Animal behavior analysis system, method, and program

Abstract

The present invention: explicates the connection between ultrasound influence on animal behavior and activity of cells and nervous system; develops a new approach for microscopically examining dynamic behavior of chemical substances and ion channels; and, for example, in the fields of drug discovery and gene therapy, reduces restriction of calculation resources and the high cost of molecular dynamics simulation.　The present invention is characterized by comprising: an observation execution step in which a sound parameter control unit controls a sound parameter, which is a characteristic of a sound, and outputs the sound toward a behavioral area of an animal, and an observation execution unit observes the behavior of the animal that received the outputted sound, and records the behavior as animal behavior information; and a simulation step in which an analysis unit analyzes the correlation between the sound parameter of the outputted sound and the behavior of the animal to acquire the analysis result, and a simulation means executes molecular dynamics simulation for a prescribed protein, by using the characteristic of the sound outputted toward the animal.

Description

Animal behavior analysis system, method, and program

[0001] The present invention relates to an animal behavior analysis system, method, and program that can observe the behavior of animals in response to high-frequency sounds such as ultrasound, and use the results to help with molecular dynamics simulations, drug discovery, and gene therapy development.

[0002] In drug discovery targeting proteins that form cells and organs in the human body, or proteins outside the body, it is necessary to explore the mechanisms of action of drugs on such proteins. For example, it is necessary to microscopically investigate the binding process and binding kinetics between the functional site of a protein and a ligand (e.g., a drug candidate molecule).

[0003] Conventionally, molecular dynamics (MD) simulation has been used as an important computational tool in such drug discovery (e.g., Patent Document 1). By using molecular dynamics simulation, it is possible to optimize the structure of a lead compound based on the results of the MD simulation, evolve it into a more effective drug, and discover potential binding sites present on the surface of a protein, thereby discovering new drug mechanisms of action and determining directions for drug design.

[0004] JP 2017-091179 A

[0005] However, molecular dynamics simulation requires large-scale computational resources, and there are certain limitations on the computational time, resources, and associated costs, especially when performing simulations of large molecular systems or long periods of time.

[0006] On the other hand, recent research has shown that ultrasound can affect animal behavior. It has become clear that one of the causes of animal behavior is the activity of cells and organs, particularly the nervous system, and research is progressing that suggests that this activity is due to, for example, the exchange of chemicals between nerve cells and the movement of ions within cells.

[0007] In recent years, molecular dynamics simulations have been used to analyze the dynamic behavior of these chemicals (e.g., neurotransmitters) and ion channels, which may lead to an understanding of how microscopic information affects animal behavior. Furthermore, many experiments have been conducted to evaluate how specific drugs affect animal behavior, and in these experiments, molecular dynamics simulations are useful for analyzing how drugs act on target cells and organs, and this information is thought to be useful for understanding the mechanisms by which drugs change animal behavior.

[0008] Therefore, the present invention has been made in consideration of the above problems, and aims to clarify the effects of ultrasound on animal behavior and the relationship with the activity of cells and the nervous system, develop a new approach to microscopically analyze the dynamic behavior of chemical substances and ion channels, and provide a method and system that can reduce the high cost and computational resource constraints of molecular dynamics simulations, for example in the fields of drug discovery and gene therapy.

[0009] (Animal Behavior Analysis System) In order to solve the above problem, the animal behavior analysis system of the present invention is characterized by comprising: an audio parameter control unit that controls audio parameters, which are characteristics of the audio output toward the animal's behavior range; an observation execution unit that observes the behavior of the animal that receives the audio output and records it as animal behavior information; an analysis unit that analyzes the correlation between the audio parameters of the output audio and the behavior of the animal; and a simulation means that obtains the analysis results by the analysis unit and performs a molecular dynamics simulation on a specified protein using the characteristics of the audio output toward the animal.

[0010] In the above invention, the sound parameter control unit preferably sets the target sound pressure as the specific sound parameter by combining volume or amplitude and frequency. Furthermore, in the above invention, the simulation means preferably places the specific protein under 1 atmosphere and applies a 1 / 100,000 fluctuation to the atmospheric pressure condition in the surrounding environment set in the molecular dynamics simulation to reproduce structural or functional changes in the cell or protein. The invention may also be configured to execute an object-based integrated psychology-physical simulation that correlates psychological activity changes of neurons based on the influence of the sound parameters with physiological parameters of the body and analyzes a set of psychological and physical states. The above invention preferably further includes a mechanism analysis unit that references a neural network trained by layering distribution patterns of features extracted from the correlation between the animal's behavior and structural or functional morphological changes of specific cells, receptors, or proteins. Based on the correlation analyzed by the analysis unit, the simulation means preferably references the neural network for animal behavior information correlated with the specific sound parameter and executes a molecular dynamics simulation for the extracted cells, receptors, or proteins.

[0011] In the above invention, it is preferable that the device further comprises a feedback control section that reflects the voice parameters used in the simulation in the control by the voice parameter control section based on the results of the simulation by the simulation means.

[0012] The above invention preferably further comprises: a drug discovery reference unit that references a drug discovery neural network trained by layering distribution patterns of features extracted from structural variations or functional morphological changes of specific cells, receptors or proteins and correlations between proteins and drug candidate molecules; and a drug candidate molecule analysis unit that selects drug candidate molecules by referring to the drug discovery neural network for animal behavior information correlated with specific acoustic parameters.

[0013] In the above invention, it is preferable that the simulation means models the structure of a specified protein that is the subject of the molecular dynamics simulation, acquires audio parameters of the audio output to the animal, calculates periodic external forces acting on the atoms of the specified protein and the water molecules surrounding them based on the audio parameters for the modeling, and adds the periodic external forces to the interactions between atoms of the specified protein and the surrounding environment, and integrates the changes over time in the relative positions and velocities between the atoms.

[0014] Furthermore, in the above invention, it is preferable that the neural network is trained by layering distribution patterns of features extracted from the correlation between the behavior of the animal and structural or functional morphological changes of specific cells, receptors or proteins in relation to the results of microbiome analysis of DNA extracted from an intestinal flora sample of the target animal, and that the mechanism analysis unit analyzes the correlation between the behavioral data of the animal, the voice parameters, the structural and functional changes of the proteins and the composition of the intestinal flora by referring to the neural network trained in relation to the results of the microbiome analysis.

[0015] (Animal behavior analysis method) Furthermore, the animal behavior analysis method of the present invention is characterized in that an audio parameter control unit controls audio parameters, which are characteristics of the audio, and outputs them toward the animal's range of behavior, an observation execution unit observes the behavior of the animal that has received the audio output and records it as animal behavior information, the analysis unit analyzes the correlation between the audio parameters of the output audio and the behavior of the animal, obtains the analysis results, and a simulation means performs a molecular dynamics simulation on a specified protein using the characteristics of the audio output toward the animal.

[0016] In the above invention, when controlling the voice parameters, it is preferable to set the target sound pressure by combining volume or amplitude and frequency.Furthermore, in the above invention, it is preferable that the simulation means reproduces structural or functional changes of cells or proteins by placing the predetermined protein under 1 atmosphere and applying a 1 / 100,000 fluctuation to the atmospheric pressure condition in the surrounding environment set in the molecular dynamics simulation, and may also execute an object-based psychology-physical integrated simulation that associates psychological activity changes of neurons based on the influence of the voice parameters with physiological parameters of the body and analyzes a set of psychological and physical states.

[0017] In the above invention, it is preferable that the mechanism analysis unit refers to a neural network trained by stacking distribution patterns of features extracted from the correlation between the animal's behavior and structural or functional morphological changes of specific cells, receptors or proteins, and that the simulation means refers to the neural network for animal behavior information correlated with specific acoustic parameters based on the correlation analyzed by the analysis unit, and performs molecular dynamics simulation on the extracted cells, receptors or proteins.

[0018] In the above invention, it is preferable that a feedback control section reflects the voice parameters used in the simulation in the control by the voice parameter control section, based on the results of the simulation by the simulation means.

[0019] In the above invention, it is preferable that the drug discovery reference unit refers to a drug discovery neural network that has been trained by layering distribution patterns of features extracted from structural variations or functional morphological changes of specific cells, receptors or proteins and the correlation between proteins and drug candidate molecules, and that the drug candidate molecule analysis unit refers to the drug discovery neural network for animal behavior information that is correlated with specific acoustic parameters to select drug candidate molecules.

[0020] In the above invention, it is preferable that the simulation means models the structure of a specified protein that is the subject of the molecular dynamics simulation, acquires audio parameters of the audio output to the animal, calculates periodic external forces acting on the atoms of the specified protein and the water molecules surrounding them based on the audio parameters for the modeling, and adds the periodic external forces to the interactions between atoms of the specified protein and the surrounding environment, and integrates the changes over time in the relative positions and velocities between the atoms.

[0021] Furthermore, in the above invention, it is preferable that the neural network is trained by layering distribution patterns of features extracted from the correlation between the behavior of the animal and structural or functional morphological changes of specific cells, receptors or proteins in relation to the results of microbiome analysis of DNA extracted from an intestinal flora sample of the target animal, and that the mechanism analysis unit analyzes the correlation between the behavioral data of the animal, the voice parameters, the structural and functional changes of the proteins and the composition of the intestinal flora by referring to the neural network trained in relation to the results of the microbiome analysis.

[0022] (Animal behavior analysis program) The above-described system and method according to the present invention can be realized by executing the program of the present invention, written in a predetermined language, on a computer. By installing the program of the present invention in an IC chip or memory device of a general-purpose computer such as a portable terminal device, smartphone, wearable terminal, mobile PC or other information processing terminal, personal computer or server computer, and executing it on the CPU, a system having the above-described functions can be constructed and the method according to the present invention can be implemented.

[0023] That is, the information processing terminal functions as: an audio parameter control unit that controls audio parameters, which are the characteristics of the audio output toward the animal's range of movement; an observation execution unit that observes the behavior of the animal that receives the audio output and records it as animal behavior information; an analysis unit that analyzes the correlation between the audio parameters of the output audio and the behavior of the animal; and a simulation means that obtains the analysis results from the analysis unit and performs a molecular dynamics simulation on a specified protein using the characteristics of the audio output toward the animal.

[0024] The program of the present invention can be distributed, for example, via a communication line, or transferred as a package application that runs on a stand-alone computer by recording it on a computer-readable recording medium. This recording medium can be recorded on a variety of recording media, including magnetic recording media such as flexible disks and cassette tapes, optical disks such as CD-ROMs and DVD-ROMs, and RAM cards. The computer-readable recording medium on which the program is recorded makes it possible to easily implement the above-described system and method using a general-purpose computer or a dedicated computer, and also makes it easy to store, transport, and install the program.

[0025] As described above, this invention clarifies the relationship between the effects of ultrasound on animal behavior and the activity of cells and the nervous system. This relationship is then used to microscopically analyze the dynamic behavior of chemical substances and ion channels in response to ultrasound on proteins and receptors in molecular dynamics simulations. For example, specific frequencies that affect animal behavior can be applied to target cells or nervous system proteins or receptors, and molecular dynamics simulations can be used to track their dynamic structural changes and movements, thereby examining the functional morphological changes of proteins and evaluating how ligands bind, how strong the binding is, and what their selectivity is. As a result, this invention can reduce the high cost and computational resource constraints of molecular dynamics simulations, for example, in the fields of drug discovery and gene therapy.

[0026] Furthermore, in controlling the sound parameters in the above invention, by setting the target sound pressure as a combination of volume (amplitude) and frequency, it is possible to precisely adjust the sound environment that is most effective for animal behavior, making it possible to efficiently elicit specific behavioral responses, and by optimizing the sound pressure, it is possible to obtain highly accurate data in behavioral analysis and observation while reducing the burden on the animal.

[0027] Furthermore, in the above invention, by reproducing the minute pressure fluctuations that proteins experience at 1 atmosphere in molecular dynamics simulation, it is possible to accurately simulate a normal physiological environment under real conditions and observe in detail the effects that fluctuations of 1 in 100,000 have on structural changes in cells and proteins, as well as on their functional changes, thereby realizing more realistic biological simulations and improving the reliability of data regarding molecular stability and dynamics.

