Generating neuron models for personalized drug therapy

Abstract

A computer-implemented method for generating neural models for selecting personalized drug therapy for a patient, comprising: receiving allele information for at least one neurophysiological coding region of a patient's genome; receiving a physiological model of a disease associated with a patient's phenotype; identifying a set of ion channels correlated with the allele information from an ion channel database; receiving a set of physiological measurement ranges, each physiological measurement range corresponding to a specific ion channel from the identified set of ion channels; and performing a simulation to generate multiple neural models that incorporate the set of ion channels with parameter values ​​within the corresponding physiological measurement ranges.Analyzing the generated neural models to identify components that influence the physiological model of a disease; and selecting a drug for the patient, at least in part, based on the identified components.

Description

TECHNICAL AREA The present invention relates to predicting drug effects on a neurophysiological feature in a patient given available information on the patient genotype by using neuronal model simulation(s) and determining a therapy decision / strategy specifically for the patient. BACKGROUND Neurophysiological features are characteristics of neuronal activity that, in a healthy individual, typically occupy a certain range but may occupy a different range in a diseased state. For example, the firing rate of a specific neuron type within a brain region may be a certain rate in a healthy individual, but an increased firing rate for that neuron type in that brain region may be a hallmark of disease. It is desirable to select a medication that most reliably aims to restore this feature to the healthy range. In this context, documents have already been published. Document WO 2009 / 152484A2 describes methods and platforms for drug detection. These methods identify an agent that corrects a phenotype associated with a health condition or predisposition to a health condition. The methods also include the detection of a diagnostic cellular phenotype, the identification of a person's risk for a health condition, methods for reducing the risk of drug toxicity in a patient, and methods for identifying potential genes that may contribute to a disease. Furthermore, document US 2009 / 0306534A1 describes a method for predicting the success of a particular treatment for a psychiatric disorder, whereby the analysis of a person's behavior or brain function is performed, in particular using the analysis of a person's genome. SUMMARY One or more embodiments of the present invention comprise a computer-implemented method for generating neural models for selecting a personalized drug therapy for a patient. The method comprises receiving allele information for at least one neurophysiological coding region of the patient's genome. The method further comprises receiving a physiological model of a disease associated with the patient's genome. The method further comprises retrieving a set of ion channels correlated with the allele information from an ion channel database. The method further comprises receiving a set of phenotypic measurement regions, each phenotypic measurement region corresponding to a specific ion channel from the retrieved set of ion channels.The procedure also includes running a simulation to generate several neuronal models that exhibit the set of ion channels with parameter values ​​within the corresponding phenotypic measurement ranges. The procedure further includes analyzing the generated neuronal models to identify components that influence the physiological model. The procedure also includes selecting a drug for the patient, at least in part, based on the identified components. One or more embodiments of the present invention comprise a system for selecting a personalized drug therapy for a patient. The system includes a memory and a processor connected to the memory for data exchange. The processor receives allele information for at least one neurophysiological coding region of the patient's genome. The processor also receives a physiological model of a disease associated with the patient's genome. Furthermore, the processor retrieves from an ion channel database a set of ion channels that correlate with the allele information. The processor also receives a set of phenotypic measurement regions, each phenotypic measurement region corresponding to a specific ion channel from the retrieved set of ion channels.Furthermore, the processor performs a simulation to generate several neural models that exhibit the set of ion channels with parameter values ​​within the corresponding phenotypic measurement ranges. The processor then analyzes the generated neural models to identify components that influence the physiological model. Finally, the processor selects a medication for the patient, at least in part, based on these identified components. One or more embodiments of the present invention comprise a computer program product containing a computer storage unit with computer-readable instructions stored therein, wherein the computer-readable instructions can be executed by a processing unit to generate neural models for selecting a personalized drug therapy for a patient. The selection includes receiving allele information for at least one neurophysiological coding region of the patient's genome. The selection also includes receiving a physiological model of a disease associated with the patient's genome. The selection further comprises retrieving a set of ion channels correlated with the allele information from an ion channel database.The selection process also includes receiving a set of phenotypic measurement ranges, where each phenotypic measurement range corresponds to a specific ion channel from the identified set of ion channels. The selection process further includes running a simulation to generate multiple neural models that exhibit the set of ion channels with parameter values ​​within the corresponding phenotypic measurement ranges. The selection process also includes analyzing the generated neural models to identify components that influence the physiological model. Finally, the selection process includes choosing a medication for the patient, at least in part, based on the identified components. Additional technical features and advantages are realized through the techniques of the present invention. Embodiments and aspects of the invention are described in detail herein and considered as part of the claimed subject matter. For better understanding, reference is made to the detailed description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS The examples described in this document are better understood with reference to the following drawings and description. The components in the figures are not necessarily to scale. Furthermore, identical numbers denote corresponding parts in all different views. Fig. 1 shows an exemplary neuron model system according to one or more embodiments of the present invention; Fig. 2 illustrates an exemplary system according to one or more embodiments of the present invention; Fig. 3 illustrates a flowchart of an exemplary method for selecting a drug combination for a patient according to one or more embodiments of the present invention; Fig. 4 shows exemplary diagrams for a pair of features that are the target during such a parameter search according to one or more embodiments of the present invention; and Fig.Figure 5 shows exemplary diagrams illustrating the results of a partial least squares regression (PLSR) performed with neural models shown in Figure 4 according to one or more embodiments of the present invention. The diagrams shown herein are for illustrative purposes only. Many changes can be made to the diagram or the processes described therein without deviating from the inventive concept. For example, the processes can be carried out in a different sequence, or processes can be added, deleted, or modified. Furthermore, the term "connected" and variations thereof describe the existence of a transmission path between two elements and does not imply a direct connection between the elements without intervening elements / connections. All such changes are considered