KR102558550B1 - Apparatus and method for generating prediction result for tcr using artificial intelligence technology - Google Patents

Abstract

Disclosed is a method for generating a prediction result using artificial intelligence technology, performed by a computing device. The method may comprise the steps of: obtaining first data corresponding to CDR3α of a TCR and second data corresponding to CDR3β of the TCR, wherein the first data includes an amino acid sequence corresponding to the CDR3α, and the second data includes an amino acid sequence corresponding to the CDR3β; receiving the first data and the second data and determining whether the first data including the CDR3α and the second data including the CDR3β correspond to a precedence/subsequence relationship using a first module based on artificial intelligence; and storing, when a result indicating that the first data and the second data correspond to the precedence/subsequence relationship is output, a combination of the first data and the second data in a CDR set candidate list. According to the present invention, the method is capable of more effectively and accurately predicting, determining, or identifying a complementarity determining region of a TCR.

Description

Method and apparatus for generating prediction results for TCR using artificial intelligence technology

The present disclosure relates to artificial intelligence technology, and more specifically, to derive a TCR (T Cell Receptor) configuration effective for inducing a cellular response using artificial intelligence technology.

The major histocompatibility complex (MHC) is a genetic locus that encodes 'MHC molecules' that act in the immune system. MHC molecules exist in type 1 (class I) and type 2 (class II). Immunopeptidome refers to a set of peptides expressed on the cell surface, and for example, the immunopeptidome may refer to a combination of MHC-related peptides. Human Leukocyte Antigen (HLA) is a glycoprotein molecule produced by the human major histocompatibility complex gene. HLA is not present in mature erythrocytes, but is expressed in immature erythroblasts and is expressed on the surface of all tissue cells in the human body, including blood cells such as leukocytes and/or platelets. The MHC gene is present in all vertebrates, and the human MHC gene is referred to as the HLA gene, and the product expressed therefrom is referred to as HLA. MHC genes are involved in self and non-self recognition, immune response to antigen stimulation, regulation of cellular and humoral immunity, and susceptibility to disease. HLA, a product of the MHC gene, is the second most important antigen next to the ABO blood type in survival of the transplanted organ in solid organ transplantation.

HLA, like MHC, can be largely classified into Class I and Class II. Class I is classified as HLA-A, HLA-B, and HLA-C and is expressed in most nucleated cells and platelets, and is essential for antigen recognition when cytotoxic T cells recognize and eliminate virus-infected cells or tumor cells. HLA Class II is classified into HLA-DR, HLA-DQ, and HLA-DP, and is expressed in B cells, monocytes, dendritic cells, and activated T cells. It interacts with the antigen receptor of helper T cells to induce cellular and humoral immune responses, and is known to be essential when recognizing antigens expressed on antigen-presenting cells. HLA is a gene that shows the largest polymorphism among genes possessed by humans, and there is a frequency difference between races and ethnic groups.

When peptides derived from proteins derived from infectious microorganisms or proteins specific to cancer cells bind to MHC and are presented on the cell surface, T cells recognize them and trigger an immune response to eliminate infected cells or cancer cells. As such, T cells are key regulators (players) that determine specific immune responses to foreign substances that do not exist in the normal human body. Therefore, the prediction of the T Cell Receptor (TCR) that binds to pMHC can be used for the development of personalized vaccines for the prevention of infectious diseases or cancer.

The manufacturing process of a T cell receptor-engineered T cell (TCR-T) includes the steps of creating a gene sequence for the TCR, introducing the TCR gene sequence into a T cell isolated from a patient cell, and proliferating and culturing the T cell into which the TCR has been introduced. A viral vector is mainly used as a method for introducing the TCR gene into T cells, and a process of activating T cells to introduce TCR-T and treating cytokines such as interleukin-2 for TCR-T proliferation is used. For cloning of the TCR gene, T cells with high affinity for MHC-tumor antigens are selected as a complex of MHC and antigen epitope peptides, and then the TCR gene of these cells is cloned. Codon optimization is performed for each of the TCR α chain and β chain genes, and this process is known to influence the increase in TCR expression on the T cell surface.

Korean Registered Patent No. 10-2322832

The present disclosure has been made in response to the above background art, and is to predict or identify a peptide-MHC complex and a TCR binding thereto in a more efficient manner and/or more accurately.

The technical problems of the present disclosure are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the description below.