[0028] In addition, the above invention can analyze the changes that sound parameters affect on the psychological activity of neurons and the physiological parameters of the body using an integrated psychological / physical simulation, making it possible to understand the psychological reactions that animals show to sound stimuli, the accompanying physical changes, and the correlation between them.For example, it is possible to quantitatively analyze how behavioral changes caused by specific frequencies or amplitudes affect physiological parameters such as the activity state of cranial nerves, heart rate, and breathing, and it is also possible to comprehensively understand the relationship between an animal's psychology and physical reactions, which is expected to contribute to elucidating the mechanisms of animal behavior based on sound stimuli and to applied research.

[0029] FIG. 1 is an explanatory diagram showing an overview of animal behavior analysis according to an embodiment. FIG. 2 is a block diagram showing the configuration of an animal behavior analysis device according to an embodiment. FIG. 3 is a block diagram showing the configuration of sound parameters provided in the animal behavior analysis device according to an embodiment. FIG. 4 is a block diagram showing the configuration of a simulation server and a new drug development server according to an embodiment. FIG. 5 is a block diagram showing the configuration of sound parameters virtually constructed on the simulation server according to an embodiment. FIG. 6 is a sequence diagram explaining animal behavior analysis according to an embodiment. FIG. 7 is a sequence diagram explaining the operation in sound parameter control in step S101. FIG. 8 is a sequence diagram explaining the operation in sound parameter control in step S203. FIG. 9 is an explanatory diagram explaining a neural network according to an embodiment. (a) is an explanatory diagram of a group of effective sound pressures of sounds used in experiments in an embodiment, and (b) is an explanatory diagram showing sound waveforms and their control interfaces for reproducing a target sound pressure.

[0030] (Outline of Animal Behavior Analysis) An embodiment of an animal behavior analysis system, method, or program according to the present invention will be described in detail below with reference to the accompanying drawings. Fig. 1 shows an outline of animal behavior analysis according to this embodiment.

[0031] In the animal behavior analysis of this embodiment, we observe how sounds with specific frequencies, such as ultrasound, affect animal behavior, and attempt to discover the activity of cells and organs, particularly the nervous system, that cause the behavior of animals affected by the sounds.We use molecular dynamics simulations to explore the correlation between the behavior of animals that receive specific sounds and, for example, the exchange of chemical substances between nerve cells or the movement of ions within cells.

[0032] Specifically, as shown in Figure 1, while controlling the sound parameters (frequency, amplitude, spectrum, etc.) that are the characteristics of the sound, the sound (acoustic sound) is output toward the animal's activity range, and the behavior of the animal that receives the sound output is observed and recorded as animal behavior information (F01). Here, the behavior of the animal that is exposed to sounds with specific sound parameters (frequency, amplitude, spectrum, etc.) is traced. This tracing observation can be done, for example, by filming the animal placed in a specified experimental field with a video camera or by recording biological information using a biological sensor.

[0033] Various experimental fields are used to evaluate various behavioral characteristics of laboratory animals, such as curiosity, anxiety, memory, and learning ability. Examples include the following: Circular open field: A large, flat, circular area, sometimes with a line separating the center from the outer edge. Experimental animals are left in the center of the field and observed for a set period of time. Measurements include time spent in the center and outer edge, distance traveled between the inner and outer areas, and running / movement speed. Square open field: A large, flat, rectangular area. Similarly, animals are left in the center of the field and observed for time spent in each corner or side, distance traveled, and running / movement speed. Elevated plus maze (EPM): A maze consisting of four cross-shaped arms. Each arm is varied, with two open arms and two surrounded by walls. Animals are left in the center and observed for time spent in the open and closed arms, number of arm visits, running / movement speed, and which arm they choose and how long they spend there. - Water maze: The field consists of a large tank filled with clear water. A platform is hidden below the water's surface. The animal is left in the water and the time it takes to find the platform, the distance traveled, and the speed at which it runs and moves are measured. - Barnes maze: A large, flat, circular area has several small holes, one of which is a hidden escape area. The animal is left in the center of the field and the time it takes to find the escape area, the number of holes visited, and the distance traveled are measured.

[0034] The following are some of the things that can be observed in behavioral analysis of this animal:・Basic behaviour The daily behaviour and lifestyle habits of animals, for example, patterns of eating, sleeping, and moving ・Social behaviour The interactions and hierarchy between animals within groups, communication methods, courtship behaviour and territorial behaviour ・Learning and memory How animals learn new skills and information, and how they remember it ・Senses and perception How each animal senses its environment and how it responds to it ・Biological clocks and rhythms Animals' circadian rhythms and seasonal changes in behaviour ・Responsiveness How animals react and adapt to changes in the environment and stressors ・Exploratory behaviour Animals' interest in and reactions to new environments and objects ・Predation and foraging How animals find, capture and ingest food ・Escape behaviour Strategies and techniques for escaping predators and dangers ・Reproductive behaviour Behaviour related to reproduction, for example, nest building, child-rearing methods, and protecting young ・Effects of disease and disability How disease and disability affect animal behaviour ・Effects of drugs and substances How various drugs and substances affect animal behaviour and neural activity

[0035] In the animal behavior analysis (F01) according to this embodiment, intestinal flora samples of the animal are collected periodically, such as by collecting fecal samples, and DNA is extracted from the collected samples, and the microbiome is analyzed using a next-generation sequencer.

[0036] Then, the correlation between the sound parameters of the output sound and the animal's behavior is analyzed, the analysis results are obtained, and the characteristics of the sound output to the animal are used to perform a molecular dynamics simulation of a specified protein (F02). For example, sound parameters when a specific behavior is performed are obtained, and the mechanisms of structural changes and functional morphological changes of the protein when the sound of these sound parameters is applied to a specified cell, receptor, or protein are analyzed.

[0037] Specifically, in the molecular dynamics simulation according to this embodiment, (1) the structure of a specific protein that is the subject of the molecular dynamics simulation is modeled, and the audio parameters of the sound output to the animal are acquired; (2) based on the audio parameters for the modeling, periodic external forces acting on the atoms of the specific protein and the water molecules around them are calculated; and (3) the periodic external forces are added to the interactions between atoms of the specific protein and the surrounding environment, and the temporal changes in the relative positions and velocities between the atoms are integrated.

[0038] In particular, in this embodiment, since the molecular dynamics simulation is performed when ultrasonic vibrations are applied to a protein, some special elements are added to the ordinary molecular dynamics simulation. The specific configuration and operation of such a simulation are described below.

[0039] First, to set the initial conditions, the protein is modeled and the environmental conditions are set. For example, the target protein structure is set, and then the protein is modeled based on experimentally obtained structural data (e.g., X-ray crystal structure analysis, NMR data), and an environment including water molecules, ions, and other molecules is set for the simulation. Next, to model the ultrasonic vibrations, the vibration parameters are set and an external force is applied. Specifically, the frequency, amplitude, and direction of the ultrasonic waves are defined to set the vibration parameters, and the temporal changes in the force applied to the protein are reproduced. Furthermore, during the simulation, external forces are applied, such as applying periodic external forces to protein atoms and nearby water molecules, based on the defined parameters.

[0040] We also perform force calculations and integrate equations of motion. This involves calculating the effects of external forces due to ultrasound, as well as the interactions between the protein and the atoms in the environment. We also use Newton's equations of motion to calculate the time evolution of the atomic positions and velocities, updating them sequentially in small time steps. We then calculate various physical quantities (e.g., structural changes, energy changes) to analyze how the vibrations affect the protein's structure, dynamics, and function. We then perform detailed analysis of the simulation results to understand the impact of ultrasonic vibrations on protein function and stability.

[0041] The simulations are then iterated and refined, for example by adjusting ultrasound parameters as needed and running simulations under different conditions, and the validity of the simulations is verified by comparing them with experimental data and other theoretical models. These simulations allow researchers to analyze the effects of ultrasound on protein structure and function, providing valuable information for ultrasound-based medical and biotechnology applications, such as drug discovery and gene therapy development.

[0042] In this embodiment, AI (artificial intelligence) is used to predict the correlation between behavior when hearing sound and structural changes or functional morphological changes of specific cells, receptors, or proteins, and a molecular dynamics simulation is performed in which sound (vibration frequency, vibration energy, etc.) is applied to a somewhat narrowed-down protein structure. This AI is equipped with a neural network trained by layering distribution patterns of features extracted from the correlation between animal behavior and structural changes or functional morphological changes of specific cells, receptors, or proteins, and refers to the neural network for animal behavior information correlated with specific sound parameters, and predicts and analyzes the mechanism of structural changes or functional morphological changes caused by virtually applying sound to the extracted cells, receptors, or proteins.

[0043] As shown in Figure 9, a neural network is a learning and recognition system that mimics the mechanisms of the human brain, with a multi-layered structure, particularly three or more layers, in which classifiers, which are neuron node units, are connected by edges with weighted coefficients. When data is input into this recognition system, the data is propagated in order from the first layer, and learning is repeated in order in each subsequent layer. During this process, the features of the data are automatically calculated.

[0044] These features are essential variables necessary for solving a problem and are variables that characterize a specific concept. In the animal behavior analysis of this embodiment, data is input, multiple feature points are hierarchically extracted from the data, and patterns are recognized based on the hierarchical combination patterns of the extracted feature points. As shown in the figure, the recognition function of the neural network is a multi-class classifier, and an object 501 containing specific feature points (here, for example, "animal behavior" and its corresponding "audio parameters") is input as input data, detected, and propagated to a recognition function module 50. This recognition function module 50 has an input unit (input layer) 503, a first weighting coefficient 504, a hidden unit (hidden layer) 505, a second weighting coefficient 506, and an output unit (output layer) 507.

[0045] At this time, a plurality of feature vectors 502 are input to the input unit 503. A first weighting coefficient 504 weights the output from the input unit 503. The hidden unit 505 performs nonlinear transformation on the linear combination of the output from the input unit 503 and the first weighting coefficient 504. A second weighting coefficient 506 weights the output from the hidden unit 505. The output unit 507 calculates a classification probability 508 for each output class (here, items related to "mechanism analysis," such as "structural variation" and "protein denaturation process"). Here, three output units 507 are shown, but this is not limiting. The number of output units 507 is the same as the number of events that the pattern classifier can detect. Increasing the number of output units 507 increases the number of events that the event classifier can detect, such as the cause of a failure.

[0046] Mechanisms analyzed by molecular dynamics simulations include: ・Structural Fluctuations: Track the dynamic structural changes and movements of proteins to investigate the functional conformational changes of proteins in detail. ・Ligand Binding: Analyze the binding process with ligands (e.g., drugs or cofactors) and the dynamic interactions involved in binding. ・Binding Free Energy: Calculate the thermodynamic stability of protein-ligand binding. ・Protein Folding: Study the folding and unfolding processes of proteins. ・Allosteric Effects: Investigate how changes in one part of a protein affect other parts of the protein. ・Hydration Dynamics: Investigate the movement and arrangement of water molecules on the surface of proteins and near active sites. ・Ion Channel Operation: Study the opening and closing mechanisms and ion passage mechanisms of ion channel and transporter proteins. ・Protein-Protein Interactions: Investigate the mechanisms of protein complex formation and disassembly in detail. ・Thermodynamic Properties: Evaluate the thermodynamic properties (e.g., enthalpy, entropy, etc.) associated with specific protein states and changes. Resonance frequencies: Detailed interpretation of structural and dynamic information is possible through comparison of molecular dynamics simulation results with experimental data from resonance Raman and NMR.

[0047] Furthermore, the mechanism analysis unit in Phase F02 may also analyze sound wave perception in the inaudible range. Specifically, for example, it may be configured with sound parameters for applying sound waves to cells other than the ear, and for the cells to sense sound waves via a piezo-type protein (e.g., Piezo1) and induce calcium ions. Because sound waves beyond the audible range may also affect cells, proteins such as Piezo1 may be used to analyze how cells sense and respond to sound waves. Simulations of molecular dynamics using sound stimulation emphasize this innovative approach, distinct from conventional drug experiments, and elucidate the mechanism by which sound action on Piezo1 induces a response in the body via calcium ions.

[0048] Furthermore, in this embodiment, the neural network is trained by layering distribution patterns of features extracted from the correlation between the behavior of the target animal and structural or functional morphological changes of specific cells, receptors, or proteins in relation to the results of microbiome analysis of DNA extracted from the target animal's intestinal flora sample. Correlation analysis and mechanism analysis then analyze the correlation between the animal's behavior data, the vocal parameters, the structural and functional changes of the proteins, and the composition of the intestinal flora, with reference to the neural network trained in relation to the results of microbiome analysis.