part of the description. DETAILED DESCRIPTION Various embodiments of the invention are described herein with reference to the accompanying drawings. Alternative embodiments of the invention may be developed without deviating from the scope of protection of this invention. Various connections and positional relationships (e.g., above, below, adjacent, etc.) are set forth in the following description and in the drawings. Unless otherwise specified, these connections and / or positional relationships may be direct or indirect, and the present invention is not intended to be limiting in this respect. Accordingly, a connection of units may refer to a direct or indirect connection, and a positional relationship between units may be a direct or an indirect positional relationship.Furthermore, the various tasks and process steps described herein can be integrated into a more comprehensive procedure or process with additional steps or functionality that is not described in full detail herein. The following definitions and abbreviations are to be used for the interpretation of the claims and the description. As used herein, the terms "comprising," "incorporating," "comprising," "has," "includes," or "containing," or any variation thereof, are intended to cover non-exclusive inclusion. For example, an arrangement, mixture, process, method, article, or apparatus that includes a list of elements is not necessarily limited exclusively to those elements but may include other elements not expressly listed or belonging to such arrangement, mixture, process, method, article, or apparatus. Furthermore, the term "exemplary" is used herein to mean "serving as an example or illustration." Any embodiment or configuration described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments or configurations. The terms "at least one" and "one or more" may be understood to include any integer greater than or equal to one, i.e., one, two, three, four, etc. The term "a plurality" may be understood to include any integer greater than or equal to two, i.e., two, three, four, five, etc. The term "connection" may include both an indirect "connection" and a direct "connection." The terms "approximately", "essentially", "about" and variations thereof are intended to encompass the degree of error inherent in measuring the respective quantity based on the equipment available at the time the application was filed. For example, "approximately" may encompass a range of ± 8%, 5%, or 2% of a given value. For the sake of brevity, conventional techniques related to creating and using aspects of the invention are not necessarily described in full detail herein. In particular, various aspects of data processing systems and specific computer programs for implementing the various technical features described herein are already known. Accordingly, for the sake of brevity, many details of conventional implementations are only briefly mentioned here or omitted entirely, without providing the already known system or process details. In neuroscience, a "neuron" is a cell capable of processing information via an active membrane that can generate and transmit electrical signals. Contact points between neurons, where information is exchanged, are called synapses. Ion pumps in the membrane create a concentration gradient. This gradient acts as an energy reservoir, used to establish and maintain a potential difference between the two sides of the membrane. Ion channels, such as aqueous pores formed by transmembrane proteins, are found throughout the membrane, allowing the transport of ions across it. There are many different types of ion channels. The basic function of ion channels is to allow the diffusion of ions across the membrane of neurons. A common characteristic of all ion channels is that they allow a high transmission rate of up to 10⁶ ions per second across the cell membrane. This flow of ions generates an electrical current on the order of 10⁻¹² to 10⁻¹⁰ amperes per channel. Such currents are large enough to produce rapid changes in the membrane potential and the electrical potential difference between the inside and outside of the cell. Since calcium and sodium ions are present extracellularly in higher concentrations than intracellularly, the opening of calcium and sodium channels allows these cations to enter the cell and depolarize the membrane potential.For analogous reasons, the cell interior becomes more negatively polarized, or hyperpolarized, when potassium leaves the cell through open channels or chloride flows into the cell. However, not all ions can flow through a given ion channel. One reason for this is that most ion channels have selective permeability for different ions. According to the Hodgkin-Huxley model, which is the de facto standard model for modeling ion flows, the gates, for example, may be controlled by voltage-sensitive particles. Most ion channels are gated, meaning they are capable of transitioning between conducting and non-conducting states. Channel gating can be induced by extracellular ligands, intracellular secondary messengers and metabolites, protein-protein interactions, phosphorylation, and other factors. Furthermore, many ion channels are controlled by another regulatory signal—the membrane protein itself. Voltage-gated ion channels respond to and modify changes in the membrane potential generated by the binding of neurotransmitters to ligand-gated ion channels at synapses. Mutations in key molecules of these channels can cause neuromuscular disorders. In many cases, distinct clinical symptoms or characteristics are observed due to the specific changes in channel activity caused by the mutations. For example, reduced activity of potassium channels in nerve fibers delays the repolarization of action potentials and decreases the stimulus intensity required to trigger them. Mutations of potassium channels with these effects underlie hereditary forms of myokymia, a spontaneous, involuntary, undulating movement of skeletal muscle resulting from abnormal spontaneous generation of action potentials in the peripheral nerve.A large number of different mutations in gene-coding subunits of the acetylcholine receptor cause congenital myasthenic syndromes, disorders associated with muscle weakness and fatigue. Furthermore, epileptic seizures are attacks resulting from the overly synchronized and excessive activity of large groups of brain neurons. Depending on the area and size of the brain regions involved in the abnormal electrical activity, the symptoms vary considerably, but can include altered or loss of consciousness, sustained or rhythmic muscle contractions, stereotyped gestural movements, and visual or somatosensory hallucinations. As previously described, neurophysiological features are characteristics of neuronal activity that, in a healthy individual, typically occupy a certain range but may occupy a different range in a diseased state. For example, the firing rate of a specific neuron type within a brain region may be a certain rate in a healthy individual, but an increased firing rate for that neuron type in that brain region may be a hallmark of disease. It is desirable to select a medication that most reliably aims to restore this feature to the healthy range. Such neurophysiological features are not identical in all individual units (e.g., single neurons or small microcircuit-forming blocks of brain activity) in a person's brain. For example, each neuron of a particular cell type in a specific brain region has a different value for this feature (e.g., a different firing rate) depending on the specific context and environment of the neuron. For each trait, there are various ways to modulate it by altering the properties of a neuron. Changes to different combinations of ion channel properties can modulate the trait in a specific direction for each individual unit. However, the most effective way to modulate the trait may differ for each individual unit. Understanding the most economical way to modulate the trait in a population of specific individual units allows for the selection / tailoring of an effective drug for the patient, achieving modulation within a healthy range. The one or more embodiments of the present invention address such technical challenges by performing neuronal model simulation(s) to determine the modulation of the trait. Furthermore, the