TWI291630B - Bio-expression system and the method of the same - Google Patents

Abstract

The bio-expression system comprising: a process system used to process data and a three-dimension image generating module embedded in the computing system, wherein while input of a set of two-dimensional individual model sections is fed into the process system, the three-dimension image generating module is responsive to the input of the two-dimension individual model sections and is capable of processing an individual model construction and model-averaging procedure, thereby generating an average model for each individual dataset. A database includes a bio-expression sub-database, cellular network sub-database and fine structure sub-database, wherein the database is coupled to the process system to store at least the average model. A stereoscopic projecting system is coupled to the process system to display stereoscopic images for active or passive virtual reality applications, thereby presenting the bio-expressions, cellular networks or bio-fine structures under the input instruction of the process system.

Description

1291630 IX. Description of the invention: [Technical field to which the invention pertains] The μt invention relates to a biological expression system, and in particular to a high-resolution cellular network image of soil or protein expression. [Prior Art] Gene analysis and behavioral analysis have been developed to recognize the merits of genes: for example, there are many ways to develop disease models, such as the production of fertile transgenic animal models (eg: Transgenic mice and other animals with special genetic characteristics). However, in practice, the main obstacle to performing gene/money analysis in transgenic mammals is that the animal's life cycle is too long (lGng life span). It takes too long to spend in the laboratory (at least a few years), so that it is impossible to accurately track the disease caused by abnormal genes in animals. Because scientists lack a quick and easy way to reach and accurately identify path〇genesis at the sub-level, so that most research on disease treatment is delayed. The H brother's remedy method uses a number of short-lived (only a few days from birth to maturity) insects as a model. For example, the Drosophilamelanogaster brain has been used to study Alzheimer's disease. Please refer to the document "A Drosophila m〇del of heimer s disease, dissecting the pathological roles of A" by K Iijima, 2004, proc. Natl. Acad. Sci. USA Journal, Vol. 1, pp. 6629-6628. 42 and A 40. Similarly, if a good correlation between genes, cell structure and disease can be successfully established in a Drosophila model, many studies of early detection and treatment of 1291630~ disease may become More effective. Many new and more precise treatments for diseases (especially those associated with genes) are of interest not only in science but also in public health. Although in some computer medical diagnostic systems Application, but so far, no research or application has shown that in computer technology, computer-aided systems are used to examine the cell level in the real environment. To observe the entire structure to the finest level, new technologies are needed to assist Traditional optical technology detection. Recently, 3D reconstruction technology has been developed to allow for the reconstruction of three-dimensional images of cells, thereby Tools can understand the fine structure of cells. However, the ones that are viewed by this technology focus on single cells or a few cells cultured in an artificial environment (in vitro) rather than a real entity (in vivo). The difference between the two is more important in the field of neuroscience. The nerve is actually three-dimensionally distributed in the body, but it is placed in two-dimensional conditions in cell culture. The two-dimensional environment is unlikely to simulate three-dimensional in vivo. Neural networks. There are currently some methods that are trying to observe the neurological system of the three-dimensional environment. However, these methods are limited by the penetration depth of the optical system, and it is difficult to see through the deeper depth of the micron meter using visible light. For information on methods for reconstructing individual wild type Drosophila larvae and adult brains, please refer to the following website: http://flybrain.neurobio.arizona.edu/Flybrain/html/cont rib/2000/ Rein/index.html. In the same kind of research, Drosophila became one of the main model systems in brain research. Drosophila brain (about 600x250x150 microns) contains large 200,000 neurons. In this very small brain, the fruit fly exhibits a 1,291,630 amazingly complex repertoire, such as: positioning, courtship, learning 4 and e-remembrance. The whole brain is dissected, sliced, and Fluorescent markers are used for easy viewing. However, in this and all previous methods, it is not possible to reorganize the entire fruit fly due to tissue damage caused by tissue dissection and limited by the observable depth of view. The neural network of the brain. Therefore, the present invention provides a complete and novel analytical method to overcome this obstacle. In addition, imaginary technology has now evolved to be implementable and applicable. This virtual reality technology has been widely used in the industry. A known application is the application of training and research. Virtual reality training applications can provide users with the development of important technologies and experiences without burdening them with the cost or risk of training on the ground. Virtual reality allows users to experience the virtual reality of computer production. The user's actions are converted into virtual environment (VE) input by the computer. The virtual environment system can simulate the feeling of nature, allowing users to navigate through the virtual environment as if they were in the real world. However, virtual reality systems have never been used to explore cellular networks of high resolution (in the range of a few microns) in a biological tissue. The current approach is to combine a high-resolution biological structure database (such as a neural network), a genetic (or protein) representation system of biological tissue, and a visual representation of a virtual entity. This system is a modular type that allows expansion of the expansion of multiple gene (protein) expressions, linking anatomical structures to functional (or dysfunction) to molecular levels. According to this method, the relationship between genes, cellular networks and 1291630 biological functions can be examined and manipulated in the most realistic environment. When detailed cellular network and genetic data are available, it is possible to achieve the level of biological function simulation. SUMMARY OF THE INVENTION In view of the above, it is an object of the present invention to disclose a biological representation system and method thereof. It is an object of the present invention to disclose a rendering system for a three-dimensional cellular network. The above biological expression system comprises a high-resolution cellular network (refer to the cell network mentioned later) database, a virtual entity projection system, which can express gene (or protein) representation and complete cell network in the whole tissue with high resolution. If there is a pleading for the road, the second person can connect to the biological performance system and database through the Internet. μ ' A further object of the present invention is to discover and document gene expression under normal or abnormal conditions. The invention also provides - technology for obtaining a wider range of biosoft groups