[0049] In the molecular dynamics simulation according to this embodiment, when the relationship between animal behavior and intestinal flora is also incorporated as a simulation element, the following additional components and processes are executed. First, in the animal behavior analysis (F01), a mechanism for periodically collecting intestinal flora samples of the animal, such as fecal sampling, is installed, DNA is extracted from the collected samples, and the microbiome is analyzed using a next-generation sequencer. Then, in the correlation analysis of phase F01, intestinal flora data analysis is performed. In this intestinal flora data analysis, microbiome data, which is the analysis result of the analyzed microbiome, is stored in database 107, and species identification, abundance measurement, diversity analysis, etc. are performed to analyze the correlation between the animal's behavior data, vocal parameters, structural and functional changes in proteins, and the composition of the intestinal flora.

[0050] Then, an expanded molecular dynamics simulation is performed by adding processing related to the intestinal flora. Here, modeling is performed taking into account the impact of changes in the intestinal flora on the structure and function of proteins and receptors, and an expanded simulation is performed that includes intestinal flora data, and the results are analyzed. Next, the animal's behavior, vocal parameters, structural and functional changes in proteins, and intestinal flora data are integrated to analyze the combined effects, and the integrated analysis results are displayed in an easy-to-understand format (e.g., graphs, charts, heat maps). Note that when adding processing related to the intestinal flora, the intestinal flora data may be added to the neural network's training data to learn more complex relationships, and the neural network model may be retuned based on the added data to make more accurate predictions.

[0051] The mechanistic analysis in Phase F02 also involves an integrated simulation of ultrasound and psychological and physical states. This integrated simulation is an object-based integrated psychological and physical simulation that links neuronal activity and physiological parameters via ultrasound to analyze sets of psychological and physical states. Its target organisms are expanded beyond animals to include primates such as humans and simple model organisms such as crickets. By targeting simple model organisms such as crickets and humans in addition to animals, the impact of "noise" in the simulation is reduced. For example, the effects of stress and stress can be understood through biomarkers such as lifespan variation in crickets.

[0052] Specifically, this integrated simulation integrates the effects of ultrasound as a "mental" and "physical" reaction through a psychology-physical interaction model, integrating them into a simulation that changes over time. In particular, neuronal activity (psychological aspects) is linked to molecular dynamics simulation (physical aspects), and state sets (blood pressure, heart rate, etc.) are used as indicators to dynamically change the simulation results. Here, object-based simulation utilizes state sets to realize an "integrated psychology-physical simulation" that explores causal relationships for each target, and analyzes the correlation between psychology and the body, focusing primarily on the effects of ultrasound.

[0053] These analyses can provide deep insights into the functional functions of proteins, their interactions with ligands, and even the effects and mechanisms of drugs. This allows for the simulation of the mechanical and chemical mechanisms of specific sounds, enabling, for example, the analysis of the dynamic behavior of neurotransmitters and ion channels, and the elucidation of how microscopic information influences animal behavior. Furthermore, in this embodiment, the effects of drugs on cells or organs exposed to specific sounds can be analyzed, providing an opportunity for the development of new drugs and gene therapies (F03).

[0054] In this embodiment, AI is used to predict the correlation between the structural or functional morphological changes of a specific cell, receptor, or protein and the protein and drug candidate molecule, and to select a narrowed-down list of proteins and drug candidate molecules. This AI is equipped with a drug discovery neural network trained by layering distribution patterns of features extracted from the correlation between the structural or functional morphological changes of a specific cell, receptor, or protein and the protein and drug candidate molecule, and selects drug candidate molecules by referring to the drug discovery neural network for animal behavior information correlated with specific sound parameters, and analyzes how the drug candidate molecule acts on cells or organs exposed to specific sounds, providing a foothold for the development of new drugs and gene therapies.

[0055] For example, the following can serve as opportunities for the development of new drugs and gene therapies: ・Elucidating the binding mechanism between proteins and ligands. Molecular dynamics simulations allow for microscopic observation of the interaction and binding process between proteins and ligands (e.g., drug candidate molecules). This allows for detailed information on binding strength, specificity, and binding sites. ・Evaluating the dynamic behavior of drugs. The dynamic behavior of drugs after binding to proteins can be evaluated, and how these changes affect drug activity. ・Optimizing lead compounds. Based on the results of molecular dynamics simulations, the structure of lead compounds (early drug candidate compounds) can be improved to achieve higher affinity and selectivity. ・Understanding the effects of resistance mutations. Some drugs are known to lose their effectiveness due to protein mutations. Molecular dynamics simulations can help understand how these mutations affect drug binding. ・Exploring new binding sites. In addition to conventional binding sites, molecular dynamics simulations can be used to explore potential binding sites on proteins, which can lead to the discovery of new drug mechanisms and the development of diverse drug design strategies. - ADMET (absorption, distribution, metabolism, excretion, toxicity) prediction By combining molecular dynamics simulations with other computational methods, it is possible to predict information about drug behavior and safety in the body.

[0056] Furthermore, based on the results of the molecular dynamics simulation in Phase F02 described above, or the design of new drugs or gene therapies in Phase F03, feedback can be performed to reflect the voice parameters used in the molecular dynamics simulation at that time in the control of the voice parameters in Phase F01 (F04).

[0057] For example, by feeding back the acoustic parameters used in molecular dynamics simulations, it is possible to analyze neural activity and behavior, such as the dynamic behavior of chemical substances and ion channels such as neurotransmitters between neurons, and to elucidate the action and mechanism of drugs. Examples of elucidate the action and mechanism of drugs include evaluating the effects of drugs in animal behavioral analysis, analyzing the binding and action of medicinal ingredients on proteins and receptors, and evaluating the mechanisms of behavioral changes in animals and their effects on physiological functions.

[0058] (Configuration of Animal Behavior Analysis System) A system for performing the above-described animal behavior analysis will now be described. Fig. 2 shows the configuration of the animal behavior analysis device 10 according to this embodiment, and Fig. 4 shows the configurations of the simulation server 3 and new drug development server 4 according to this embodiment.

[0059] (1) Animal Behavior Analysis Device 10 As shown in Figure 2, animal behavior analysis device 10 is a device that observes and records the behavior of animal A1, such as a mouse, and can be implemented using an information processing terminal such as a personal computer, tablet PC, smartphone, etc. Connected to this animal behavior analysis device 10 are a camera 11 for capturing images, a speaker 13 for audio output, and a receiver 12 for receiving signals from various sensors.

[0060] By executing the animal behavior analysis application of the present invention on an information processing terminal, various functional modules can be virtually constructed on the CPU of animal behavior analysis device 10. Specifically, animal behavior analysis device 10 includes an imaging control unit 101, a sensor control unit 102, a sound parameter control unit 105, and an analysis data transmission unit 108 as input / output system interfaces.

[0061] The imaging control unit 101 is an interface to which the camera 11 is connected, and controls the operation of the camera 11, acquires images captured by the camera 11, and inputs the image files together with timestamps to the video recognition unit 104. This camera 11 can be an ordinary visible light camera, an infrared camera, an ultraviolet camera, an underwater camera, or the like. The camera 11 is also equipped with an audio recorder that records sounds and communication of animals, and by analyzing the sounds and communication, it is possible to evaluate the social behavior and state of the animals.

[0062] The sensor control unit 102 is a communication interface to which the sensor 21 attached to the animal A1 is connected wirelessly or via a wire. It controls the operation of the sensor 21, acquires various detection signals detected by the sensor 21, and inputs these signals along with timestamps to the observation execution unit 103. The following sensors 21 can be used: RFID tag (radio frequency identification) A small tag is attached to the animal's body, recording information when the animal passes through a specific point, thereby tracking movement and activity patterns. Accelerometer A sensor for measuring an animal's movement and activity level, analyzing the animal's movements and behavior patterns. Heart rate monitor Used to evaluate a living organism's stress level and activity level, and can evaluate the animal's health and response. Temperature sensor Used to monitor an animal's body temperature, evaluating its health and adaptability to the environment. Underwater sensors Pressure sensors, water temperature sensors, hydrophones that capture underwater sounds, acoustic tags, and other sensors can be used to track an animal's behavior and movement underwater.

[0063] The voice parameter control unit 105 is a module that controls parameters of the voice output from the speaker 13. Specifically, the voice parameter control unit 105 controls frequency, the "pitch" of the voice, formants (resonance peaks), temporal features (voice duration and fixed phoneme intervals), spectral envelope (shape and intensity of frequency components in a spectrogram), energy (intensity and amplitude of the voice signal), zero-crossing rate, MFCC (Mel Frequency Cepstral Coefficients), spectral features, delta features (temporal changes in features), voice quality, acoustic spectrum (energy distribution in each frequency band of the voice), LTAS (Long-Term Average Spectrum), etc.

[0064] In particular, in this embodiment, the voice parameter control unit 105 has a function of reproducing a specific "sound pressure" as a part of the "specific voice parameter" by combining volume (amplitude) and frequency. Here, sound pressure refers to the pressure fluctuation that occurs when sound waves propagate through air or other media. Specifically, in the atmosphere, the generation of sound waves changes the pressure of the surrounding air, and this change is expressed as the magnitude (dB) of the change, and is usually measured as a ratio to a reference sound pressure (20 μPa).

[0065] 3, the sound parameter control unit 105 includes an interface control unit 105a, an environment setting unit 105b, a sound pressure setting unit 105c, an amplitude / frequency control unit 105d, and an output control unit 105e. The interface control unit 105a is a module that controls an interface for manually or automatically inputting a target sound pressure and a volume (amplitude) or frequency set to reproduce that sound pressure, and is connected to a keyboard 15, a display device 14, etc. as a user interface. The interface control unit 105a also functions as control means that sequentially calculates and automatically sets or presents other parameters required to reproduce the target sound pressure based on the calculation results by the sound pressure conversion unit 109a in conjunction with input settings such as volume or frequency.

[0066] The environment setting unit 105b is a module that acquires and records information about the environment in which the sound is to be output. This environmental information includes not only arbitrary input values ​​by the operator but also the current atmospheric pressure, temperature, and humidity detected by various sensors. The sound pressure setting unit 105c is a module that sets the sound pressure of the sound to be output in the experiment, and sets a sound pressure that has a fluctuation of 1 / 100,000 of the current atmospheric pressure conditions included in the environmental information acquired by the environment setting unit 105b. This atmospheric pressure condition is usually 1 atmosphere. As shown in Figure 10(a), the sound that needs to be output in the experiment is included in an effective sound pressure group. This effective sound pressure group is set to fall within a fluctuation range of 1 / 100,000 of the current atmospheric pressure (usually 1 atmosphere). An arbitrary sound pressure is selected from this effective sound pressure group and set as the target sound pressure.

[0067] The "fluctuation of 1 / 100,000 relative to atmospheric pressure conditions" refers to a state in which air fluctuations can occur in quiet environments equivalent to a sound pressure (dB) of 1 / 100,000 of 1 atmosphere, and is a quantitative air pressure fluctuation (pressure fluctuation) that can be correlated between Pa and dB using the conversion formula above. While sound pressure is also measured in Pascals, sound pressure level (dB) is measured on a logarithmic scale and used to indicate perceived intensity. The equivalent amount can be calculated using the conversion formula from air pressure (Pa) to sound pressure level (dB): "1 atmosphere (Pa) = 20 μdB = 2.5 × 10 ?? dB." For example, in quiet environments such as libraries, sound pressure levels are typically below 15 dB. While this is a very small fluctuation in sound pressure level (dB), in terms of atmospheric pressure, it is a pressure fluctuation equivalent to "approximately 1 / 10,000 of 1 atmosphere," which corresponds to a volume that can produce a sound pressure effect that resonates with proteins.

[0068] The amplitude / frequency control unit 105d is a module that sets the range of amplitude and frequency when reproducing the specific target sound pressure set by the sound pressure setting unit 105c by combining volume or amplitude and frequency. For example, when determining the amplitude (volume) and frequency for reproducing the sound pressure required in an experiment, the module provides a user interface for setting the upper (or lower) limits of the volume and frequency that can be output in order to suppress the output volume. In this case, either the amplitude or the frequency may be selected, and the reproducible sound pressure may be calculated from the selected amplitude or frequency.