properties of individual units vary across populations of specific units, yet they share a common origin in that they all derive from the individual's genotype. For example, a neuron of a particular type requires a specific complement of ion channels to function as a member of that class of neurons and to fulfill its corresponding role in the brain. Types of neurons can include, for example, sensory neurons, motor neurons, and interneurons. It should be noted that the classifications of neurons can differ in their various forms. Gene expression controls how these ion channels are assembled and inserted into the cell membrane. However, the precise expression of individual genes regulates this complement of ion channels differently between individuals, with a particular cell type being regulated in a different way. Differences in genotype can be represented as differences in the possible parameter values ​​that the components of a model with a single unit can assume. For example, a single neuron of a particular cell type may conform to the stereotypical behavior required for that cell type to perform its functional task in a variety of ways, utilizing a variety of ion channel combinations. The specific parameter combinations required to regulate a single unit within the functionally realizable range for a given genotype can be differentially affected by disturbances aimed at altering the neurophysiological properties of that genotype. Consequently, setting the parameter combinations accessible within a given set of parameters of a genotype can enable the identification of healthy trait values.The embodiments of the present invention address such technical challenges and enable the setting of parameter combinations to determine healthy feature values. The embodiments of the present invention employ an evolutionary algorithm in conjunction with soft thresholding of error values, combined with a penalty term for crowding in the feature space, to determine which parameter combinations can generate the "healthy" region for a feature in a given genotype. The described embodiments improve performance, leading to increased speed in identifying and / or tailoring a drug for a patient. This performance improvement arises from the evolutionary algorithm (described below) for simulating neuronal models, which is a more efficient parameter search than the grid search typically used.The parameter search also improves the efficiency of tailoring a drug to a specific patient by providing a database of model parameter sets to analyze optimal control of the range of traits. The algorithm improves performance compared to other tailoring methods based on responses to specific drug treatments, trial and error, or machine learning of gene / expression data, which aim to predict drug efficacy directly from genetic information. Furthermore, the described embodiments of the invention improve the reliability of determining the parameter combinations for generating the "healthy" range for a trait.The reliability results from the ability to modulate multiple ion channels simultaneously along optimal control axes in the simulation, thereby enabling better control over a feature. Furthermore, embodiments of the present invention employ statistical modeling techniques such as partial least squares regression (PLSR) to regress parameter sets against the determined feature values ​​generated by the model with that parameter set. Accordingly, embodiments of the invention enable the identification of functional regulatory units that specify particular ion channel parameter combinations which, in the event of a disturbance, economically modify the desired feature value in most individual units of the genotype. Neuromuscular synaptic channel dysfunction illustrates how a wide variety of disease phenotypes can arise from mutations in channels that function together at a single anatomical site. The brain contains a far greater diversity of channels, and the role played by specific channels remains largely unknown. Embodiments of the present invention enable the analysis of the contributions of a large number of channel genes at the cellular and neuronal network levels. Furthermore, embodiments of the present invention allow the identification of preparations that modify the activity of individual channel types with higher specificity. For example, embodiments of the present invention can be used to diagnose a specific affective disorder (binocular switching task). Information about the performance of the task can then be used, together with genetic factors, to predict a drug response and determine a drug dosage based on the underlying assumptions of the theoretical model of a described bipolar disorder. Consequently, embodiments of the present invention allow the use of patient data (genotype) in combination with an unbiased approach to parameterize a simulation that automatically generates a therapeutic target. The embodiments of the present invention achieve this by modeling and simulating a system involved in the disorder and the parameters required to modulate the system back into a specific target range.The genotypic information is used to narrow down ranges of specific parameters in the simulation, and the drug / dosage is selected based on the simulation results, which identify targets with respect to neurophysiological properties, such as proteins or neurotransmitters, that are most likely to restore the simulated dysfunctional individual to a normal state. Therefore, instead of relying directly on genetic information in the drug search, embodiments of the present invention use the genetic information to generate and narrow down a computer model / simulation of the patient's neural models and subsequently analyze these neural models to identify drug targets. Fig. 1 shows an example of a neuron model system according to one or more embodiments of the present invention. Among other components, the system 100 comprises a genotype analyzer 110, an ion channel extractor 120, a neuron model simulator 130, a drug tailor 140, and a data repository 180. The data repository 180 comprises one or more databases, for example, an ion channel database, a drug efficacy database, and the like. The one or more components of the system 100 exchange data with the data repository via a wired and / or wireless data transmission network. System 100 receives as input a person's genotype, for example, that of a patient for whom a medication needs to be selected / tailored. The Genotype Analyzer 110 analyzes the input and extracts estimates of gene expression in relation to the set of membrane proteins of a nerve cell, defining a specific model of a single entity for the patient. These gene expression estimates provide ranges for parameters to generate a neuronal model for the patient. The term "genotype" refers to the alleles present in the patient's genomic DNA, where an allele is defined by the specific nucleotide(s) present at one or more specific locations in a nucleic acid sequence. A "genotype" is the nucleotide(s) present at a single polymorphic location that is known to vary within a population. The "genotype information" received as input is desired information concerning variations or changes in the genetic structure of a gene or locus of interest. Genotype information can indicate the presence or absence of a specific allele. Furthermore, "loci of interest" can be a gene, an allele, or a polymorphism of interest.Genes or loci of interest include genes that encode a) specific drug-metabolizing enzymes, b) specific drug transporters, c) specific drug receptors, d) enzymes, transporters or receptors that affect other drugs that interact with the medication in question, or e) body functions that affect the activities of the medication in question. Furthermore, the system receives as input disease-related characteristics that represent therapeutic targets of interest to the patient. Examples of disease-related characteristics that represent therapeutic targets of interest to the patient include the rheobase of striatal projection neurons, which is lower in Huntington's disease than in healthy controls; another is the firing rate of dopamine neurons in Parkinson's disease, which needs to be increased to compensate for the loss of dopamine neurons; and another is the burst / single-spike firing modes of dopamine neurons, which require modulation in schizophrenia.It should be noted that the above are examples of the various possible disease characteristics that can be received by System 100. Using the estimated parameter ranges determined from the individual's genotype, the Ion Channel Extractor 120 searches an ion channel database for ion channels that fall within "healthy" ranges of the characteristics of interest. In one or more examples, the Ion Channel Extractor 120 performs the search using an evolutionary algorithm with a combination of soft thresholding of error values ​​and density tightening in the feature space.