High-resolution images of ysoft bi〇l〇gical tissues), such as the mammalian brain. According to the present invention, the 'biological performance system includes a processing system for the shell material' - the two-dimensional image generation module is embedded in the processing system, and the two-dimensional individual model section of the bean cattle group is fed into the processing system, and the three-dimensional image is The group responds to the cross-section input of the two-dimensional individual model, and performs the over-type averaging calculation of the individual model. The results are generated from the individual model of the group--the average model/, the database can be classified into the gene (or protein) performance secondary database, one ^Network sub-database and fine structure times: #料库, + the above data shoulder to the processing system to store at least the average model; - Stereo projection system table 1291630 The above processing system is used to display a 3D image for active or passive virtual Entity applications, while presenting gene (or protein) representations or fine biological structures within the cellular network under the input instructions of the processing system described above. The above model averaging procedure consists of two stages of different levels of stages. The above system also includes a standard biogeographic index sub-database that can be calibrated and compared between different individual data. In one embodiment, the biological network is classified according to gene (or protein) expression, individual growth, development, disease, or certain empirically derived procedures. Cell network sub-databases can create functional links between cell entities to reflect a function (or dysfunction). The invention discloses a method for presenting by a biological expression system, including: inputting an individual model cross section (secti〇ns) to one of the above biological performance systems to process data; and utilizing the above-mentioned processing system An average model generation module of the performance system, wherein the average mode i generates a module to respond to the input of the individual model cross section, to perform the individual model building and the model average calculation, and then generate an average structure of the cell structure from all the data. A c〇arse level model is implemented by a system that introduces an average model generation right group. The above procedure first divides each individual model into several important parts. By using the above average model generation module, a skeleton is determined from the corresponding major axes of each important part of the individual model, and a local coordinate system is established for each individual model. Then perform a translation and rotation to import each local coordinate into a global coordinate system. After considering all the principals of all individual models, an average skeleton can be calculated. 11 1291630 Then, the original model can be remodeled with a 3D field-based warping algorithm built into the above average model generation module. Then, construct an average model. Next, the hollow body is transformed into a solid by using a 3D seed. algorithm. Finally, all the geometrical points of the average karyotype are determined as the best continuation of the above average model. [Implementation] I _ reference diagram and the following description, the purpose of which is only to illustrate the present invention. The preferred embodiment is not intended to limit the scope of the invention. The present invention provides a cellular network database and a gene expression system for a biological tissue, and a comparison example is a gene expression of a fruit rope brain. This performance system is a modular module and can be extended to a variety of genes with different functional types. System for Presenting Gene Expressions | Referring to Figure 1, the biological performance system 1Q of the present invention includes a computing processing system 100 for processing and computing data and information under certain instructions. This biological performance system collects and presents biological properties. In one embodiment, the above-described system can be used to analyze and define a neural circuit involving any scientific research, medical related diagnostics, or art exhibits. According to the prior art, the above object can be achieved by using a high-performance computer with an advanced central processing unit (CPU). An averaging model generation module 2 is embedded in the processing system 1 , to facilitate conversion of the input 2D image data, such as a set of individual model sections (secti〇ns) into 3D 12 1291630 images. In a preferred embodiment, a commercial application product or software such as AMIRA (3·1 Edition, manufactured by Mercury Computer Systems, USA) can be introduced to achieve the above objectives. In the conventional three-dimensional drawing technique, image data can be calculated by introducing a processing system 10〇 in the average model generating module 2 to generate a reconstructed three-dimensional image. First, the input data is prepared by the sample preparation system 500 and the image data generating system 4〇〇. The above sample preparation system 5 can be used to generate a target sample for supply to the biological performance system. In one embodiment, the Drosophila brain (a mature brain of about 6〇〇χ25〇χ15〇 microns) is used to illustrate. As will be appreciated by those skilled in the art, the above-described brain tissue of Drosophila is intended to illustrate the invention and is not intended to limit the scope of the invention. The whole tissue of the fruit rope brain can be obtained by a known method, and then by increasing the transparency of the light to about 毫米15 mm (mm) or more, the technique can be referred to the following documents, for example, referring to the inventors of the present invention - Invented technology · Refer to US Patent No. 6472216 B1, in which the Japanese name is October 29, 2002, and the patent name is Aque〇us." The above reference poor materials are stated here for reference. Inject fluorescent signs § 〇Γ labelmg) 分? 'In order to facilitate the labeling of the predetermined part of the sample: knife (4) is called some of the functions of the function. These procedures can be accomplished by genetic engineering of the prior art. φ , , " production system The 400 includes a scanning microscope equipped with a plurality of right-handed lasers. During this procedure, the laser is scanned and labeled to activate the fluorescent molecules.

