In this case, the biological activity values ​​were multiplied by the corresponding numerical importance factor according to step 2.1a, thereby obtaining the final list values ​​used for the final calculation of the benefit score (Figure 9 (values ​​in column b of cluster b)).

[0314] The sum of each value for each treatment provided the benefit score (Figure 9).

[0315] Risk scores were assessed as previously indicated.

[0316] 9.4 Benefit / Risk Score Calculation Once the risk and benefit scores are calculated, they are applied to the following formula: JPEG0007758403000015.jpg12165 (see Figures 11C and D).

[0317] Therefore, the value of the ratio must be interpreted as follows: the higher the value obtained, the safer the treatment applied, since the benefits outweigh the risks.

[0318] Figure 11 summarizes the results of the benefit / risk scores obtained with Osteoredux and DIBASE treatment (A). Calculation of the fold change of the score obtained with Osteoredux compared to the score obtained with DIBASE: Osteoredux score is 3 times higher than DIBASE (B).

[0319] When calculating the benefit / risk score taking into account the optional steps described above (numerical coefficients of importance), the results obtained are shown in Figure 11C, showing the calculation of the fold change of the score obtained with Osteoredux compared to the score obtained with DIBASE: the Osteoredux score is 6 times higher than that of DIBASE (Figure 11D).

[0320] Therefore, benefit / risk scores calculated with and without the optional step are consistent with each other.

Claims

1. 1. A computer-implemented method for providing a benefit / risk score for a therapeutic product of interest or for the manufacture of a therapeutic product of interest, comprising determining the benefit / risk score based on data obtained by transcriptomics analysis, the method comprising steps 1 to 4 of:

1. Perform transcriptomic analysis by: 1.1 providing a sample of a biological substrate representative of the pathology or pathological state to be treated by the product of interest; a. Treating one or more of said samples (hereinafter referred to as sample a) with said product of interest; b. Treating one or more of said samples (hereinafter referred to as sample b) with a reference agent for treating said disease state or pathological condition; and c. One or more samples of said biological substrate (hereinafter referred to as sample c) are used as relevant controls; 1.2 Extract RNA from each of the samples a, b and c; 1.3 performing a computational analysis of the transcriptome raw data from the RNA extracted in step 1.2 to identify, for each of samples a and b, differentially expressed genes (DEGs) relative to sample c and quantify their expression fold changes relative to sample c, thereby obtaining a fold change value for each of the DEGs (where the DEGs relative to sample c for each of samples a and b are genes whose expression is differential in samples a and b compared to sample c); 1.4 performing pathway enrichment analysis and functional analysis of the transcriptome data on the fold change values ​​obtained for each of the DEGs obtained by the transcriptome raw data analysis in step 1.3, thereby obtaining a numerical value representing the variation in magnitude and directionality of one or more biological activities associated with the differential expression of the DEGs in each of samples a and b normalized to sample c; 2. The computer determines the benefit score by: 2.1 selecting, from the one or more biological activities of step 1.4, one or more biological activities relevant to the therapeutic indication of the product of interest; 2.2 Identify a value that represents the variation in magnitude and direction of the one or more biological activities selected in step 2.1 and convert the value to an absolute value. 2.3 summing the absolute values ​​obtained in step 2.2 for each of the one or more biological activities, thereby obtaining a benefit score for the product of interest; 3. The computer determines the risk score by: 3.1 Provide a list of known side effects associated with the reference drug for the treatment of said condition 3.2 Determine one or more biological activities associated with said side effects by pathway and functional analysis; 3.3 For samples a and b, construct an in silico model of risk using pathway and functional analysis by establishing a relationship between the gene expression pattern obtained in step 1.3 and one or more biological activities determined in step 3.2, where the gene expression pattern is the associated fold-change value for each DEG and the combination of DEGs; 3.4 Input the associated fold change values ​​of each DEG in samples a and b obtained in step 1.3 into the in silico model of said risk obtained in step 3.3, and use pathway analysis and functional analysis to obtain data representing the direction (positive or negative) and magnitude of regulation for each of the one or two biological activities determined in step 3.2, and convert the data into corresponding absolute values; 3.5 The absolute values ​​obtained in step 3.4 are summed to obtain a final value representing the risk score of the product of interest being inspected; if the final value is less than a predetermined positive minimum value, the final value is automatically corrected to the predetermined positive minimum value.

4. A computer provides a benefit / risk score for the product of interest as a ratio of the benefit score obtained in step 2.3 to the risk score obtained in step 3.5, to determine and manufacture the therapeutic product of interest based on the benefit / risk score of the therapeutic product of interest.

2. The method of claim 1, wherein the pathway enrichment analysis and functional analysis of transcriptomics are performed using commercial or open source bioinformatics tools, such as QIAGEN Ingenuity Pathway Analysis (IPA), MetaCore (Clarivate Analytics), GeneGo (MetaCore), Qiagen OmicSoft, Elsevier Pathway Studio, Ingenuity Variant Analysis (IVA, Qiagen), DAVID (Database for Annotation, Visualization, and Integrated Discovery), GSEA (Gene Set Enrichment Analysis), STRING (Search Tool for the Retrieval of Interacting Genes / Proteins), KEGG Pathway Analysis, Metascape, Cytoscape with enrichment plugins (e.g., ClueGO, ReactomeFI), etc.