“Soft thresholding” here refers to the elimination of errors, so that all ion channel parameters found to produce acceptable feature values ​​within the “healthy” range for that feature are considered equally valid and consequently all receive an error value of zero. The density straitforcer then influences the evolutionary search towards less densely populated regions of a feature space, thereby enabling a homogeneous selection of the “healthy” feature values. The Neuron Model Simulator 130 further analyzes the ion channels that have been determined to be located in the "healthy" feature space. In one or more examples, the Neuron Model Simulator 130 uses partial least squares regression (PLSR), a two-gate hyperplane normal algorithm, or any other such algorithm to regress feature values ​​against parameter combinations from the identified ion channels. For example, PLSR is a statistical algorithm similar to principal component regression that combines functions of principal component analysis (PCA) and multiple linear regression (MLR). PLSR enables the finding of a linear regression model that predicts a set of dependent variables (DV) from a set of independent variables (IV).This is achieved by projecting the independent and dependent variables onto new latent variable spaces that have the optimal predictive power. Unlike observable variables such as the conductance of ion channels, latent variables are variables that are not observed but instead inferred from other observables. In another example, a 2-gate hyperplane normal algorithm is used. Here, at least two sets of features (i.e., "gates") are sought through optimization, identifying sets of parameters that carry the label of the gate. For example, one gate might be labeled "wild type," while the other might be labeled "disease type." By inserting a hyperplane into the parameter space, thereby minimizing the loss of categorization of a parameter set by one label or the other, System 100 next represents the movement through the parameter space required to transform a neuron feature from "wild type" to "disease type" as a normal vector to the hyperplane.The hyperplane is defined such that it categorizes parameter sets based on the gate for which they were optimized and has a normal to the hyperplane in the direction of parameter sets derived from one of the gates, for example, the "wild type". In one or more examples, the hyperplane is defined using techniques such as the perceptron algorithm if the two (or more) gates are linearly separable. If the two gates are not linearly separable, multiple perceptrons can be used alternatively or additionally. Alternatively or additionally, a support vector machine algorithm can be used to define the hyperplane in one or more examples. The algorithms for defining the hyperplane can be implemented using an artificial neural network in one or more examples. Accordingly, the neuron model simulator determines a set of regression coefficients or normal vector coefficients with a normalized coefficient for each parameter in the system, which explain a maximum range of variation in the feature space. The neuron model simulator 130 outputs the set of regression coefficients that represent a specific combination of parameters that can change a feature value in the individual target unit. System 100 therefore allows for the prediction that a therapeutic intervention targeting precisely the parameters represented by the regression coefficients, in the specific ratio shown by the coefficients, will most likely be able to change the patient's characteristics in a desired direction. It should be noted that the prediction is personalized, meaning that the prediction may differ for different patients. The Device 140 for tailoring medications identifies a drug or combination of drugs that manipulate parameters based on the patient's genotype. For example, the Device 140 searches a drug database containing information on various drugs and the efficacy of each drug in influencing one or more channel-specific characteristics to identify the one or more drugs to be used that affect the determined parameters of the ion channels. Alternatively or additionally, the drug tailoring device 140 predicts the effects of a particular therapeutic drug / therapeutic combination on other characteristics, thereby predicting side effects of a treatment with respect to modifying alternative behaviors of a single entity by the treatment that were not the target of the therapeutic planning process. Fig. 2 illustrates an exemplary system 200 according to one or more embodiments of the present invention. The system 200 may be a data transmission device such as a computer. For example, the system 200 may be a desktop computer, a tablet computer, a laptop computer, a telephone such as a smartphone, a server computer, or any other device that exchanges data over a network. The system 200 comprises hardware, for example, electronic circuits. In one or more examples, the system 100 and / or each component of the system 100 may be represented by the system 200. Among other components, the system 200 comprises a processor 205, a memory 210 connected to a memory control unit 215, and one or more input units 245 and / or output units 240, such as peripheral or control units, connected for data exchange via a local I / O control unit 235. These units 240 and 245 may include, for example, battery sensors, position sensors (altimeter, accelerometer, GPS), indicator lights, and the like. Input units such as a conventional keyboard 250 and a mouse 255 may be connected to the I / O control unit 235. The I / O control unit may, for example, be one or more buses or other wired or wireless connections known in the prior art.To enable data exchange operations, the I / O control unit can have 235 additional elements such as control units, buffers (cache memory), drivers, amplifiers and receivers, which are omitted for the sake of simplicity. The I / O units 240, 245 may also include units that transmit both input and output signals, for example disk and tape storage, a network interface card (NIC) or a modulator / demodulator (for accessing other files, units, systems or a network), a radio frequency (RF) or other receiver, a telephone interface, a bridge, a router and the like. The Processor 205 is a hardware unit for executing hardware instructions or software, especially those stored in Memory 210. The Processor 205 may be a custom-built or commercially available processor, a central processing unit (CPU), an additional processor alongside various processors belonging to System 200, a semiconductor-based microprocessor (in the form of a microchip or chipset), a macroprocessor, or any other instruction-executing unit.The 205 processor includes a 270-byte cache memory, comprising an instruction cache for accelerating the retrieval of executable instructions, a data cache for accelerating the retrieval and storage of data, and a translation lookaside buffer (TLB) used to accelerate virtual-physical address translation for both executable instructions and data, but not limited to them. The 270-byte cache memory can be organized as a hierarchy of multiple cache memory levels (L1, L2, and so on). The memory 210 can comprise one or combinations of volatile memory elements (for example, random-access memory (RAM) such as DRAM, SRAM, SDRAM) and non-volatile memory elements (for example, a ROM, an erasable programmable read-only memory (EPROM), an electronically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a tape, a compact disc read-only memory (CD-ROM), a disk, a floppy disk, a plug-in module, a cassette, or the like). Furthermore, the memory 210 can include electronic, magnetic, optical, or