13 S 1291630 Cross-sections (cr〇ss-sections) of different depths of the sample can be scanned completely (or partially) with a laser. Therefore, the scanned image data contains a plurality of surface images of different depths. Different parts of the same cross-section can be reconstructed with the aid of a computer software such as AMIRA. The resulting image data is then fed to the processing system 1 described above to facilitate subsequent processing. As previously described, the average model generation module 2 is coupled to the processing system 100 to process the input data to produce a three dimensional image or an average model. The resulting average model or 3D image data may be stored in the database 600. The above database may include a plurality of secondary databases, such as a biological performance secondary database 610, a cellular network secondary database 62, and a biological fine structure secondary database 630. It is to be understood by those skilled in the art that the preferred embodiments described herein are intended to illustrate the invention and not to limit the scope of the invention. The biological performance secondary database 610 contains information on biological characteristics, such as gene (or protein) expression. The database 61〇 also contains a standard biogeography index (bio_ge〇graphic index) sub-database that can be calibrated and compared between different individuals. } White matter) performance, individual growth, development, and classification. All data can be derived from genes (or eggs, diseases or some empirically derived procedures)

^^(dysfunctions)) ^ ^ ^ ^ M ^ 0 言^ temporarily inhibits the function of dynamin in neurons, blocking

14 1291630

Encodes a Phospholipase-A2 and Define a Neural Circuit, Involved in Anesthesia-Resistant Memory", from Current Biology, Vol. 14, pp. 263-272, published February 17, 2004, by Ann-ShynChiang This document is hereby incorporated by reference. The above-mentioned fine structure sub-database 630 contains biologically fine structured materials. A preferred embodiment constructs a Drosophila brain database. Gene expression of the Drosophila brain It can be represented by Gal4/UAS-GFP technology, which is known to those skilled in the art. Figure 6 illustrates an example of generating a three-dimensional fruit fly brain, which is hereby incorporated by reference. Three-dimensional image, this figure shows an image of many nerve cells showing the radish gene in the brain of the adult male fruit in the whole brain tissue. The green part is the expression of the radish gene indicated by green fluorescent protein. The medium to large green dots and the connecting nerve fibers in between are clearly visible. The larger spots are neurons with a width of about 10 microns, and the smallest points. Φ (dots) is considered to be synapses, and its size is approximately in the micrometer range. The brown part is the mushroom body. The braided system is constructed according to the algorithm of the average model module in the above processing system. All cells can be placed in the calculated standard brainframe by 3D image generation techniques. Figure 7 shows a standard Drosophila brain model constructed according to its scorpion (purple), orange for GH146 The location of the gene neuron. The yellow part is the optical lobe, which is the collection of visual signals. The blue part is called the central complex. In Figure 8, there are several cell networks that display different genes. In a brain, the line shows (: s 15 1291630 for GiH46 (green), tim (color) and 2 for the biological clock of Drosophila. (Line color). The above genes are from Figure 6 to Figure 8, biological The gentleman arrives. The bio-network image can be stored and constructed by the present invention - a computer readable storage medium for storage - :::: body. The invention is disclosed by the use - laser scanning microscope: sweeping The fluorescent molecules in the sample are 吝:/, the