[0069] The output control unit 105e is a module that generates a frequency WAV file to actually reproduce a specified sound pressure and outputs it using an output device such as a speaker at a specified volume and frequency. In particular, in this embodiment, it is equipped with an interface that allows for fine volume and frequency control of the hardware so that a fluctuation of 1 atmosphere + 1 / 100,000 can be achieved.

[0070] The voice parameter control unit 105 according to this embodiment also includes a sound pressure control unit 109 that sets a predetermined target sound pressure and controls the volume (amplitude) and frequency to achieve that target sound pressure. This sound pressure control unit 109 is a module that outputs sound to achieve the desired sound pressure according to the set sound pressure, volume, and frequency. In this embodiment, the frequency band output for experimental purposes can be set to a range that humans cannot hear (ultrasonic waves (20 kHz or higher) and low-frequency sounds (20 Hz or lower)) or a frequency band that is highly efficient and safe. To achieve this, the voice parameter control unit 105 amplifies the low-frequency range to ensure a certain sound pressure, and filters and compresses the high-frequency audible range, as shown in FIG. 10B, thereby achieving a sound pressure effect without excessively increasing the volume.

[0071] The "highly efficient and non-hazardous frequency bands" mentioned above include those within the human audible range (20 Hz to 20 kHz), particularly the 1 kHz to 4 kHz frequency band, which efficiently transmits energy in acoustic equipment and everyday sounds and are considered to be comfortable and safe for humans, while low frequencies below 20 Hz are difficult for humans to hear under normal use, have limited energy transmission, and are considered to have little impact on human health and safety. Generally, the acoustic band from 1 kHz to 4 kHz and ultrasound above 20 kHz, which is managed in medical settings, are frequency bands that can be used safely and efficiently.

[0072] Specifically, the sound pressure control unit 109 includes a sound pressure conversion unit 109a, a frequency filter unit 109b, a specific frequency amplification unit 109d, an antiphase sound generation unit 109c, and a compressor unit 109e. The sound pressure conversion unit 109a is a module that converts a desired sound pressure into an amplitude and frequency to be output by performing calculations based on a decibel-to-Pascal conversion formula to reproduce a target sound pressure as a specific sound parameter using a combination of volume or amplitude and frequency. For example, the sound pressure conversion unit 109a uses an FFT (fast Fourier transform) to decompose the audio signal into frequency components, analyze the volume (amplitude) of each component, calculate a target sound pressure based on the amplitude for each frequency, and convert it into a reference sound pressure across the entire frequency band. This sound pressure conversion unit 109a performs calculations based on a decibel-to-Pascal conversion formula. In this case, the sound pressure control unit 109 sets a specific frequency and controls the sound pressure without excessively increasing the volume, thereby applying sound parameters that affect the object of observation in the low frequency range of 20 Hz or less.

[0073] In particular, in this embodiment, the output frequency band is experimentally set to a range inaudible to humans (below 20 Hz) and a frequency band that is highly efficient and not dangerous. Furthermore, it is necessary to ensure a predetermined sound pressure effect at the minimum necessary volume. The sound pressure conversion unit 109a applies sound parameters that affect the observation target in the low frequency range below 20 Hz by setting a specific frequency and controlling the sound pressure without excessively increasing the volume. Note that the sound pressure conversion unit 109a can also convert the amplitude and frequency into a sound pressure reproduced by a combination of these by arbitrarily setting the amplitude and frequency. It is also possible to calculate the sound pressure from a specific amplitude and frequency, adjust the calculated sound pressure, and recursively calculate the amplitude and frequency required to reproduce the adjusted sound pressure.

[0074] The frequency filter unit 109b is a module that uses an equalizer (EQ) function to emphasize (boost) or suppress (cut) specific frequency bands to reproduce the target sound pressure. Specifically, the frequency filter unit 109b filters the low frequency range (20 Hz to 200 Hz) and the high frequency range (2 kHz to 20 kHz) intensively to reproduce the target sound pressure at specific frequencies, and also limits the output sound to a required band while suppressing the overall volume by limiting it to a range that is inaudible to humans, cutting or reducing the audible high frequency range, etc.

[0075] The specific frequency amplification unit 109d is a module that amplifies a specific band (e.g., low or mid-range) from among the frequency components set by the sound pressure conversion unit 109a, thereby achieving the required sound pressure. It can adjust the amount of amplification required at the target frequency from the output signal of the frequency filter unit 109b, and compensate for any output amount that is insufficient in sound pressure, thereby increasing the sound density in line with the target while suppressing the sound pressure.

[0076] The anti-phase sound generation unit 109c is a module that uses an impulse response to generate anti-phase sound waves that cancel out unnecessary reflected sound and interference.It generates anti-phase sound waves in a specific frequency band, suppresses unnecessary fluctuations in sound pressure due to environmental sounds, etc., and achieves clear sound quality and accurate target sound pressure.

[0077] The compressor unit 109e is a module that suppresses sound peaks (maximum volume) and adjusts RMS (average sound pressure). For example, if the volume exceeds a set threshold, it compresses the peak volume and controls the volume in conjunction with a limiter, maintaining a uniform sound pressure while reaching a predetermined sound pressure.

[0078] The analysis data sending unit 108 is a module that sends the results of the analysis performed by the analysis unit 106 to the simulation server 3, and includes devices that send data via wireless or wired transmission or media such as a USB memory.

[0079] The animal behavior analysis device 10 also includes a video image recognition unit 104, an observation execution unit 103, an analysis unit 106, and a database 107 as modules of the analysis system.

[0080] The video recognition unit 104 is a module that recognizes video captured by the camera 11 and extracts and traces information such as the subject's movements. Specifically, the video recognition unit 104 breaks down the video into frames, extracts features from each frame, and extracts information about various elements in the image, such as edges, colors, textures, object shapes, and faces, as well as temporal changes and object movements, to identify the presence of objects or entities in specific frames. Based on the results of this identification, the unit tracks the object's movement between consecutive frames, performs action recognition to determine how the object is behaving in the video, analyzes the context and scenes of the entire video, distinguishes between background and foreground, and classifies the video or its parts into specific categories based on the extracted features and information.

[0081] The observation execution unit 103 is a module that oversees the operations of the shooting control unit 101, sensor control unit 102, and audio parameter control unit 105, and synchronizes the operations of the camera 11 and various sensors 21 with the audio output from the speaker 13, and inputs the images and signals input from the camera 11 and sensors 21 into the analysis unit 106 as observation data that associates the audio output with the images and signals input from the camera 11 and sensors 21.

[0082] The observation execution unit 103 includes a feedback control unit 103a, which, if the results of a previously performed molecular dynamics simulation are available, controls various settings to reflect the mechanism by which a unique phenomenon was observed in the molecular dynamics simulation and the audio parameters at that time in the next observation of animal behavior. For example, by feeding back the audio parameters used in the molecular dynamics simulation, it is possible to analyze neural activity and behavior, such as analyzing the dynamic behavior of chemical substances and ion channels such as neurotransmitters between neurons, or to elucidate the action and mechanism of drugs. Examples of elucidating the action and mechanism of drugs include evaluating the effects of drugs in animal behavior analysis, analyzing the binding and action of medicinal ingredients to proteins and receptors, and evaluating the mechanism of behavioral changes in animals and their effects on physiological functions.

[0083] The analysis unit 106 is a module that analyzes the correlation between the sound parameters of the output sound and the behavior of the animal. Specifically, when sound is output with predetermined parameter settings, the analysis unit 106 extracts animal movements such as turning toward the sound source, pricking up ears, or tilting its head in a specific direction, escape and exploration behaviors such as showing interest in the sound and moving toward or away from the sound source or searching for the sound source, signs of surprise or stress such as dilated pupils, bristling dorsal hair, and trembling, and reactions such as vocalizations, growling, or barking by the animal in response to the sound. The analysis unit 106 also analyzes territorial behaviors such as whether an animal interrupts eating or resting due to the sound and then resumes, whether herd animals change their collective behavior or the interactions between individuals, and whether behaviors to protect boundaries with other animals or the same species are reinforced or reduced by the sound.

[0084] Furthermore, in this embodiment, in observing this animal behavior, work is carried out to periodically collect intestinal flora samples of the animal, such as collecting fecal samples, and DNA is extracted from the collected samples, and in correlation analysis in the analysis unit 106, microbiome analysis is performed using a next-generation sequencer, as well as intestinal flora data analysis. In this intestinal flora data analysis, microbiome data, which is the analysis result of the analyzed microbiome, is stored in database 107, and species identification, measurement of abundance, analysis of diversity, etc. are performed to analyze the correlation between the animal's behavior data, vocal parameters, structural and functional changes in proteins, and the composition of the intestinal flora.

[0085] The analysis unit 106 analyzes the correlation between the behavior of the target animal and structural variations or functional morphological changes in specific cells, receptors, or proteins in relation to the results of microbiome analysis of DNA extracted from the intestinal flora sample of the target animal. Specifically, the analysis unit 106 refers to a neural network trained by stacking distribution patterns of features extracted from the correlation between the behavior of the animal and structural variations or functional morphological changes in specific cells, receptors, or proteins in relation to the results of microbiome analysis of DNA extracted from the intestinal flora sample of the target animal, and refers to the neural network based on the observed animal behavior and the results of microbiome analysis of DNA extracted from the intestinal flora sample of the animal, identifies proteins and receptors that are highly correlated with the animal behavior, and includes them in the analysis data as correlation data.

[0086] The molecular dynamics simulation engine 34 of the simulation server 3 then executes an expanded molecular dynamics simulation with the addition of intestinal flora processing. Here, modeling is performed that takes into account the impact of changes in intestinal flora on the structure and function of proteins and receptors, and an expanded simulation is executed that includes intestinal flora data, and the results are analyzed. Next, the animal's behavior, vocal parameters, structural and functional changes in proteins, and intestinal flora data are integrated to analyze the combined effects, and the integrated analysis results are displayed in an easy-to-understand format (e.g., graph, chart, heat map). When adding this intestinal flora processing, the intestinal flora data may be added to the neural network's training data to learn more complex relationships, and the neural network model may be retuned based on the added data to enable more accurate predictions.

[0087] The database 107 is a storage device that stores the observation records performed by the observation execution unit 103 and also stores the results of the analysis performed by the analysis unit 106 in association with the observation records. The AI ​​neural network described above may be stored in this database 107 and referenced by this database 107.

[0088] (2) Simulation Server 3 The simulation server 3 is equipped with a molecular dynamics simulation engine 34 and is a simulation means that acquires the analysis results from the analysis unit 106 of the animal behavior analysis device 10 and executes a molecular dynamics simulation for a predetermined protein using the characteristics of the voice output to the animal. Specifically, as shown in FIG. 4, the simulation server 3 is equipped with a communication I / F 31, an animal behavior acquisition unit 32, the molecular dynamics simulation engine 34, a result transmission unit 26, and various databases 35a-c.

[0089] The communication I / F 31 is a module that communicates with other devices and servers via a communication network, and the animal behavior acquisition unit 32 is a module that acquires, from the animal behavior analysis device 10, the observation records and the correlation between the sound parameters analyzed by the analysis unit 106 related to these observation records and the animal behavior, via the communication I / F 31. The animal behavior acquisition unit 32 is equipped with a correlation extraction unit 33, which is a module that extracts the correlation between the sound parameters and the behavior of the animal that has been played that sound, and inputs this to the molecular dynamics simulation engine 34.

[0090] The correlation extraction unit 33 uses, for example, AI (artificial intelligence) to predict the correlation between behavior when hearing sound and structural or functional morphological changes in specific cells, receptors, or proteins, extracts the correlation between sound parameters and the behavior of animals that have been exposed to the sound, and performs a molecular dynamics simulation in which the sound (vibration frequency, vibration energy, etc.) acts on a protein structure that has been narrowed down to a certain extent.

[0091] The molecular dynamics simulation engine 34 is software that references a neural network for animal behavior information correlated with a specific voice parameter based on the correlation between the voice parameter and animal behavior analyzed by the analysis unit 106, and executes a molecular dynamics simulation for the extracted cells, receptors, or proteins. Note that this neural network may be an AI service provided by a server located on a network accessible via an API (Application Programming Interface), or may be stored in the molecular dynamics database 35b.