According to one embodiment of the present disclosure, a method performed by a computing device is disclosed. The method includes: acquiring first data corresponding to CDR3α of T cell receptor (TCR) and second data corresponding to CDR3β of TCR, wherein the first data includes an amino acid sequence corresponding to CDR3α, and the second data includes an amino acid sequence corresponding to CDR3β -, receiving the first data and the second data using an artificial intelligence-based first model, and including the CDR3α Determining whether data 1 and the second data including the CDR3β correspond to a precedence relationship, and when a result is output that the first data and the second data correspond to the precedence relationship, storing a combination of the first data and the second data in a CDR set candidate list.

In one embodiment, the first data corresponding to CDR3α of the TCR is generated by a second model different from the first model, the second model is an artificial intelligence-based model pretrained to receive a peptide-MHC conjugate (pMHC) and output the first data including CDR3α corresponding to the pMHC, and the second data corresponding to CDR3β of the TCR is a third model different from the first model , and the third model may be an artificial intelligence-based model pretrained to take the pMHC as an input and output the second data including CDR3β corresponding to the pMHC.

In one embodiment, the first data corresponding to CDR3α of the TCR and the second data corresponding to CDR3β of the TCR may be generated by a fourth model different from the first model.

In one embodiment, the fourth model may include an encoder and a decoder.

In one embodiment, the fourth model may be an artificial intelligence-based model pre-trained such that a first input dataset including an amino acid sequence corresponding to a peptide and MHC is input to the encoder, an amino acid sequence corresponding to CDR3α and CDR3β related to the peptide and MHC is input to the decoder, and the first data and a prediction result including the first data are output from the decoder.

In one embodiment, the first model comprises: generating a first negative dataset by randomly combining a first dataset corresponding to CDR3α or a second dataset corresponding to CDR3β of TCR; identifying CDR3α and CDR3β identified as binding to each other in the randomly combined first negative dataset; It can be pretrained based on the steps of generating a second negative dataset by excluding from one negative dataset, and including the combination of CDR3α and CDR3β identified as binding in the first positive dataset.

In one embodiment, combinations of CDR3α and CDR3β included in the second speech dataset may be labeled as correct answer data that do not correspond to a precedence relationship in the learning process of the first model.

In one embodiment, combinations of CDR3α and CDR3β included in the first positive dataset may be labeled as correct answer data corresponding to a precedence relationship in the learning process of the first model.

In one embodiment, the first model may be pre-trained based on additional steps: identifying at least one similar combination having a degree of similarity equal to or greater than a predetermined threshold to the CDR3α and CDR3β combination included in the first positive dataset, among the combinations of CDR3α and CDR3β included in the second negative dataset, and including the identified similar combination in the first positive dataset.

In one embodiment, the first model may include a Recurrent Neural Network (RNN), a Long Short Term Memory (LSTM) network, a Bidirectional Long Short Term Memory (BiLSTM) network, a Generative Pre-trained Transformer (GPT), Diffusion model, Bidirectional Encoder Representations from Transformers (BERT), spanBERT, Gated Recurrent Unit (GRU), or Bidirectional Gated Recurrent Unit (BiGRU).

In one embodiment, the CDR set candidate list may refer to a combination of TCR α chain and *?* included in TCRs capable of binding to pMHC.

In one embodiment, a combination of the first data and the second data stored in the CDR set candidate list may be used to generate a TCR-T.

In one embodiment, a computer program stored on a computer readable storage medium is disclosed. The computer program, when executed by a computing device, causes the computing device to perform operations for generating a prediction result using artificial intelligence technology, wherein the operations include: obtaining first data corresponding to CDR3α of TCR and second data corresponding to CDR3β of TCR, the first data including an amino acid sequence corresponding to the CDR3α, and the second data including an amino acid sequence corresponding to the CDR3β, using a first artificial intelligence-based model The method may include receiving the first data and the second data and determining whether the first data including the CDR3α and the second data including the CDR3β correspond to a precedence relationship, and storing a combination of the first data and the second data in a CDR set candidate list when a result that the first data and the second data correspond to the precedence relationship is output.

A computing device according to an embodiment is disclosed. The computing device may include at least one processor and memory. The at least one processor: obtaining first data corresponding to CDR3α of TCR and second data corresponding to CDR3β of TCR, the first data including an amino acid sequence corresponding to CDR3α, and the second data including an amino acid sequence corresponding to CDR3β, receiving the first data and the second data using an artificial intelligence-based first model, and receiving the first data and the second data including the CDR3α An operation of determining whether data and the second data including the CDR3β correspond to a precedence relationship, and an operation of storing a combination of the first data and the second data in a CDR set candidate list when a result is output that the first data and the second data correspond to the precedence relationship.