other types of storage media. It should be noted that the memory 210 can have a distributed architecture, with various components located remotely, but accessible to the processor 205. The instructions in memory 210 can comprise one or more separate programs, each containing an ordered list of executable instructions for implementing logical functions. In the example of Fig. 2, the instructions in memory 210 comprise a suitable operating system (OS) 211. The operating system can essentially control the execution of other computer programs and provides scheduling, input / output control, file and data management, memory management, data exchange control, and related services. Additional data, including, for example, instructions for the processor 205 or other retrievable information, may be stored in memory 220, which is a storage unit such as a hard disk drive or a semiconductor storage medium. The instructions stored in memory 210 or in memory 220 may include those that enable the processor to execute one or more aspects of the systems and procedures described herein. Furthermore, the system 200 may include a display control unit 225 connected to a user interface or display 230. In some embodiments, the display 230 may be a liquid crystal display (LCD). In other embodiments, the display 230 may include a plurality of LED status lights. In some embodiments, the system 200 may also include a network interface 260 for connecting to a network 265. The network 265 may be an IP-based network for exchanging data between the system 200 and an external server, client, or the like via a broadband connection. In one embodiment, the network 265 may be a satellite network. The network 265 transmits and receives data between the system 200 and external systems. In some embodiments, the network 265 may be a managed IP network administered by a service provider.Network 265 can be implemented wirelessly, for example, using wireless protocols and technologies such as WLAN, WiMAX, or any others. Furthermore, Network 265 can be a packet-switched network, such as a local area network, a wide area network, a large area network, the internet, or another similar type of network environment. Network 265 can be a fixed wireless network, a wireless local area network (LAN), a wireless wide area network (WAN), a personal area network (PAN), a virtual private network (VPN), an intranet, or another suitable network system, and it can include devices for receiving and transmitting signals. Fig. 3 illustrates a flowchart of an exemplary method for selecting a drug combination for a patient according to one or more embodiments of the present invention. At 305, the method comprises receiving / accessing genotype information for at least one neurophysiologically relevant coding region of a patient genome. Examples include: allelic information of ion channel subunit genes, receptor subunit gene allelic information, etc. At 310, the method further comprises accessing a database of model parameter ranges of ion channels that correlate with the allelic information. The ion channel database is part of the data repository 180. Accessing the ion channel database comprises identifying a set of ion channels with a parameter within a defined measurement range.In one or more examples, two or more parameter measurements can be used to select the ion channels, for example, oscillation frequency and amplitude. The functionality involves configuring one or more measurement ranges for a neuronal model relevant to a phenotypic measure that is the target of one or more medications to be selected for the patient. For example, the phenotypic measure might be an oscillation frequency, oscillation amplitude, and similar characteristics of the neuron, encompassing the features being analyzed. It should be noted that in one or more examples, multiple sets of measurement ranges can be used to create a "gate." The measurement ranges can be entered by a user in one or more examples. In 320, the method further comprises performing a population-based evolutionary search with softmax tightening, wherein the search is performed in the population of ion channels identified in the phenotypic measurement space constrained by measurement ranges of the patient genotype. The resulting ion channels identified by the search are used in 330 to find neuronal models that generate phenotypic measures about the likely current and desired neuronal states of patients. The one or more embodiments of the present invention thus employ a search using an evolutionary algorithm to find ion channel parameters in neuronal models for generating a feature and also apply a population-based search with softmax thresholding to generate neuronal models for tailoring drugs.In one or more examples, the neural models are also configured as nerve tissue models and brain circuit models, with further optimization and phenotypic measures being derived from these combined models. Fig. 4 shows exemplary diagrams for a pair of features targeted by such a parameter search according to one or more embodiments of the present invention. In the example scenario shown, the features being searched for are the oscillation amplitude and frequency of a dopamine neuron in the substantia nigra pars compacta. A scatter plot 410 shows ion channels with different values ​​of the two features from the ion channel database (corresponding to block 310). Rectangle 415 shows the ion channels that can meet the search criteria, for example, ion channels that generate neuronal features such as oscillation amplitude and frequency, which represent the "healthy" range in which error values ​​are zero and selection is controlled by the tightness-strengthening mechanism.Furthermore, a scatter plot 420 shows the ion channels that are sought and selected from among the ion channels shown by rectangle 415. In this example, the density stiffener has influenced the evolution, so that the current generation includes models that lead to new models, which are generated approximately uniformly in this feature space, as shown by histograms 423 and 427. It should be noted that the diagrams in Fig. 4 are exemplary and that other diagrams can be generated in other examples, and other feature values ​​and measurement ranges can be used. Referring again to the procedure outline, the procedure at 340 also includes analyzing the neuronal models, for example, using PLSR, to identify components—that is, parameters of the neuronal model—that influence the traits analyzed for the patient. The analysis identifies components that can convert patient phenotypes into desired “healthy” states and their coefficients via ion channel parameters, ranked according to the contributions of each component to the trait. For example, extracellular K+, Na+, and Ca2+ ions can all influence the resting membrane potential of the neuron; furthermore, NALCN, in conjunction with UNC79 and UNC80, contributes to basal Na+ leak conductance in neurons. Consequently, a combination of one or more ion channels can contribute to a particular trait.The analysis performed identifies the parameters that influence the “healthy” measurements as determined by the subpopulation of ion channels (rectangle 415). Fig. 5 shows exemplary graphs illustrating the results of a PSLR performed with neural models shown in Fig. 4 according to one or more embodiments of the present invention. All parameter sets that generate a "healthy" combination of features are regressed from their feature values. This provides a set of coefficients for each feature that show how variations in the parameter space are likely to affect the feature. Graphs 510 and 515 show exemplary coefficient vectors for the amplitude and frequency features shown in Fig. 4. These coefficient vectors represent a prediction of the precise combination of intracellular parameters in the real neurons that should be regulated by a therapeutic agent to influence the frequency and amplitude features.Furthermore, Figures 520 and 525 show the predicted feature values ​​calculated by multiplying the normalized parameter values ​​of each neural model by the amplitude or frequency to predict the components shown in the upper part of the figure. Both figures show the same set of models; however, the predicted feature values ​​are orthogonal, indicating that independent control of these