[0092] Specifically, the molecular dynamics simulation engine 34 according to this embodiment (1) models the structure of a specified protein that is the subject of molecular dynamics simulation, and acquires the audio parameters of the sound output to the animal; (2) calculates the periodic external forces acting on the atoms of the specified protein and the water molecules around them based on the audio parameters for the modeling; and (3) adds the periodic external forces to the interactions between atoms of the specified protein and the surrounding environment, and integrates the changes over time in the relative positions and velocities between the atoms.

[0093] In particular, this embodiment is a molecular dynamics simulation in which ultrasonic vibrations are applied to a protein, and some special elements are added to a normal simulation method. The specific configuration and operation of such a simulation are described below.

[0094] First, to set the initial conditions, the protein and environmental conditions are modeled. Then, to model the ultrasonic vibrations, vibration parameters are set and an external force is applied. Specifically, the frequency, amplitude, and direction of the ultrasonic waves are defined to set the vibration parameters, and the temporal changes in the force applied to the protein are reproduced. Furthermore, during the simulation, external forces are applied, such as applying periodic external forces to protein atoms and nearby water molecules, based on the defined parameters.

[0095] We also calculate forces and integrate equations of motion. This involves calculating the effects of external forces due to ultrasound, as well as the interactions between the protein and the environment atoms. We also calculate the time evolution of the atomic positions and velocities, updating them sequentially in small time steps. We then calculate various physical quantities (e.g., structural changes, energy changes) to analyze how the vibrations affect the protein's structure, dynamics, and function. We also perform detailed analysis of the simulation results to understand the impact of ultrasonic vibrations on the protein's function and stability. Finally, we iterate and improve the simulation, including these processes.

[0096] The molecular dynamics simulation engine 34 according to this embodiment also performs an integrated simulation of ultrasound and psychological and physical states during mechanism analysis in the mechanism analysis unit 34b. This integrated simulation is an object-based psychological and physical integrated simulation that correlates neuronal activity and physiological parameters of the body through ultrasound and analyzes a set of psychological and physical states, and performs behavioral analysis not only on animals but also on primates such as humans, and even on simple organisms such as insects.

[0097] Specifically, this integrated simulation integrates the effects of ultrasound as a "mental" and "physical" reaction via a psychology-physical interaction model into a simulation that changes over time. In particular, neuronal activity (psychological aspects) is linked to molecular dynamics simulation (physical aspects), and the simulation results are dynamically changed using state sets (blood pressure, heart rate, etc.) as indicators. In this case, by utilizing state sets in object-based simulation, an "integrated psychology-physical simulation" is realized that explores causal relationships for each target, and the correlation between psychology and the body is analyzed with the influence of ultrasound as the main axis.

[0098] The mechanism analysis unit 34b also analyzes sound wave perception in the inaudible range. Specifically, for example, it is equipped with sound parameters for applying sound waves to cells other than the ear, and is configured to enable the cells to sense sound waves via a piezoelectric protein (e.g., Piezo1) and induce calcium ions. Because sound waves beyond the audible range may also affect cells, proteins such as Piezo1 are used to analyze how cells sense and respond to sound waves. Simulations of molecular dynamics using sound stimulation emphasize this innovative approach, different from conventional drug experiments, and elucidate the mechanism by which sound action on Piezo1 induces a response in the body via calcium ions.

[0099] The molecular dynamics simulation engine 34 according to this embodiment also includes a voice parameter control unit 34 a and a mechanism analysis unit 34 b. The voice parameter control unit 34 a is a module that controls the characteristics of voice (acoustics) that act on a protein structure in the simulation executed by the molecular dynamics simulation engine 34. Parameters for controlling these voice characteristics include frequency, voice "pitch," formants (resonance peaks), temporal features (voice duration and fixed phoneme intervals), spectral envelope (shape and intensity of frequency components in a spectrogram), energy (intensity and amplitude of a voice signal), zero-crossing rate, MFCC (Mel Frequency Cepstral Coefficients), spectral features, delta features (temporal changes in features), voice quality, acoustic spectrum (energy distribution in each frequency band of voice), LTAS (Long-Term Average Spectrum), and the like.

[0100] In particular, in this embodiment, the voice parameter control unit 34a reproduces a specific "sound pressure" as a combination of volume (amplitude) and frequency as part of a "specific voice parameter" in the simulation performed by the molecular dynamics simulation engine 34. Here, the frequency band used experimentally is a range inaudible to humans (below 20 Hz) or a frequency band that is highly efficient and not dangerous. Furthermore, during the simulation, the voice parameter control unit 34a performs calculations based on a decibel-to-Pascal conversion formula, including measures to ensure the sound pressure effect without excessively increasing the volume. At this time, the voice parameter control unit 34a sets a specific frequency and applies voice parameters that affect the observation subject in the low-frequency range below 20 Hz by controlling the sound pressure without excessively increasing the volume.

[0101] In this embodiment, the voice parameter control unit 34a includes a simulation condition reference unit 37a, a sound pressure setting unit 37b, an environment setting unit 37c, an amplitude / frequency control unit 37d, a parameter output unit 37e, and a sound pressure control unit 38, as shown in FIG. 5.

[0102] The simulation condition reference unit 37a is a module that acquires the target sound pressure and the volume (amplitude) or frequency set to reproduce that sound pressure from a database or arbitrarily input setting values ​​as simulation conditions, and has a sound pressure control unit 38. The simulation condition reference unit 37a also functions as control means that sequentially calculates and automatically sets or presents other parameters required to reproduce the target sound pressure using the calculation results by the sound pressure conversion unit 38a of the sound pressure control unit 38 in conjunction with setting inputs such as volume or frequency.

[0103] The environment setting unit 37c is a module that acquires and sets information about the environment in which the audio is output. This environmental information includes any input value by the operator, as well as the current atmospheric pressure, temperature, and humidity detected by various sensors. In particular, the atmospheric pressure is set to 1 atmosphere.

[0104] The sound pressure setting unit 37b is a module that sets the sound pressure of the sound to be output in the experiment, and 1 atmosphere is set as the current atmospheric pressure condition (environmental condition) included in the environmental information acquired by the environment setting unit 37c, and the sound pressure setting unit 37b sets a sound pressure that fluctuates by 1 / 100,000 of this 1 atmosphere. This atmospheric pressure condition is usually 1 atmosphere, and as shown in Figure 10(a) , the sound that needs to be output in the experiment is included in an effective sound pressure group, which is set in simulation to fall within a range of fluctuation of 1 / 100,000 of 1 atmosphere, and an arbitrary sound pressure is selected from this effective sound pressure group and set as the target sound pressure.

[0105] The amplitude / frequency control unit 37d is a module that sets the range of amplitude and frequency when reproducing the specific target sound pressure set by the sound pressure setting unit 37b by combining volume or amplitude and frequency. For example, when determining the amplitude (volume) and frequency for reproducing the sound pressure required in an experiment, the module provides a user interface for setting the upper (or lower) limits of the volume and frequency that can be output in order to suppress the output volume. In this case, either the amplitude or frequency may be selected, and the reproducible sound pressure may be calculated from the selected amplitude or frequency.

[0106] The parameter output unit 37e is a module that applies a specified volume and frequency in order to actually reproduce a predetermined sound pressure in a molecular dynamics simulation. In particular, in this embodiment, as shown in FIG. 10(a), a fluctuation of 1 atmosphere + 1 / 100,000 can be realized, enabling fine volume and frequency control in a molecular dynamics simulation.

[0107] The voice parameter control unit 34a according to this embodiment also includes a sound pressure control unit 38 that sets a predetermined target sound pressure in the molecular dynamics simulation and controls the volume (amplitude) and frequency to achieve that sound pressure. This sound pressure control unit 38 is a module that outputs sound to achieve the desired sound pressure according to the set sound pressure, volume, and frequency. In this embodiment, the frequency band output for experimental purposes can be set to a range that humans cannot hear (below 20 Hz) or a frequency band that is highly efficient and safe. To achieve this, the voice parameter control unit 34a amplifies the low-frequency range to ensure a certain sound pressure, and filters and compresses the audible high-frequency range to ensure a sound pressure effect without excessively increasing the volume, as shown in FIG. 10(b).

[0108] Specifically, the sound pressure control unit 38 includes a sound pressure conversion unit 38a, which virtually simulates the functions of a frequency filter, specific frequency amplification, inverse phase sound generation, and compressor. By virtually simulating these functions, the sound pressure conversion unit 38a converts the sound pressure into a volume and frequency that achieves a predetermined target sound pressure. Specifically, the sound pressure conversion unit 38a converts the desired sound pressure into an amplitude and frequency to be output by performing calculations based on a decibel-to-Pascal conversion formula to reproduce the target sound pressure as a specific audio parameter using a combination of volume or amplitude and frequency. For example, the sound pressure conversion unit 38a uses an FFT (fast Fourier transform) to decompose the audio signal into frequency components, analyze the volume (amplitude) of each component, calculate the target sound pressure based on the amplitude for each frequency, and convert it into a reference sound pressure across the entire frequency range. The sound pressure conversion unit 38a performs calculations based on a decibel-to-Pascal conversion formula. In this case, the sound pressure control unit 38 sets a specific frequency and controls the sound pressure without excessively increasing the volume, thereby applying sound parameters that affect the object of observation in the low frequency range of 20 Hz or less.

[0109] In particular, in this embodiment, the frequency band used in the molecular dynamics simulation is experimentally set to a range inaudible to humans (below 20 Hz) and a frequency band that is highly efficient and not dangerous, and furthermore, it is necessary to ensure a predetermined sound pressure effect at the minimum required volume. For this reason, the sound pressure conversion unit 38a uses functions of a frequency filter, specific frequency amplification, inverse phase sound generation, and compressor to set a specific frequency and control the sound pressure without excessively increasing the volume, thereby applying sound parameters that affect the object of observation in the low frequency range below 20 Hz.

[0110] The sound pressure conversion unit 38a can also convert any amplitude and frequency into a sound pressure that is reproduced by combining these values, and can also calculate sound pressure from a specific amplitude and frequency, adjust the calculated sound pressure, and recursively calculate the amplitude and frequency required to reproduce the adjusted sound pressure.

[0111] Specifically, the frequency filter function uses the equalizer (EQ) function to emphasize (boost) or suppress (cut) specific frequency bands to reproduce the target sound pressure. For example, by filtering the low frequency range (20Hz-200Hz) or high frequency range (2kHz-20kHz) intensively, the target sound pressure can be reproduced at specific frequencies, and the output sound can be limited to the required band while suppressing the overall volume by limiting it to the inaudible range for humans, cutting or reducing the audible high frequency range, etc.

[0112] The specific frequency amplification function is an emphasis (amplification) process known as an amplifier, which amplifies specific bands (e.g., low and midrange frequencies) among the frequency components calculated by the sound pressure conversion unit 38a to achieve the required sound pressure. For example, the required amplification amount at a target frequency of the output signal from the frequency filter can be adjusted to compensate for insufficient sound pressure, thereby increasing the sound density in line with the target while suppressing sound pressure. The compressor function suppresses sound peaks (maximum volume) and adjusts RMS (average sound pressure). For example, if the volume exceeds a set threshold, the peak volume is compressed and a limiter is used to control the volume, maintaining a uniform sound pressure while reaching the desired sound pressure.

[0113] The inverse phase sound generation function uses impulse responses to generate inverse phase sound waves that cancel out unwanted reflected sounds and interference. For example, in a specific frequency band, it generates inverse phase sound waves to suppress unnecessary fluctuations in sound pressure caused by environmental sounds, etc., achieving clear sound quality and accurate target sound pressure.

[0114] Furthermore, in the simulations that the molecular dynamics simulation engine 34 according to this embodiment performs, as part of the sensing of gravity and sound pressure and cell stimulation using sound pressure, sound waves are applied to a "piezo-type protein (e.g., Piezo1)" to reproduce the way in which minute vibrations caused by sound waves act on cells. At this time, the sound parameter control unit 34a reproduces the sound pressure by a combination of volume (amplitude) and frequency when simulating structural or functional changes in animal or human cells or proteins using "sound pressure" as a specific sound parameter.

[0115] Furthermore, the molecular dynamics simulation engine 34 introduces minute fluctuations under one atmospheric pressure condition. More specifically, in the molecular dynamics simulation, by applying a 1 / 100,000 air fluctuation (pressure fluctuation) to the cells or proteins being observed under an atmospheric pressure environment of one atmospheric pressure, the structural or functional response of the target cells or proteins is reproduced. This is also intended to reflect minute changes in the actual environment, and the effect of applying sound pressure can be verified.