sample that has a private (不abel) has a plurality of laser light sources; the scanning process is scanned by laser light, and the cross-section of the depth is Get = scan different scanning image data according to a predetermined sequence. (4) Multiple surface images of the same depth.. u. The images of different parts of the rake face can be combined - Μ) as a whole. Three-dimensional objects can be reconstructed by the use of computer software such as ships (10). The present invention further comprises the steps of: preparing a sample and injecting a molecule that potentially produces camp light into the sample prior to scanning the sample: and genetically engineering to predict the target sample (tons to 眶_). The light transparency of the above sample was increased to about 〇 15 mm (mm) or more after the file. Referring to Figure 2, the flow of the average model (3D atlas) from a group of initial individual models is shown in the present invention. As shown in Figure 2, this method consists of two main steps, the individual model construction and the two model averaging procedures at different levels. The first step is to construct a two-dimensional framework model (wireframe model;) for each individual data set. Individual model Performs two-dimensional segmentation and contouring on the original data set.

16 1291630 After the contour analysis and the rim analysis, the surface model reconstruction algorithm can be used to obtain a wireframe model representing the individual. The above can be achieved by the assistance of computer software such as AMIRA. Spindle selection and spindle The second step is the coarse level model averaging. Each individual model is cut into several important parts (sub-models) by performing a cutting step using the user interface. For each individual model, a set of spindles can be picked to serve as a skeleton for the above model. In this step, the individual model ® is fed into the processing system described above, and the system can process the cutting program under the input command of the user. Each individual wireframe model is manually segmented into several important parts, and then the corresponding main axis of each sub-model can be obtained by the technique of PCA, which is disclosed by Ian T Jolliffe and its content is published. In 1986 in New York, Springer-Verlag revealed the "Principal Component Analysis". By calculating the eigenvalues of the following sample variation matrix (covariance matrix) and the corresponding eigenvectors, the direction of the principal axis can be determined by:

τ where m is the number of vertices on the secondary model, X is the position vector of the above vertices, and // X is the sample average of X. The direction D of the main axis is the eigenvector with the smallest eigenvalue S. The major axis of the secondary model is the rotational axis with the smallest rotational inertia in the secondary model, which can be represented as a parameter line segment with one of the parameters t: 17 1291630 v(;)-A + I^D ^ ^ <Γ<^ where A is One point on the spindle, set to / to / / x. The boundaries, tmin and tmax, can be determined by the projection of all sub-model vertices to the main axis (pr〇j ecti〇n). For each individual model, the selected set of spindles can be considered as the model skeleton for that individual. Spindle average Refer to Figure 2, after the spindle is determined by the above processing system or computer. The registration of each individual model is performed prior to the execution of the averaging procedure. Each individual model can be built from its own principal axis two correlation locations—the local coordinate system, and then each individual model is translated and rotated by the computing system. After applying some translation and rotation, the original local coordinate system according to the global axis is recorded. The above parameter line segment can be calculated by using D α where Μ is the midpoint of the main axis, and