two features is possible and that a drug could target one or the other set of parameter coefficients responsible for independently controlling each feature. Referring again to the flowchart, the procedure at 350 also includes sequencing and grouping the one or more medications that must be selected based on their ability to target the complete set of coefficients determined by the analysis for the patient. The procedure also includes outputting the ranked list of medications at 360. This output may also include grouping the medications, with the grouping providing a combination of medications that collectively influence the parameters and characteristics. In one or more examples, selecting the medication for the patient based on the identified coefficients involves, at least in part, accessing a drug database to identify medications that modify the ion channels associated with the identified components. For example, if the identified coefficients apply to voltage-gated sodium (Nav1) channels, which play a key role in generating and propagating action potentials of sensory nerves necessary for transmitting pain signals, identified medications may include local applications of non-subtype-selective sodium channel blockers, such as novocaine, which provide pain relief through conduction block. Accordingly, the present invention enables the use of genotypic information to parameterize simulations, thereby allowing the prediction of the impact of various disorders on the individual. Furthermore, the present invention enables the establishment of a set of intracellular parameter limits from the allele information. Additionally, the present invention enables a process for selecting a drug / dose that does not rely directly on the existing phenotype, but rather on simulations that target the electrophysiological feature values ​​representing the desired effect of the drug. In one or more embodiments of the present invention, regression is used to derive the elements of the neural system that are to be modified by a drug in order to control a behavioral feature of this system that is representative of a patient. In other words, the present invention enables a simulation to be performed with a specific test drug, which is pre-selected as a potential therapeutic agent for modifying features in a particular tissue, for example, in the brain. The simulation predicts the drug's efficacy in altering the extent of receptor occupancy that the drug could influence in this tissue. The embodiments of the present invention thus enable predictions of suitable drug targets with respect to elements of the simulated system that can be regulated to produce robust feature modifications. The present invention enables computer-aided generation of neural models by first modeling a person's phenotype using a genotype-constrained mechanistic simulation and then performing an automated parameter sensitivity analysis to establish a recovery vector in the parameter space to transform the neural model from a diseased to a healthy state, whereby this can be used to select a drug or a 'cocktail' of combined drugs by another method that compares known drug targets with the recovery vector. It should be noted that the present invention is not specific to any particular neurophysiological model. Figures 4 and 5 illustrate, by way of example, a model of dopamine neurons of the substantia nigra; however, the present invention can be applied equally to any model of dopamine neurons, as well as to previously published models of any neuron type from any brain region affected by a neurophysiological dysfunction. The specific neurophysiological model used depends on the question or problem that the drug is intended to address. Furthermore, the generated neuronal models can include multiple ion channels, for example, 10 ion channels or 15 ion channels, and thousands and millions of such neuronal models are analyzed to determine the dosage of the drug combination for the patient.Consequently, the system described herein analyzes large amounts of data in a powerful manner, thereby providing an improvement in computer technology, particularly in systems for identifying personalized medicines. One or more embodiments of the present invention comprise a method for selecting a drug combination comprising at least one dose with a known physiological target based on a prediction of a clinically advantageous phenotypic change. The method comprises specifying the changes to parameters of a model of physiological components to represent the target of the drug combination. The method further comprises providing inputs from the model of physiological components to a simulation of a neuronal cell model and receiving at least one additional input from the model of physiological components.Furthermore, the procedure includes generating models that produce a range of observed phenotypes from multiple simulations of the neuronal cell model, driven by a population-based evolutionary search algorithm, while varying the parameters of the physiological component model. The procedure also includes analyzing the combined effects of parameter changes in the at least two physiological component models on the neuronal cell model by determining coefficients of parameter change components, using partial least squares regression of model parameter sets on the clinically beneficial phenotypic changes. Other algorithms may also be used to determine the coefficients.The procedure also includes selecting the drug combination based on maximizing the vector projection of an expected drug target change on parameters of at least two models of physiological components onto the coefficients of parameter changes that most strongly correlate with the beneficial phenotypic change. Furthermore, one or more embodiments of the present invention comprise a system for selecting a drug combination that includes at least one dose of a drug with a known physiological target, based on a prediction of a clinically advantageous phenotypic change. The system includes a searchable database of changes to parameters of a model of physiological components to represent the target of the drug combination. The system also includes simulation software for simulating the model of physiological components and for providing inputs from the model and at least one additional model of physiological components to a neuronal cell model.Furthermore, the system includes simulation hardware for repeatedly calculating models to generate a range of observed phenotypes from multiple simulations of the neuronal cell model, driven by a population-based evolutionary search algorithm that varies the parameters of the physiological component models. Additionally, analysis software within the system determines the combined effects of parameter changes in the at least two physiological parameter models on the neuronal cell model by calculating coefficients of parameter change components using partial least squares regression of model parameter sets on clinically beneficial phenotypic changes.Furthermore, the system includes ranking hardware for calculating partial least squares regression, evaluating the vector projection of an expected drug target change on parameters of at least two models of physiological components onto the coefficients of parameter changes that most strongly correlate with the beneficial phenotypic change, and creating a ranking of drug combination targets based on the evaluation. In one or more examples, the neuronal cell model provides inputs to a neural tissue simulation, and a measurement of the neural tissue simulation corresponds to the range of observed phenotypes and the beneficial clinical change. Alternatively or additionally, in one or more examples, the neuronal cell model provides inputs to a brain model simulation, and a measurement of the brain model simulation corresponds to the range of observed phenotypes and the beneficial clinical change. The present invention may be a system, a method, and / or a computer program product with any possible degree of technical integration. The computer program product may comprise a computer-readable storage medium (or media) containing computer-readable program instructions stored thereon to induce a processor to execute aspects of the present invention.A computer-readable storage medium can be a physical