[0116] The mechanism analysis unit 34b is a module that refers to AI and a database and analyzes mechanisms due to structural variations and functional morphological changes of cells, receptors, or proteins obtained as a result of execution by the molecular dynamics simulation engine 34. Specifically, the molecular dynamics simulation engine 34 performs a molecular dynamics simulation in which sound (vibration frequency, vibration energy, etc.) acts on the protein structure narrowed down to a certain extent by the correlation extraction unit 33, and the mechanism analysis unit 34b analyzes mechanisms due to structural variations and functional morphological changes of cells, receptors, or proteins obtained as a result of the simulation with reference to AI.

[0117] This AI is equipped with a neural network trained by layering distribution patterns of features extracted from the correlation between animal behavior and structural or functional morphological changes of specific cells, receptors, or proteins. The mechanism analysis unit 34b references this neural network for animal behavior information correlated with specific sound parameters, and predicts and analyzes mechanisms such as ligand binding, binding free energy, and allosteric effects due to structural or functional morphological changes obtained as a result of performing a molecular dynamics simulation on the extracted cells, receptors, or proteins.

[0118] The AI ​​of the mechanism analysis unit 34b is also linked to a neural network trained by layering distribution patterns of features extracted from the correlation between the animal's behavior and structural or functional morphological changes of specific cells, receptors, or proteins in association with the microbiome analysis results of DNA extracted from the target animal's intestinal flora sample.The mechanism analysis unit 34b then refers to the neural network trained in association with the microbiome analysis results to analyze the correlation between the animal's behavior data, the voice parameters, the structural and functional changes of the proteins, and the composition of the intestinal flora.

[0119] Then, the molecular dynamics simulation engine 34 of the simulation server 3 executes an expanded molecular dynamics simulation with the addition of processing related to the intestinal flora. Here, modeling is performed taking into account the effects of changes in the intestinal flora on the structure and function of proteins and receptors, an expanded simulation including data on the intestinal flora is executed, and the results are analyzed.

[0120] Next, the data on animal behavior, vocal parameters, structural and functional changes in proteins, and intestinal flora are integrated to analyze the combined effects, and the integrated analysis results are displayed in an easy-to-understand format (e.g., graph, chart, heat map). Note that when adding this intestinal flora processing, the intestinal flora data may be added to the neural network's training data to learn more complex relationships, and the neural network model may be readjusted based on the added data to make more accurate predictions.

[0121] The result transmission unit 26 is a communication module that sends the results of the simulation by the molecular dynamics simulation engine 34 and the mechanism analysis results obtained as a result thereof to the animal behavior analysis device 10 and the new drug development server 4. The information sent from the result transmission unit 26 is used as feedback to the animal behavior analysis device 10 and for new drug development in the new drug development server 4.

[0122] The various databases 35a-c include an animal behavior information database 35a, a molecular dynamics database 35b, and a simulation database 35c. These databases may be a single database device, or may be a group of devices in which multiple databases are linked by relationships.

[0123] The animal behavior information database 35a records animal information such as the type, age, sex, and weight of the animals that were the subject of the experiment, experimental environment information such as the frequency, intensity, and duration of the sound, the environmental temperature, and humidity, behavioral information such as the type, frequency, duration, and fluctuations of behavior before and after the sound stimulus, and physiological responses such as heart rate, brain wave fluctuations, and the amount of stress hormone secretion.

[0124] The molecular dynamics database 35b stores molecular information such as structural information of target proteins and neurons (amino acid sequence, 3D structure, etc.), acoustic parameters, simulation conditions such as the initial structure of the protein, temperature, pressure, pH, and ion concentration, structural fluctuations such as structural fluctuations, folding, and binding information of proteins due to acoustic stimuli, as well as dynamic information such as energy fluctuations, binding energy, and patterns of dynamic fluctuations.

[0125] The simulation database 35c includes simulation results such as structural fluctuations, energy fluctuations, dynamic fluctuations, and other changes over time obtained through simulation, correlation information such as the correlation between animal behavior and molecular dynamic fluctuations, and which structural fluctuations are associated with which behavioral fluctuations, as well as information on analytical tools such as tools and software for analyzing the simulation results and information on the algorithms and methods used.

[0126] (3) New Drug Development Server 4 The new drug development server 4 is a server device that supports the development of new drugs and gene therapies using the simulation results from the simulation server 3. It supports the development of new drugs and gene therapies by, for example, referring to AI trained for drug discovery and selecting drug candidate molecules based on animal behavior information that correlates with specific voice parameters.

[0127] This AI trained for drug discovery is equipped with a drug discovery neural network trained by layering distribution patterns of features extracted from the structural or functional morphological changes of specific cells, receptors, or proteins, and the correlation between proteins and drug candidate molecules, and references the drug discovery neural network to animal behavior information correlated with specific sound parameters to select drug candidate molecules, analyzes how the drug candidate molecules act on cells or organs to which specific sounds have been applied, and serves as an opportunity for the development of new drugs and gene therapies. Note that this drug discovery neural network may be an AI service provided by a server located on a network accessible via an API, or may be stored in drug discovery database 45b.

[0128] Specifically, as shown in Figure 4, the new drug development server 4 includes a communication I / F 41, a simulation result acquisition unit 42, a new drug development and design unit 44, a development information transmission unit 46, and various databases 45a-c. The communication I / F 41 is a module that communicates with other devices and servers via a communication network, and the simulation result acquisition unit 42 is a module that acquires simulation results and mechanism analysis results from the simulation server 3 via the communication I / F 41. The simulation result acquisition unit 42 includes a drug discovery reference unit 42a for referencing drug discovery information. The drug discovery reference unit 42a extracts useful information based on the results obtained from the simulation and the mechanism analysis results, using, for example, AI (artificial intelligence), and inputs the information to the new drug development and design unit 44.

[0129] The new drug development and design unit 44 is a module that supports development policies and designs in the development of new drugs and gene therapies, and includes a clinical information collection unit 44a, an evaluation unit 44b, and a drug candidate molecule analysis unit 44c.

[0130] The clinical information collection unit 44a is a module that collects information obtained at clinical sites, stores it in a clinical information database 45c, and organizes and analyzes it. Specifically, the clinical information collection unit 44a collects animal behavior data on changes in the reactions and behavior of animals when they are exposed to specific sounds, collects biomarker data from the blood and urine of animals, and records data on functional and structural changes in nerve cells and related proteins, as well as the type, frequency, amplitude, etc. of the sounds used.

[0131] The evaluation unit 44b is a module that uses AI to evaluate the effectiveness and policy of new drugs and gene therapies based on collected clinical information and simulation results. Specifically, the evaluation unit 44b performs correlation analysis to evaluate the correlation between animal behavior and nervous system fluctuations, predicts the potential effectiveness of new drugs and gene therapies based on simulation results, evaluates potential side effects and risks based on simulation results, and proposes the most effective treatment methods and drug dosages.

[0132] The drug candidate molecule analysis unit 44c is a module that plays a role in predicting and analyzing molecules that will be candidate for new drugs by referring to AI, and in this embodiment, drug candidate molecules are selected by referring to a drug discovery neural network for animal behavior information that is correlated with specific sound parameters. Specifically, the drug candidate molecule analysis unit 44c can select proteins or molecules associated with nervous system fluctuations as targets, predict molecules that may bind to the target molecule, calculate binding affinity (binding strength between the candidate molecule and the target) using simulation, analyze drug pharmacokinetics (absorption, distribution, metabolism, excretion) using simulation, and propose new drug design guidelines based on the properties of the most promising drug candidate molecules.

[0133] The development information transmission unit 46 is a communication module that sends new drug development plans and gene therapy plans designed by the new drug development and design unit 44 to the animal behavior analysis device 10 and other information processing terminals. The various databases 45a-c include a simulation result database 45a, a drug discovery database 45b, and a clinical information database 45c. These databases may be a single database device, or a group of devices in which multiple databases are linked by relationships.

[0134] The simulation result database 45a is a storage device for saving the results of molecular dynamics simulations, and records structural information about proteins and cells, the initial and final structures of the simulation, structural fluctuations during the simulation, dynamic fluctuations such as energy fluctuations, interaction information such as details of interactions with proteins and molecules inside and outside the cells, simulation parameters such as the force field used, temperature, and pressure, and acoustic parameters such as the type of acoustics used, frequency, and amplitude.

[0135] The drug discovery database 45b is a storage device for storing information on new drug designs and drug candidates. Specifically, it stores and holds information on drug candidate molecules such as chemical structure, physicochemical properties, and biological activity; target information such as information on the target protein or cell; binding affinity such as the binding strength and binding mode between the drug candidate and the target; drug discovery policies and plans such as the methods and techniques used, the desired therapeutic effects, and potential side effects; and synthesis routes such as the synthesis procedures for drug candidates and the necessary reagents.

[0136] The clinical information database 45c is a storage device for storing clinical trial and patient information, and includes patient information such as age, sex, disease history, and genetic background, information on the drugs and treatments used, their effects and side effects, the effects of sound such as the patient's reactions and changes after acoustic stimulation, information on nervous system fluctuations such as details of structural and functional fluctuations in nervous system cells and proteins, and biomarker information such as the concentrations of specific markers in blood and urine. These databases are interrelated, and exchanging and integrating information enables more efficient and accurate decisions to be made in the development of new drugs and gene therapies.

[0137] (Animal Behavior Analysis Method) The animal behavior analysis method of the present invention can be implemented by operating the animal behavior analysis system described above. Figure 6 is a sequence diagram showing the operation of the animal behavior analysis system. Note that the processing procedures described below are merely examples, and each process may be modified as much as possible. Furthermore, steps in the processing procedures described below may be omitted, replaced, or added as appropriate depending on the embodiment.

[0138] 6 , first, in animal behavior analysis device 10, a specific sound with controlled audio parameters (S101) is played to a target animal, and observation such as filming and tracing the animal's behavior is started (S102). Specifically, observation execution unit 103 controls the operations of imaging control unit 101, sensor control unit 102, and audio parameter control unit 105, synchronizes the operations of camera 11 and various sensors 21 with the audio output from speaker 13, and inputs the images and signals input from camera 11 and sensors 21 and the audio output as associated observation data to analysis unit 106.

[0139] At this time, the voice parameter control unit 105 controls the parameters of the voice output from the speaker 13. Here, the voice parameter control unit 105 controls the frequency, voice "pitch," formants, temporal features, spectral envelope, energy (intensity and amplitude of the voice signal), zero-crossing rate, MFCC (Mel Frequency Cepstral Coefficients), spectral features, delta features, voice quality, acoustic spectrum, LTAS (Long-Term Average Spectrum), etc. To describe in detail the processing by this voice parameter control unit 105, in the above step S101, the processing shown in FIG. 7 is executed.

[0140] First, a target sound pressure is set (S301), and then the volume (amplitude) and frequency for reproducing that sound pressure are adjusted (S302). Specifically, the interface control unit 105a manually or automatically inputs the target sound pressure and the volume (amplitude) or frequency set for reproducing that sound pressure. At this time, the environment setting unit 105b acquires and inputs information about the environment in which the sound will be output. Based on the input environment, the sound pressure setting unit 105c sets the sound pressure of the sound to be output in the experiment, and the amplitude / frequency control unit 105d sets the range of amplitude and frequency so that the specific target sound pressure set by the sound pressure setting unit 105c is reproduced by a combination of volume or amplitude and frequency. In this case, an upper limit (or lower limit) of the volume or frequency that can be output can be set to limit the output volume. Alternatively, either the amplitude or frequency can be selected, and the reproducible sound pressure can be calculated from the selected amplitude or frequency.

[0141] In adjusting the volume and frequency, filtering of specific frequencies (S303), compression and amplification of specific frequencies (S304), and generation of inverse phase sound (S305) are performed as necessary. Specifically, the equalizer (EQ) function of the frequency filter unit 109b is used to emphasize (boost) or suppress (cut) specific frequency bands, and filtering is focused on the low frequency range (20 Hz to 200 Hz) and the high frequency range (2 kHz to 20 kHz), and the output sound is limited to a required band while suppressing the overall volume by limiting it to a range that is inaudible to humans, cutting or reducing the audible high frequency range, etc.