The average processing of the spindle averaging program can be expressed as: by calculating the average position of the midpoint, the average length &', the result of the computer or the invention, the average spindle can be expressed as:

18 1291630 v ( I) .= μ.Μ + ’ * μΒ,.’ draw s I s I leg where μΜ =

Where n is the number of individual models. After the processing system has completed calculating the average major axis of the sub-model, the processing system generates an average skeleton of the original data set and stores it in the memory of the processing system 100 or in the database 600 of Figure 1. Figure 3 shows the wireframe model of the mushroom body. Figure 4 shows the main shaft and frame of the same model as Figure 3. Figure 5 shows a partial coordinate system for a set of spindles. Three Dimension Field-Based Warping After the average skeleton of the original data set is generated, the next program then deforms the stored three-dimensional individual skeleton data set by the processing system 100 of the present invention. Based on the field-based 0 warping algorithm disclosed by J. Gomes et al., entitled "Warping and Morphing of Graphical Objects", published by the Morgan Kaufman publisher in San Francisco in 1999, executed by the processing system 100. The calculation can be transformed for each individual model to its relative pseudo-average model. The comparison function processed by the computer is defined as: rr(p):=p + ^J7-, Δρ|=%(ρ)-p \ i 19 1291630 where P is the position vector of the vertex on a volume model, And r is the value of the major axis in a set of skeletons. (8) is the position of P after the change. The weight of W(8) in the Kth main axis is defined as:

Ka + ^J where lk is the length of the major axis, and its importance can be adjusted by the constant c. Is the fairy from? The distance from the point to the main axis. The constant a represents the dependence of the main axis, and the constant b can be regarded as the strength of the main axis strength. The final level model average average model can be produced by the above described processing system 100 in accordance with the above model. Then, the average model is recorded according to the general average skeleton. The final syllabic modal 3L can be obtained by the geometric intermediate point of the averaging model on the processing system (10) described above. For the frame model, t-e patches can be converted into stereoscopic volume pixels by sampling. The three-dimensional filling algorithm is used to convert this hollow body into a solid object. The model will generate a solid 'to indicate the stacking volume of all its solid volume U β from 1 to N pixel values, which can be superimposed by N like-average models ^. The geometric intermediate point is located at a pixel value of N/2 Once the geometric intermediate point is determined by the above processing system 1 and 3, the final average model can be established. The generated virtual reality device example full color three-dimensional neuron icon can be operated by the device. View 20 1291630 • In order to demonstrate a very fine extension of the nerve, several devices are provided in the data generation system 400. A Zeiss LSM 510 confocal microscope is equipped with four laser sources, including one Argon laser (emitting at 364 nm), an argon-helium laser (wavelength 458, 488 or 514 nm) and two xenon lasers (wavelengths 543 and 633 nm). This system allows simultaneous detection of four fluorescent lights. The Zeiss LSM 510 META confocal two-photon microscope system is equipped with four laser sources, which contain a gas 氪 laser (wavelength 458, 488 or 5 14 nm), two xenon lasers (wavelengths 543 and 633 nm), and a coherent Mira that is supplied to a nonlinear optical microscope (2-photon) (femtosecond) T-Sapphire lasers, which can be set-tuned at 700-100 Onrn. This is a vivo observation of fluorescent signals designed for use in thick living tissue. The Zeiss LSM 510 ΜΕΤΑ confocal two-photon microscope system is equipped with three laser sources, including an argon-helium laser (wavelength 458, 488 or 514 nm) and two xenon lasers (wavelengths 543 and 633 nm). Three photomultipliers and a ΜΕΤΑ detector to allow simultaneous collection of full-spectrum fluorescent signals. The system does not have a penetrating light detector. The system has an image collage (image montage) An automated stage scanner and an infrared optical system. A stereoscopic image projection system 300 is coupled to the processing system 100 for stereoscopic image presentation. The processing system 100 described above can access the database ' under the input command and transmit the image to a video card to present a graphical output of the various 21 1291630 (e.g., NVIDIA Quadro 4-980 or better). The central processing unit (CPU) in processing system 100 can be one (or more) 32-bit or 64-bit (or higher order) units with sufficient memory for image data processing. The various output images described above are individually fed into a plurality of projectors such that stereoscopic presentation and operation of the front or rear projection can be implemented. The above procedure can be controlled by commercial software (such as amira ν·3.1) and hardware (such as three-dimensional mouse). A special lens of a known technique can be provided for generating a virtual three dimensional image. It is a well-known technique known to all, so it is not described here. The present invention has been described above in terms of preferred embodiments, but it is not intended to limit the scope of patent rights claimed by the invention. The scope of patent protection depends on the scope of the patent: and the scope of the equivalent. Those who are familiar with the technical attack in this field are all in the spirit or scope of the patent, and the changes or refinements made: = the equivalent m material completed in the spirit of the (4), Wei Yingxi is included in the following patent application scope. . [Simple description of the drawing] By the following detailed description of the McCaw test, the above viewpoints and the advantages will be able to replace the tips of the secrets, and it is easy to understand the spirit of the present invention by the following description and additional drawings. . Wherein the first figure shows a system diagram in accordance with the present invention. The second figure shows a flow chart in accordance with the present invention. The third figure shows the frame model of the scorpion. : 3 shows: The main axis and line architecture of the same model in Figure 3. The system shows a local coordinate system of a set of spindles. < s 22 1291630 The sixth figure shows an example of a three-dimensional substructure and a network generated in the brain of Drosophila. ^ The seventh figure shows the Drosophila brain model with some major substructures. g Figure 8 shows some of the distribution within the neural network in the Drosophila brain. Soil [main component symbol description] ίο biological performance system 100 processing system 200 average model generation module 300 stereoscopic image projection system 400 image data generation system 500 sample preparation system 600 database 61 〇 biological performance secondary database 620 cell network times Database 630 Biological Fine Structure Sub-Database 23

Claims (1)