unit capable of retaining and storing instructions for use by an execution unit. For example, a computer-readable storage medium can be an electronic storage unit, a magnetic storage unit, an optical storage unit, an electromagnetic storage unit, a semiconductor storage unit, or any suitable combination thereof, without limitation. A non-exhaustive list of more specific examples of computer-readable storage media includes the following: a portable computer disk, a hard disk, random-access memory (RAM), read-only memory (ROM), and erasable programmable read-only memory (EPROM).Flash memory), static random-access memory (SRAM), portable compact storage disk-read-only memory (CD-ROM), DVD (digital versatile disc), USB flash drive, floppy disk, a mechanically coded unit such as punched cards or raised structures in a groove on which instructions are stored, and any suitable combination thereof. A computer-readable storage medium shall not, in its use herein, be understood as volatile signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission medium (e.g., light pulses traveling through an optical fiber cable), or electrical signals transmitted by a wire. The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to individual data processing units or, via a network such as the internet, a local area network, a wide area network, and / or a wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission lines, wireless transmission, routing computers, firewalls, switching units, gateway computers, and / or edge servers. A network adapter card or network interface in each data processing unit receives computer-readable program instructions from the network and forwards them for storage on a computer-readable storage medium within the respective data processing unit. Computer-readable program instructions for executing the steps of the present invention can be assembly instructions, ISA (Instruction Set Architecture) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., as well as conventional procedural programming languages ​​such as C or similar languages. The computer-readable program instructions can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server.In the latter case, the remotely located computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be established with an external computer (for example, via the internet using an internet service provider). In some embodiments, electronic circuits, including, for example, programmable logic circuits, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), can execute the computer-readable program instructions by using state information from the computer-readable program instructions to personalize the electronic circuits to perform aspects of the present invention. Aspects of the present invention are described herein with reference to flowcharts and / or block diagrams or charts of methods, devices (systems), and computer program products according to embodiments of the invention. It is pointed out that each block of the flowcharts and / or block diagrams or charts, as well as combinations of blocks in the flowcharts and / or block diagrams or charts, can be executed by means of computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a specialized computer, or another programmable data processing device to create a machine, such that the instructions executed via the processor of the computer or other programmable data processing device generate a means of implementing the functions / steps specified in the block(s) of the flowcharts and / or block diagrams or charts.These computer-readable program instructions may also be stored on a computer-readable storage medium capable of controlling a computer, programmable data processing device, and / or other units to function in a particular manner, such that the computer-readable storage medium on which instructions are stored has a manufactured product, including instructions that implement aspects of the function / step specified in the block(s) of the flowchart and / or block diagrams or charts. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or any other programmable data processing device to create a machine such that the instructions executed by the processor of the computer or other programmable data processing device will generate a means of implementing the functions / steps specified in the block(s) of the flowcharts and / or block diagrams or charts.These computer-readable program instructions may also be stored on a computer-readable storage medium capable of controlling a computer, programmable data processing device, and / or other units to function in a particular manner, such that the computer-readable storage medium on which instructions are stored has a manufactured product, including instructions that implement aspects of the function / step specified in the block(s) of the flowchart and / or block diagrams or charts. The computer-readable program instructions can also be loaded onto a computer, other programmable data processing device, or other unit to cause the execution of a series of process steps on the computer or other programmable device or other unit in order to generate a process executed on a computer, such that the instructions executed on the computer, other programmable device, or other unit implement the functions / steps specified in the block(s) of the flowcharts and / or block diagrams or charts. The flowcharts and block diagrams or charts in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this context, each block in the flowcharts or block diagrams or charts can represent a module, segment, or part of instructions that includes one or more executable instructions for performing the specific logical function(s). In some alternative embodiments, the functions specified in the block may occur in a different order than shown in the figures. For example, two blocks shown consecutively may in reality be executed essentially simultaneously, or the blocks may sometimes be executed in reverse order depending on the corresponding functionality.It should also be noted that each block of the block diagrams or charts and / or flowcharts, as well as combinations of blocks in the block diagrams or charts and / or flowcharts, can be implemented by special hardware-based systems that perform the specified functions or steps, or execute combinations of special hardware and computer instructions. A second operation can be considered "in response to" a first operation, regardless of whether the second operation results directly or indirectly from the first. The second operation can occur significantly later than the first and still be in response to the first. Similarly, the second operation can be considered in response to the first even if other operations occur between the first and second operations, and even if one or more of the intervening operations directly cause the second operation to occur. For example, a second operation can occur in response to a first operation if the first operation sets a marker and a third operation subsequently initiates the second operation each time the marker is set. To avoid using the expressions "at least one of" , , ... and <n>" or "at least one of< / n> , , ... <n> or combinations thereof" or "< / n> , , ... and / or <n>"To clarify and thus make publicly known, these terms shall be interpreted in the broadest sense as meaning one or more elements selected from a group comprising A, B, ... and N, superseding all other definitions implied herein, either before or after, unless expressly stated otherwise. In other words, the terms mean any combination of one or more of the elements A, B, ... or N, including any single element or one element in combination with one or more of the other elements, which may also include additional unlisted elements." It is also clear that every module, unit, component, server, computer, port, or entity mentioned herein as an example and executing instructions may include or otherwise have access to computer-readable media, such as storage media, computer storage media, or data storage units (removable and / or non-removable), such as magnetic disks, optical disks, or tape. Computer storage media may include volatile and non-volatile, removable and non-removable media implemented by any method or technology for storing data, such as computer-readable instructions, data structures, program modules, or other data. Such computer storage media may be part of the entity or accessible by or connectable to it.Each application or module described herein may be implemented using computer-readable / executable instructions that may be stored on such computer-readable media or otherwise cached. The descriptions of the various embodiments of the present invention are provided for illustrative purposes only and are not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and changes are apparent to those skilled in the art without altering the scope of protection of the described embodiments. The terminology used herein has been chosen to best explain the basic concept of the embodiments, their practical application, or technical improvements over commercially available technologies, or to enable other skilled individuals to understand the embodiments described herein.< / n>