[0142] In accordance with this filtering, the specific frequency amplifier 109d adjusts the amplification amount required at the target frequency of the output signal from the frequency filter 109b to compensate for any insufficient sound pressure, thereby increasing the sound density in line with the target while suppressing sound pressure. The compressor 109e suppresses the sound peak (maximum volume) and adjusts the RMS (average sound pressure). Furthermore, the antiphase sound generator 109c uses the impulse response to generate antiphase sound waves that cancel out unwanted reflected sounds and interference, thereby suppressing unnecessary sound pressure fluctuations due to environmental sounds and the like in a specific frequency band, thereby achieving clear sound quality and accurate target sound pressure. The results of these processes are then integrated and the sound pressure converter 109a sequentially converts parameters that reproduce the target sound pressure (S306), and the calculated results are automatically set or presented as audio parameters (S307).

[0143] The processes of steps S301 to S307 can be repeated as needed ("N" in S308) for adjustment. After the amplitude and frequency adjustment is completed in step S308 ("Y" in S308), the output control unit 105e generates an audio (WAV) file to actually reproduce the specified sound pressure (S309), and outputs this WAV file as audio parameters (S310). In particular, in this embodiment, the audio parameters include detailed hardware volume and frequency control to achieve a fluctuation of 1 atmosphere + 1 / 100,000.

[0144] After controlling the audio parameters in this manner, as shown in FIG. 6 , image recognition is performed on the captured image (S103), and the correlation between the audio parameters of the output audio and the animal's behavior is analyzed (S104). Specifically, the analysis unit 106 analyzes the correlation between the audio parameters of the output audio and the animal's behavior. More specifically, when audio is output with predetermined parameter settings, the analysis unit 106 extracts animal movements, escape / exploration behavior, signs of surprise or stress such as pupil dilation, and reactions. The analysis unit 106 also analyzes the effects of audio on animals while they are eating or resting, changes in the collective behavior of animals in a group, changes in interactions between individuals, and territorial behavior.

[0145] In step S104, it is possible to refer to an AI (neural network) trained by layering distribution patterns of features extracted from the correlation between the behavior of the animal and structural or functional / morphological changes of specific cells, receptors, or proteins in relation to the microbiome analysis results of DNA extracted from the intestinal flora sample of the target animal. This correlation analysis, or in the mechanism analysis S206 described below, may analyze the correlation between the behavioral data of the animal, the voice parameters, the structural / functional changes of the proteins, and the composition of the intestinal flora by referring to the neural network trained in relation to the microbiome analysis results.

[0146] The behavioral analysis process of steps S101 to S104 is repeated while changing the voice parameters ("N" in S105), and when a certain amount of data has been recorded, the behavioral analysis process is terminated ("Y" in S105), and the accumulated behavioral information and analysis data are sent to the simulation server 3 (S106).

[0147] In the simulation server 3 that receives this behavioral information and analysis data, the animal behavior acquisition unit 32 acquires the observation record and the correlation between the audio parameters analyzed by the analysis unit 106 related to this observation record and the animal behavior from the animal behavior analysis device 10 via the communication I / F 31 (S201).

[0148] Next, the correlation extraction unit 33 extracts correlations between the sound parameters and the behavior of the animals that have been exposed to the sound, and inputs the correlations into the molecular dynamics simulation engine 34 (S202). At this time, the correlation extraction unit 33 uses, for example, AI (artificial intelligence) to predict the correlation between the behavior of the animals when they hear the sound and structural fluctuations or functional morphological changes of specific cells, receptors, or proteins, extracts the correlations between the sound parameters and the behavior of the animals that have been exposed to the sound, and executes a molecular dynamics simulation in which the sound (vibration frequency, vibration energy, etc.) acts on protein structures that have been narrowed down to a certain extent.

[0149] In executing this molecular dynamics simulation, the voice parameter control unit 34a controls the voice (acoustic) characteristics to be applied to the protein structure (S203). Parameters for controlling these voice characteristics include frequency, voice "pitch," formants (resonance peaks), temporal features (voice duration and fixed phoneme intervals), spectral envelope (shape and intensity of frequency components in a spectrogram), energy (intensity and amplitude of a voice signal), zero-crossing rate, MFCC (Mel Frequency Cepstral Coefficients), spectral features, delta features (temporal changes in features), voice quality, acoustic spectrum (energy distribution in each frequency band of voice), LTAS (Long-Term Average Spectrum), etc.

[0150] The processing by the voice parameter control unit 34a will be described in detail below. In step S204, the processing shown in Fig. 8 is executed. First, the simulation conditions are referenced (S401), and the environment that is the premise of the simulation is set (S402), and the target sound pressure is set (S403). Here, atmospheric pressure and fluctuation range are set as environmental parameters.

[0151] Specifically, the simulation condition reference unit 37a acquires the target sound pressure and the volume (amplitude) or frequency set to reproduce that sound pressure from a database or arbitrarily input settings as simulation conditions. According to the acquired simulation conditions, the environment setting unit 37c sets atmospheric pressure, temperature, and humidity as environmental information for audio output. In particular, 1 atmosphere is set as the atmospheric pressure condition, and a fluctuation range of the target sound pressure is set based on this atmospheric pressure condition. Under these environmental conditions, the sound pressure setting unit 37b sets a sound pressure that fluctuates by 1 / 100,000 of the current atmospheric pressure condition included in the environmental information acquired by the environment setting unit 37c. This atmospheric pressure condition is typically 1 atmosphere. As shown in FIG. 10( a), the sound that needs to be output in the experiment is included in an effective sound pressure group. This effective sound pressure group is set to fall within a fluctuation range of 1 / 100,000 of 1 atmosphere in the simulation. An arbitrary sound pressure is selected from this effective sound pressure group and set as the target sound pressure.

[0152] After the target sound pressure is set, the volume (amplitude) and frequency for reproducing that sound pressure are adjusted (S404). Specifically, the target sound pressure to be output in the simulation, set by the sound pressure setting unit 37b, is reproduced by a combination of volume or amplitude and frequency through the amplitude / frequency control unit 37d. In this case, in order to suppress the output volume, an upper (or lower) limit of the volume or frequency that can be output may be set, or either the amplitude or frequency may be selected, and the reproducible sound pressure may be calculated from the selected amplitude or frequency.

[0153] In adjusting the volume and frequency, filtering of specific frequencies (S405), compression and amplification of specific frequencies (S406), and generation of inverse phase sound (S407) are performed as necessary, and the results of these processes are integrated and sequentially converted into parameters that reproduce the target sound pressure by the sound pressure conversion unit 38a (S408), and the calculation results are output as audio parameters (S409).

[0154] After the sound parameters are determined by these processes, as shown in Figure 6, the molecular dynamics simulation engine 34 refers to the neural network for animal behavior information that is correlated with specific sound parameters based on the correlation between the sound parameters analyzed by the analysis unit 106 and animal behavior, and performs molecular dynamics simulation on the extracted cells, receptors, or proteins (S204).

[0155] Specifically, the molecular dynamics simulation according to this embodiment involves: (1) modeling the structure of a given protein that is the subject of the molecular dynamics simulation, and acquiring audio parameters of the sound output to the animal; (2) calculating periodic external forces acting on the atoms of the given protein and the water molecules around it based on the audio parameters for the modeling; and (3) adding the periodic external forces to the interactions between the atoms of the given protein and the surrounding environment, and integrating the temporal changes in the relative positions and velocities between the atoms. In particular, since this embodiment is a molecular dynamics simulation in which ultrasonic vibrations are applied to a protein, several special elements are added to a normal molecular dynamics simulation.

[0156] In addition, in the molecular dynamics simulation, the mechanism analysis unit 34b refers to AI and a database (S205) and analyzes the mechanisms resulting from structural variations and functional morphological changes of cells, receptors, or proteins obtained as a result of execution by the molecular dynamics simulation engine 34 (S206).

[0157] Specifically, the molecular dynamics simulation engine 34 performs a molecular dynamics simulation in which sound (vibration frequency, vibration energy, etc.) is applied to the protein structure narrowed down to a certain extent by the correlation extraction unit 33, and the mechanism analysis unit 34b analyzes the mechanisms due to structural variations and functional morphological changes of cells, receptors, or proteins obtained as a result of the simulation with reference to AI. The mechanism analysis unit 34b refers to a neural network for animal behavior information correlated with specific sound parameters, and predicts and analyzes mechanisms such as ligand binding, binding free energy, and allosteric effects due to structural variations or functional morphological changes obtained as a result of performing a molecular dynamics simulation on the extracted cells, receptors, or proteins.

[0158] The simulation-related processes of steps S201 to S206 are repeated while changing the sound parameters ("N" in S207), thereby iterating and improving the simulation. For example, if necessary, the ultrasound parameters are adjusted to run a simulation under different conditions, and the validity of the simulation is verified by comparing it with experimental data or other theoretical models. The simulation process is then terminated when a certain amount of data has been recorded ("Y" in S207), and the results of the simulation by the molecular dynamics simulation engine 34 and the resulting mechanism analysis results are sent from the result transmission unit 26 to the animal behavior analysis device 10 and the new drug development server 4 (S208). The information transmitted from the result transmission unit 26 is used for new drug development in the new drug development server 4 or for feedback control of the animal behavior analysis device 10 (S209).

[0159] The new drug development server 4 uses AI to predict the correlation between the structural or functional morphological changes of specific cells, receptors, or proteins and the proteins and drug candidate molecules, and selects a narrowed-down list of proteins and drug candidate molecules. This AI is equipped with a drug discovery neural network trained by layering distribution patterns of features extracted from the correlation between the structural or functional morphological changes of specific cells, receptors, or proteins and the proteins and drug candidate molecules, and selects drug candidate molecules by referring to the drug discovery neural network for animal behavior information correlated with specific sound parameters, and analyzes how the drug candidate molecules act on cells or organs to which specific sounds have been applied, providing a foothold for the development of new drugs and gene therapies.

[0160] Feedback control, for example, allows analysis of neural activity and behavior, such as the dynamic behavior of chemical substances and ion channels, such as neurotransmitters, between neurons, and the elucidation of the action and mechanism of drugs, by reflecting the sound parameters used in the molecular dynamics simulation in the animal behavior analysis device 10. Examples of elucidation of the action and mechanism of drugs include evaluating the effects of drugs in animal behavior analysis, analyzing the binding and action of medicinal ingredients on proteins and receptors, and evaluating the mechanism of behavioral changes in animals and their effects on physiological functions.

[0161] (Actions and Effects) This embodiment clarifies the relationship between the effects of ultrasound on animal behavior and the activity of cells and the nervous system, and uses this relationship to microscopically analyze the dynamic behavior of chemical substances and ion channels in response to ultrasound on proteins and receptors in molecular dynamics simulations. For example, a specific frequency that affects animal behavior can be applied to target cells or nervous system proteins or receptors, and molecular dynamics simulations can be used to track their dynamic structural changes and movements, examining functional morphological changes in proteins, and evaluating how ligands bind, how strong the binding is, and what selectivity they possess. As a result, this invention can reduce the high cost and computational resource constraints of molecular dynamics simulations, for example, in the fields of drug discovery and gene therapy.

[0162] In particular, in this embodiment, by setting the target sound pressure as a combination of volume (amplitude) and frequency using the sound parameter control unit 105 in animal experiments and the sound parameter control unit 34a in the simulation in molecular dynamics simulations, it is possible to finely adjust the sound environment that is most effective for animal behavior and efficiently elicit a specific behavioral response. Furthermore, by individually optimizing the volume and frequency, such as by filtering audible sounds or amplifying low-frequency sounds, it is possible to obtain highly accurate data in behavioral analysis and observation while reducing the burden on subjects such as animals.

[0163] Furthermore, in this embodiment, by reflecting the minute pressure fluctuations that proteins experience at 1 atmosphere in the molecular dynamics simulation environment, it is possible to more accurately simulate a normal physiological environment and reproduce the effects that fluctuations of 1 in 100,000 have on structural changes in cells and proteins, as well as on their functional changes. Furthermore, by generating acoustic pressure within the range of minute fluctuations, it is possible to observe in detail how the dynamic behavior of biomolecules changes in a natural environment, thereby realizing more realistic biological simulations and improving the reliability of data regarding the stability and dynamics of macromolecules such as proteins and cells.