Patent application scope: A biological expression system comprising: 1291630 a sample preparation system for providing a sample of the biological representation to the biological expression system; - an image data generation system for feeding the target sample; a processing system for processing data of the image data generating system; an average model generating module embedded in the processing system, wherein when the two-dimensional individual model is input into the processing system, the average model generating module: Violation:: the input of the individual model, to carry out the individual model construction and the model average calculation 'to form a set of individual models to generate an average model; database & 表现 biological performance sub-data 细胞, cell network sub-database and biology a fine structure secondary database, wherein the data county is integrated into the processing, and the average model can be stored at least; and the two stereoscopic projection system is coupled to the processing system for displaying a stereoscopic image 1 for active or Passive virtual reality applications, while presenting biological representations, cellular networks, or fine biological structures under the input of the processing system. 2. The biological performance system of claim 1 of the patent scope, wherein the model average procedure comprises an averaging phase of two different levels. 3. The biological expression system of claim i, wherein the cell network comprises a neural circuit. 4. The biological expression system of claim 3, wherein the biological form 24 S 1291630. now comprises a gene or protein expression. 5. The biological performance system of claim 1 of the patent scope, wherein the database further comprises a standard biogeographic index sub-database, which can be calibrated and compared between different individuals. 6. A biological expression system as claimed in claim 1, wherein the biological expression is classified according to procedures for gene expression, individual growth, development, disease, or certain empirical processes. The female claims the biological expression system of item 1 of the patent scope, wherein the cell network library can establish a functional link between the cells to reflect a certain function (or dysfunction). 8. The biological expression system of claim 3, further comprising a two-dimensional %-shirt image data generation system comprising a laser scanning microscope configured with a plurality of laser light sources, wherein a prepared sample is fed into the image data The generating system 'at least a portion of the cross-section of the prepared sample can be scanned by the laser source, and the cross-sections of different depths can be scanned according to a predetermined sequence, resulting in a plurality of surface images being generated at different depths to feed the process SiS. 9. A method of producing a biological structure presentation of a biological performance system, comprising: inputting an individual model to a processing system of the biological performance system to process 25 1291630 data; and quoting an average embedded in the biological performance system through the processing system a type generating weight group, wherein the average model generating module should input an individual type to perform an individual model construction and a model average calculation, and the result generates an average model from a set of inputs; wherein the model average program includes executing each of the One of the individual models is a rough hierarchical model average to pick a set of major axes, and the main axis can be regarded as a skeleton of a model; and by using the average model generating module, an average is determined from all of the skeletons of all the individual models a skeleton, wherein the individual model is fed into the processing system; wherein the model averaging program includes performing registration of each of the individual models; and by using the average model generating module, each part of the individual model is established from a relevant position of the main axis Coordinate system; and using the average generation model module, borrowing Each of the individual models is translated and rotated by the processing system; wherein after applying the translations and rotations, the original local coordinate systems ordered by the respective coordinate axes are registered. 10. The method for generating a biological structure presentation of a biological expression system according to claim 9 wherein the coarse-level model comprises performing a segmentation step to distinguish each of the individual models into a plurality of important parts (sub-model) . '11. The method for generating a biological structure of a biological expression system according to claim 9 of the patent application, wherein the model averaging program comprises: 26 1291630, the processing system is processed in all the original models by all the main axes of the average sub-model a midpoint, a direction and a length to generate an average skeleton of the original model, and storing the data set; using the average model to generate a module according to the three-dimensional curvature and deforming the original model according to the three-dimensional average skeleton to generate the relative The average model is used; and the average model is used to generate a module to determine a geometrical point between the geometric model to perform a fine-level model averaging, thereby obtaining a final average model of the biological structure presentation. & A method for producing a biological structure of a biological expression system according to the scope of claim U, wherein the model averaging program comprises: converting a triangle patches into a solid hollow body by sampling; applying a three-dimensional filling algorithm (3D seed_fill) Alg〇rithm) converts the solid hollow body into a solid body, wherein each of the similar average models produces the solid body; and the voxel value is N/2 as a critical value of the cumulative volume to determine the midpoint, wherein N is like an average model superposition value. 13. The current production method of the biological expression system of claim 9 includes a production-two-dimensional image data, which is produced by a confocal microscope equipped with a plurality of laser light sources, at least two, and prepared separately. The sample is scanned by the laser source, and the cross section of the sample is scanned according to a predetermined sequence, and the plurality of surface image results are generated at different depths: 55 27 1291630 Feed into the processing system. 14. A medium that can be read and stored by a computer, which can use an image of the network, which is scanned by a laser scanning microscope to indicate that it is produced in a sample. a light molecule, wherein the laser scanning microscope 褒^ 稷 a plurality of laser light sources; during the scanning process, at least ― eight of the samples are scanned by the laser light, and according to the predetermined scanning depth of the horizontal Cut the surface, the result is the image of the calf surface; + 丨 "Cut the number of images/woodiness of different parts of the same cross section of the table; and the coincidence of all 2D cross-section images of different depths to become a three-dimensional object (10) The computer of the scope of item 14 can read the storage medium, and before the sample is scanned, the method further comprises the steps of: preparing the sample; implanting a labeled molecule capable of generating glory in the sample, and marking the substrate by a project a predetermined portion of the target sample; and increasing the light transparency of the sample to about 毫米15 mm or more. The student 16. The computer readable storage medium as claimed in claim 14 The passage comprising a neural network of the brain. 28 Ί291630 17. The scope of the patent application 16 of the first computer readable storage medium, the raw material web 1 comprising a neural network of the Drosophila brain. 29