The method and device according to one embodiment of the present disclosure can predict, determine or identify the complementarity determining region of the TCR in a more efficient manner and/or more accurately.

1 schematically illustrates a block configuration diagram of a computing device according to an embodiment of the present disclosure.
2 illustrates an exemplary structure of an artificial intelligence-based model according to an embodiment of the present disclosure.
3 illustrates an exemplary method for obtaining predictive results related to CDR3α and CDR3β of a TCR from input data using an artificial intelligence-based predictive model according to an embodiment of the present disclosure.
4 exemplarily illustrates a method for generating a prediction result including combination information for CDR3α of TCR and CDR3β of TCR, according to an embodiment of the present disclosure.
5 exemplarily illustrates a method of generating data including information of CDR3α of TCR or CDR3β of TCR according to an embodiment of the present disclosure.
6 illustratively illustrates first positive data, first negative data, and second negative data including information related to a combination of CDR3α of TCR and CDR3β of TCR, according to an embodiment of the present disclosure.
7 illustratively illustrates storing a combination of CDR3α and CDR3β in a CDR set candidate list, according to an embodiment of the present disclosure.
8 is a schematic diagram of a computing environment according to one embodiment of the present disclosure.

Various embodiments are described with reference to the drawings. In this specification, various descriptions are presented to provide an understanding of the present disclosure. Prior to describing specific details for the implementation of the present disclosure, it should be noted that configurations not directly related to the technical gist of the present disclosure have been omitted within the scope of not distracting from the technical gist of the present invention. In addition, the terms or words used in the present specification and claims should be interpreted as meanings and concepts consistent with the technical idea of the present invention based on the principle that the inventor can define the concept of appropriate terms in order to best describe his/her invention.

As used herein, the terms “component,” “module,” “system,” “unit,” and the like refer to a computer-related entity, hardware, firmware, software, a combination of software and hardware, or an execution of software, and may be used interchangeably. For example, a component may be, but is not limited to, a procedure, processor, object, thread of execution, program, and/or computer running on a processor. For example, both an application running on a computing device and a computing device may be components. One or more components may reside within a processor and/or thread of execution. A component can be localized within a single computer. A component may be distributed between two or more computers. Also, these components can execute from various computer readable media having various data structures stored thereon. Components may communicate via local and/or remote processes, e.g., according to a signal with one or more packets of data (e.g., data from one component interacting with another component in a local system, distributed system, and/or data transmitted via a signal to another system and over a network such as the Internet).

In addition, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or”. That is, unless otherwise specified or clear from the context, “X employs A or B” is intended to mean one of the natural inclusive substitutions. That is, X uses A; X uses B; Or, if X uses both A and B, "X uses either A or B" may apply to either of these cases. Also, the term "and/or" as used herein should be understood to refer to and include all possible combinations of one or more of the listed related items.

Also, the terms "comprises" and/or "comprising" should be understood to mean that the features and/or components are present. However, it should be understood that the terms "comprises" and/or "comprising" do not exclude the presence or addition of one or more other features, elements, and/or groups thereof. Also, unless otherwise specified or where the context clearly indicates that a singular form is indicated, the singular in this specification and claims should generally be construed to mean "one or more".

In addition, the term "at least one of A or B" or "at least one of A and B" is to be interpreted as meaning "includes only A", "includes only B", and "combines the configuration of A and B".

Those of skill should further appreciate that the various illustrative logical components, blocks, modules, circuits, means, logics, and algorithms described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate the interchangeability of hardware and software, various illustrative components, blocks, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented in hardware or as software depends on the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of this disclosure.

The description of the presented embodiments is provided to enable any person skilled in the art to use or practice the present invention. Various modifications to these embodiments will be apparent to those skilled in the art of this disclosure. The general principles defined herein may be applied to other embodiments without departing from the scope of this disclosure. Thus, the present invention is not limited to the embodiments presented herein. The present invention is to be accorded the widest scope consistent with the principles and novel features set forth herein.

In the present disclosure, terms expressed as Nth, such as first, second, or third, are used to distinguish at least one entity. For example, the first and second entities may be the same as or different from each other.