[0164] Furthermore, this embodiment can analyze the changes caused by sound parameters on the psychological activity of neurons and physiological parameters using an integrated psychology-physiology simulation, making it possible to understand the psychological responses of animals to sound stimuli, the accompanying physical changes, and the correlation between them. For example, it is possible to quantitatively analyze how behavioral changes caused by specific frequencies or amplitudes affect physiological parameters such as cranial nerve activity, heart rate, and respiration. As a result, this embodiment provides a comprehensive understanding of the relationship between animal psychology and physical responses, and is expected to contribute to elucidating the mechanisms of animal behavior in response to sound stimuli and to applied research.

[0165] The present invention is not limited to the above-described embodiments, and the components can be modified and embodied in practice without departing from the spirit of the invention. Furthermore, various inventions can be created by appropriately combining multiple components disclosed in the above-described embodiments. For example, some components may be omitted from all the components shown in the embodiments.

[0166] 3...Simulation server 4...New drug development server 10...Animal behavior analysis device 11...Camera 12...Receiver 13...Speaker 14...Display 15...Keyboard 21...Sensor 26...Result transmission unit 32...Animal behavior acquisition unit 33...Correlation extraction unit 34...Molecular dynamics simulation engine 34a...Sound parameter control unit 34b...Mechanism analysis unit 35a...Animal behavior information database 35b...Molecular dynamics database 35c...Simulation database 37a...Simulation condition reference unit 37b...Sound pressure setting unit 37c...Environment setting unit 37d...Amplitude / frequency control unit 37e...Parameter output unit 38...Sound pressure control unit 38a...Sound pressure conversion unit 42...Simulation result acquisition unit 42a...Drug discovery reference unit 44...New drug development and design unit 44a...Clinical information collection unit 44b...Evaluation unit 44c...Drug candidate molecule analysis unit 45a...Simulation result database 45b...Drug discovery database 45c...Clinical information database 46...Development information transmission unit 101...Photography control unit 102...Sensor control unit 103...Observation execution unit 103a...Feedback control unit 104...Moving image recognition unit 105...Sound parameter control unit 105a...Interface control unit 105b...Environment setting unit 105c...Sound pressure setting unit 105d...Amplitude / frequency control unit 105e...Output control unit 106...Analysis unit 107...Database 108...Analysis data transmission unit 109...Sound pressure control unit 109a...Sound pressure conversion unit 109b...Frequency filter unit 109c...Inverse phase sound generation unit 109d...Specific frequency amplification unit 109e...Compressor unit

Claims

1. An animal behavior analysis system comprising: a sound parameter control unit that controls sound parameters, which are characteristics of the sound output toward the animal's movement range; an observation execution unit that observes the behavior of the animal that receives the sound output and records it as animal behavior information; an analysis unit that analyzes the correlation between the sound parameters of the output sound and the animal's behavior; and a simulation means that obtains the analysis results by the analysis unit and performs a molecular dynamics simulation of a specified protein using the characteristics of the sound output toward the animal.

2. The animal behavior analysis system according to claim 1, characterized in that the sound parameter control unit sets a target sound pressure as the specific sound parameter by a combination of volume or amplitude and frequency.

3. The animal behavior analysis system of claim 1, characterized in that the simulation means reproduces structural or functional changes in cells or proteins by placing the specified protein under 1 atmosphere of pressure in the surrounding environment set in the molecular dynamics simulation and applying a 1 / 100,000th fluctuation to the atmospheric pressure conditions.

4. The animal behavior analysis system described in claim 1, characterized in that the simulation means performs an object-type psychological and physical integrated simulation that associates psychological activity changes of neurons based on the influence of the sound parameters with physiological parameters of the body and analyzes a set of psychological and physical states.

5. The animal behavior analysis system described in claim 1, further comprising a mechanism analysis unit that references a neural network trained by stacking distribution patterns of features extracted from the correlation between the animal's behavior and structural or functional morphological changes in specific cells, receptors or proteins, and the simulation means performs a molecular dynamics simulation on the cells, receptors or proteins extracted by reference to the neural network for animal behavior information correlated with specific sound parameters based on the correlation analyzed by the analysis unit.

6. The animal behavior analysis system according to claim 1, further comprising a feedback control unit that reflects the voice parameters used in the simulation in the control by the voice parameter control unit based on the simulation results by the simulation means.

7. The animal behavior analysis system described in claim 1, characterized in that it comprises a drug discovery reference unit that references a drug discovery neural network trained by stacking distribution patterns of features extracted from structural variations or functional morphological changes in specific cells, receptors or proteins, and the correlation between proteins and drug candidate molecules, and a drug candidate molecule analysis unit that selects drug candidate molecules by referring to the drug discovery neural network for animal behavior information correlated with specific sound parameters.

8. The animal behavior analysis system described in claim 1, characterized in that the simulation means models the structure of a specified protein that is the subject of the molecular dynamics simulation, acquires audio parameters of the sound output to the animal, calculates periodic external forces acting on the atoms of the specified protein and the water molecules surrounding them based on the audio parameters for the modeling, and adds the periodic external forces to the interactions between the atoms of the specified protein and the surrounding environment, and integrates the changes over time in the relative positions and velocities between the atoms.

9. The animal behavior analysis system described in claim 5, characterized in that the neural network is trained by layering distribution patterns of features extracted from the correlation between the animal's behavior and structural or functional morphological changes of specific cells, receptors or proteins in relation to the results of microbiome analysis of DNA extracted from an intestinal flora sample of the target animal, and the mechanism analysis unit analyzes the correlation between the animal's behavior data, the voice parameters, the structural and functional changes of the protein and the composition of the intestinal flora by referring to the neural network trained in relation to the results of the microbiome analysis.

10. A method for analyzing animal behavior, comprising: an observation execution step in which a sound parameter control unit controls sound parameters, which are characteristics of the sound, and outputs them toward the animal's range of movement, and an observation execution unit observes the behavior of the animal that has received the sound output and records it as animal behavior information; and a simulation step in which an analysis unit analyzes the correlation between the sound parameters of the output sound and the animal's behavior, obtains the analysis results, and uses the characteristics of the sound output toward the animal to perform a molecular dynamics simulation of a specified protein with a simulation means.

11. The animal behavior analysis method described in claim 10, characterized in that in the observation execution step, the sound parameter control unit sets a target sound pressure as the specific sound parameter by a combination of volume or amplitude and frequency.

12. The animal behavior analysis method described in claim 10, characterized in that in the simulation step, the simulation means places the specified protein under 1 atmosphere of pressure in the surrounding environment set in the molecular dynamics simulation, and applies a 1 / 100,000th fluctuation to the atmospheric pressure conditions to reproduce structural or functional changes in cells or proteins.

13. The animal behavior analysis method according to claim 1, characterized in that in the simulation step, the simulation means executes an object-based integrated psychology-physical simulation that correlates psychological activity changes of neurons based on the influence of the sound parameters with physiological parameters of the body and analyzes a set of psychological and physical states.

14. The animal behavior analysis method of claim 10, further comprising a mechanism reference step in which a mechanism analysis unit refers to a neural network trained by stacking distribution patterns of features extracted from the correlation between the animal's behavior and structural or functional morphological changes of specific cells, receptors or proteins, and in the simulation step, the simulation means refers to the neural network for animal behavior information correlated with specific sound parameters based on the correlation analyzed by the analysis unit, and performs a molecular dynamics simulation on the extracted cells, receptors or proteins.

15. The animal behavior analysis method according to claim 10, further comprising a feedback control step in which a feedback control unit reflects the voice parameters used in the simulation in the control by the voice parameter control unit based on the simulation results by the simulation means.

16. The animal behavior analysis method described in claim 10, further comprising: a drug discovery referencing step in which a drug discovery reference unit refers to a drug discovery neural network trained by stacking distribution patterns of features extracted from structural variations or functional morphological changes of specific cells, receptors or proteins, and the correlation between proteins and drug candidate molecules; and a drug candidate molecule analysis step in which a drug candidate molecule analysis unit refers to the drug discovery neural network for animal behavior information correlated with specific sound parameters and selects drug candidate molecules.

17. The animal behavior analysis method described in claim 10, characterized in that in the simulation step, the simulation means models the structure of a specified protein that is the subject of the molecular dynamics simulation, acquires audio parameters of the sound output to the animal, calculates periodic external forces acting on the atoms of the specified protein and the water molecules surrounding them based on the audio parameters for the modeling, and adds the periodic external forces to the interactions between the atoms of the specified protein and the surrounding environment, and integrates the changes over time in the relative positions and velocities between the atoms.

18. The animal behavior analysis method described in claim 14, characterized in that the neural network is trained by layering distribution patterns of features extracted from the correlation between the behavior of the animal and structural or functional morphological changes of specific cells, receptors or proteins in association with the results of microbiome analysis of DNA extracted from an intestinal flora sample of the target animal, and in the mechanism reference step, the mechanism analysis unit refers to the neural network trained in association with the results of the microbiome analysis to analyze the correlation between the behavioral data of the animal, the voice parameters, the structural and functional changes of the protein and the composition of the intestinal flora.

19. An animal behavior analysis program that causes an information processing terminal to function as: a sound parameter control unit that controls sound parameters, which are characteristics of the sound output toward the animal's movement range; an observation execution unit that observes the behavior of the animal that receives the sound output and records it as animal behavior information; an analysis unit that analyzes the correlation between the sound parameters of the output sound and the animal's behavior; and a simulation means that obtains the analysis results by the analysis unit and performs a molecular dynamics simulation of a specified protein using the characteristics of the sound output toward the animal.

20. The animal behavior analysis program according to claim 19, characterized in that the sound parameter control unit sets a target sound pressure as the specific sound parameter by a combination of volume or amplitude and frequency.

21. The animal behavior analysis program described in claim 19, characterized in that the simulation means reproduces structural or functional changes in cells or proteins by placing the specified protein under 1 atmosphere of pressure in the surrounding environment set in the molecular dynamics simulation and applying a 1 / 100,000th fluctuation to the atmospheric pressure conditions.

22. The animal behavior analysis program described in claim 19, characterized in that the simulation means executes an object-type integrated psychology-physical simulation that correlates psychological activity changes of neurons based on the influence of the sound parameters with physiological parameters of the body and analyzes a set of psychology and physical states.

23. The animal behavior analysis program of claim 19, further comprising causing the information processing terminal to function as a mechanism analysis unit that references a neural network learned by stacking distribution patterns of features extracted from the correlation between the animal's behavior and structural or functional morphological changes of specific cells, receptors or proteins, and wherein the simulation means references the neural network for animal behavior information correlated with specific sound parameters based on the correlation analyzed by the analysis unit, and performs a molecular dynamics simulation on the extracted cells, receptors or proteins.

24. The animal behavior analysis program according to claim 19, further comprising causing the information processing terminal to function as a feedback control unit that reflects the voice parameters used in the simulation in the control by the voice parameter control unit based on the simulation results by the simulation means.

25. The animal behavior analysis program according to claim 19, characterized in that the information processing terminal further functions as: a drug discovery reference unit that refers to a drug discovery neural network trained by stacking distribution patterns of features extracted from structural variations or functional morphological changes in specific cells, receptors or proteins, and the correlation between proteins and drug candidate molecules; and a drug candidate molecule analysis unit that refers to the drug discovery neural network for animal behavior information correlated with specific sound parameters and selects drug candidate molecules.

26. The animal behavior analysis program described in claim 19, characterized in that the simulation means models the structure of a specified protein that is the subject of the molecular dynamics simulation, acquires audio parameters of the sound output to the animal, calculates periodic external forces acting on the atoms of the specified protein and the water molecules surrounding them based on the audio parameters for the modeling, and adds the periodic external forces to the interactions between the atoms of the specified protein and the surrounding environment, thereby integrating the changes over time in the relative positions and velocities between the atoms.

27. The animal behavior analysis program described in claim 24, characterized in that the neural network is trained by layering distribution patterns of features extracted from the correlation between the animal's behavior and structural variations or functional and morphological changes in specific cells, receptors or proteins in association with the results of microbiome analysis of DNA extracted from an intestinal flora sample of the target animal, and the mechanism analysis unit analyzes the correlation between the animal's behavior data, the voice parameters, the structural and functional changes in the protein and the composition of the intestinal flora by referring to the neural network trained in association with the results of the microbiome analysis.

Images