3. The method described in claim 1, wherein the fold change value obtained in step 1.3 is expressed as a Z-score value.

4. 10. The method of claim 1 further comprising: In Step 2.1, as Step 2.1.a, clustering the one or more biological activities selected in Step 2.1 into hallmarks of a disease state or pathological condition treated by the product of interest and providing a numerical coefficient of importance for each of the hallmarks; and In step 2.2, each of the numerical values ​​representing the variation in the magnitude and direction of the biological activity obtained in step 1.4 is multiplied by the numerical coefficient of the clustered hallmark provided in step 2.1.a, and the resulting values ​​are converted into absolute values.

5. The method of claim 4 , wherein the numerical importance coefficient is determined based on a frequency of association with a pathological condition, a statistical association, or a machine learning-based feature importance.

6. 5. The method of claim 4, wherein step 2.1.a and / or step 3.2 are performed using a machine learning model trained on a transcriptome dataset, such as a neural network or a support vector machine trained on a known transcriptome dataset.

7. 10. The method of claim 1 further comprising: assigning a numerical importance coefficient to each side effect in the list in step 3.1 and multiplying the fold-change value for each of the DEGs obtained in step 1.3 associated with each side effect by the corresponding numerical importance coefficient; However, the numerical coefficient of the importance is determined by the following parameters: The severity of the side effect according to the clinical classification system, its frequency in clinical data, its impact on patient-reported quality of life, and the duration of the side effect; is calculated based on one or more of The fold-change value for each of the DEGs may be multiplied by the corresponding importance numerical coefficient to obtain a numerical value that may be used as a further fold-change value for each of the DEGs.

8. 8. The method of claim 7, wherein the numerical importance coefficient is calculated based on one or more of the following parameters: severity of the side effect according to a clinical classification system, impact on patient-reported quality of life, duration of the side effect.

9. 10. The method of claim 1, further comprising performing steps 1-4 for group b using a reference drug, thereby providing a benefit score and a risk score for the reference drug, and a benefit / risk score for the reference drug.

10. 10. The method of claim 1, further comprising providing a benefit / risk score for the therapeutic product of interest for treating a given pathology based on data previously obtained from an animal model. wherein providing a benefit / risk score for the therapeutic product of interest includes: I. Provide values ​​obtained from the following animal models that represent the condition: a. a group treated with the therapeutic product of interest; and b. A group treated with a reference drug for the treatment of the condition, and c. Relevant control group The values ​​obtained in step I represent each of the following parameters: treatment effect, behavioral changes indicating animal distress, and weight loss indicating animal distress, and are the values ​​observed in each of groups a and b, normalized to the control group c. II. Determine a benefit score by summing the values ​​representing the therapeutic effects to arrive at a benefit score for the product of interest under test. III. Determining the Risk Score III.1 Identify the behavioral changes and weight loss that indicate distress in animals, as quantified in Step I, as risk parameters. III.

2. Summing the values ​​from step III.1 (for group a) to obtain a final value representing the risk score of the product of interest; where if the final value is less than a predetermined positive minimum value, it is automatically corrected to the positive minimum value. IV. Provide a benefit / risk score for the product of interest as the ratio of the benefit score obtained in Step II to the risk score obtained in Step III.

2.

11. 11. The method of claim 10, wherein the method is based on existing data obtained from an animal model, and the numerical value provided in step I. is derived from data previously obtained for the animal model.

12. 11. The method of claim 10, further comprising performing steps I-IV for group b using the reference drug, thereby providing its benefit score and risk score, and its benefit / risk score.

13. 1. A method for selecting one or more new therapeutic or beneficial products for clinical development, comprising: Implementing all of steps 1 to 4 defined in claim 1 by a computer to provide a benefit / risk score for one or more therapeutic products of interest for the treatment of a given disease state or pathological condition and a reference drug for the treatment of said disease state or pathological condition, wherein at least one of said therapeutic products of interest is selected as one or more new therapeutic or beneficial products for clinical development if the provided benefit / risk score for at least one of said therapeutic products of interest is equal to or greater than the benefit / risk score provided for said reference drug.

14. 14. The method of claim 13, further comprising providing a benefit / risk score for the one or more therapeutic products of interest and the reference drug by performing all of the steps defined in any of claims 10 to 12, wherein if each of the benefit / risk scores provided for at least one of the therapeutic products of interest is equal to or greater than each of the relative benefit / risk scores provided for the reference drug, then at least one of the therapeutic products of interest is selected as one or more new therapeutic or beneficial products for clinical development.

15. A method or screening method for the manufacture of a therapeutic or beneficial product, comprising a computer performing steps 1 to 4 defined in claim 1 to assess its benefit / risk score.

16. 20. The method of claim 1 or 15, wherein the therapeutic or beneficial product of interest is a product comprising one or more natural matrices, such as natural matrices obtained from eukaryotic or prokaryotic sources, such as plants, marine, bacterial, fungal, yeast, animal, or natural sources.

17. 17. The method of claim 16, wherein the natural matrix is ​​selected from one or more of: chopped or crushed plant parts, plant extracts, fractions of said extracts, microorganisms, honey, propolis, silk, wax, plant resins, plant gums, plant exudates, plant oils, plant essential oils, animal tissue lysates, or plant or animal body fluids.

Images