Claims

A computer-implemented method for generating neural models for selecting personalized drug therapy for a patient, comprising: receiving allele information for at least one neurophysiological coding region of a patient's genome; receiving a physiological model of a disease associated with a patient's phenotype; identifying a set of ion channels correlated with the allele information from an ion channel database; receiving a set of physiological measurement ranges, each physiological measurement range corresponding to a specific ion channel from the identified set of ion channels; and performing a simulation to generate multiple neural models that incorporate the set of ion channels with parameter values ​​within the corresponding physiological measurement ranges.Analyzing the generated neural models to identify components that influence the physiological model of a disease; and selecting a drug for the patient, at least in part, based on the identified components. A computer-implemented method according to claim 1, wherein the set of physiological measurement ranges includes ranges corresponding to neuronal models that generate healthy neuronal responses. A computer-implemented method according to claim 2, wherein the set of physiological measurement ranges also includes ranges corresponding to neuronal models that generate pathological neuronal responses. A computer-implemented method according to claim 1, wherein the analysis of the generated neuronal models comprises performing a partial least squares regression using the ion channel parameter values ​​of the generated neuronal models and the physiological model. A computer-implemented method according to claim 1, wherein the simulation for generating multiple neural models uses an optimization that at least partially incorporates a soft thresholding of error values ​​and a penalty term for tightness. A computer-implemented method according to claim 1, wherein the physiological model of a disease has at least two feature values. A computer-implemented method according to claim 6, wherein the selection of the drug for the patient comprises accessing a drug database to identify drugs with ion channel alterations that are associated with the identified components, at least in part based on the identified components. A computer system for generating neural models for selecting a personalized drug therapy for a patient, comprising: a memory; and a processor connected to the memory for data exchange, the processor being configured to: receive allele information for at least one neurophysiological coding region of a patient's genome; receive a physiological model of a disease associated with a patient's phenotype; identify a set of ion channels correlated with the allele information from an ion channel database; receive a set of physiological measurement regions, each physiological measurement region corresponding to a specific ion channel from the identified set of ion channels;Performing a simulation to generate several neuronal models that feature the set of ion channels with parameter values ​​within the corresponding physiological measurement ranges; analyzing the generated neuronal models to identify components that influence the physiological model of a disease; and selecting a drug for the patient, at least in part, based on the identified components. System according to claim 8, wherein the set of physiological measurement areas comprises a first set of areas corresponding to neuronal models that generate healthy neuronal responses and a second set of areas corresponding to neuronal models that generate pathological neuronal responses. System according to claim 8, wherein the analysis of the generated neuronal models comprises performing a partial least squares regression using the ion channel parameter values ​​of the generated neuronal models and the physiological model. System according to claim 9, wherein the analysis of the generated neural models comprises using a 2-Gate Hyperplane Normal algorithm to determine a hyperplane between a first set of neural models corresponding to the first set of areas and a second set of neural models corresponding to the second set of areas. System according to claim 8, wherein the simulation for generating multiple neural models uses an optimization that at least partially features a soft thresholding of error values ​​in combination with a penalty term for density. System according to claim 8, wherein the physiological model of a disease has at least two feature values. System according to claim 13, wherein the selection of the drug for the patient, at least partially based on the identified components, comprises accessing a drug database to identify drugs with ion channel alterations that are associated with the identified components. A computer program product comprising a computer storage unit containing computer-readable instructions, wherein the computer-readable instructions are executable by a processing unit to generate neural models for selecting a personalized drug therapy for a patient, the selection comprising: receiving allele information for at least one neurophysiological coding region of a patient's genome; receiving a physiological model of a disease associated with a patient's phenotype; retrieving a set of ion channels correlated with the allele information from an ion channel database; receiving a set of physiological measurement regions, each physiological measurement region corresponding to a specific ion channel from the retrieved set of ion channels;Performing a simulation to generate several neuronal models that feature the set of ion channels with parameter values ​​within the corresponding physiological measurement ranges; analyzing the generated neuronal models to identify components that influence the physiological model of a disease; and selecting a drug for the patient, at least in part, based on the identified components. Computer program product according to claim 15, wherein the set of physiological measurement areas comprises a first set of areas corresponding to neuronal models that generate healthy neuronal responses and a second set of areas corresponding to neuronal models that generate pathological neuronal responses. Computer program product according to claim 15, wherein analyzing the generated neuronal models comprises performing a partial least squares regression using the ion channel parameter values ​​of the generated neuronal models and the physiological model of a disease. Computer program product according to claim 16, wherein analyzing the generated neural models comprises using a 2-gate hyperplane normal algorithm to determine a hyperplane between a first set of neural models corresponding to the first set of areas and a second set of neural models corresponding to the second set of areas. Computer program product according to claim 15, wherein the simulation for generating multiple neural models uses an optimization that at least partially features a soft thresholding of error values ​​in combination with a penalty term for density. Computer program product according to claim 15, wherein the physiological model has at least two feature values ​​and wherein selecting the drug for the patient, at least partially based on the identified components, includes accessing a drug database to identify drugs with ion channel changes that are associated with the identified components.

Images