In the present disclosure, for convenience of explanation, human leukocyte antigen (HLA) will be used as an example for MHC (Major Histocompatibility Complex) as an example. Therefore, the description of HLA or MHC used below is an example for expressing a description of MHC or HLA, and the scope of the present disclosure will be determined based on the content described in the claims, and through examples of HLA, the scope of rights will not be construed as limited to HLA. As such, HLA and MHC in the present disclosure may be used interchangeably.

As used in the present disclosure, the term “human leukocyte antigen (HLA)” is a glycoprotein molecule produced by the human MHC gene, and is a gene that shows the largest polymorphism among genes possessed by humans. HLA typing, which determines the HLA type, can be used very actively in various fields such as organ transplantation, immunotherapy, disease-related research, paternity tests such as paternity, forensic use, and genetic research.

An HLA type in the present disclosure may include, for example, an HLA-A type, an HLA-B type, and/or an HLA-C type.

The conjugate of MHC and peptide in the present disclosure may refer to a peptide antigen presenting complex through MHC class I molecules through processing through proteasome in antigen presenting cells. The proteasome is composed of two units, LMP-2 and LMP-7 (low molecular weight polypeptide). These two proteasome units are located near the TAP-1 and TAP-2 genes in the MHC gene. The proteasome subunits are particularly important for the degradation of peptides bound to MHC I molecules. When cells are treated with interferon gamma (IFN-γ), a cytokine, the expression of LMP-2 and LMP-7 can be induced. Expression of LMP-2 and LMP-7 results in a change in the substrate specificity of the proteasome, increasing its ability to degrade into peptides. IFN-γ increases the expression of proteins related to antigen presentation, such as MHC I and MECL-1, as well as LMP proteins, so that antigen presentation can be increased in antigen presenting cells.

MHC I molecules bind not only to peptide antigens degraded by the proteasome, but also to peptides produced by proteolytic enzymes present in the endoplasmic reticulum.

Tapasin serves as a bridge between TAP-1 (transporter associated with antigen processing) and MHC I, which has a stable tertiary structure in the endoplasmic reticulum (ER).

There are two types of T cell receptor (TCR), mainly composed of TCRα and TCRβ. The TCRα chain is located on chromosome 14, and the TCRβ chain is located at an independent locus on chromosome 7. TCRβ is composed of a V gene segment, a D gene segment, a J gene segment, and a C gene segment, similar to an immunoglobulin (Ig) heavy chain, and a TCRα chain does not have a D gene segment, similar to an Ig light chain, and consists of V, J, and C gene segments.

In the recombination process of TCR, the TCRα chain is encoded by the V, J, and C gene segments, similar to the immunoglobulin light chain, and the TCRβ chain is formed by a gene recombination process in which C binds to V-D-J formed by DJ binding and then V linkage. The linkage of V-(D)-J is mediated by RAG-1 and RAG-2 (recombination-activating gene).

The hairpin structure is cleaved by an endonuclease, and a complementary P-nucleotide is added to the short strand of DNA formed at this time. In addition, 1 to 20 nucleotides may be randomly added to the cleavage ends, which are referred to as N-nucleotides, and this process is mediated by TdT (terminal deoxynucleotide transferase). P- and N-nucleotides increase the diversity of T-cell receptors.

T cells recognize only foreign antigens presented on the cell surface, and the activity of mature peripheral T cells is initiated by the interaction between antigens and peptides presented at the peptide binding cleft of TCR and MHC molecules.

In particular, most immune profiling is focused on the analysis of CDR3 (complementarity determining region 3, complementarity determining region) region in the TCR sequence. The CDR3 region is an important region involved in the interaction between an antigen and a receptor, and the most mutations are identified.

Examples in the present disclosure include the step of deriving a sequence of a TCR including a peptide including an antigenic peptide such as neoantigen, CMV, EBV, and KRAS, and a CDR3 corresponding to a conjugate of MHC molecules (pMHC) according to the type of HLA and the phenotype by race/individual, with the main purpose of immunoactivation of T cells induced through the interaction between the peptide-MHC complex (pMHC) and the TCR.

1 schematically illustrates a block configuration diagram of a computing device 100 according to one embodiment of the present disclosure.

Computing device 100 according to an embodiment of the present disclosure may include a processor 110 and a memory 130 .

The configuration of the computing device 100 shown in FIG. 1 is only a simplified example. In one embodiment of the present disclosure, the computing device 100 may include other components for performing a computing environment of the computing device 100, and only some of the disclosed components may constitute the computing device 100.

The computing device 100 in the present disclosure may mean any type of node constituting a system for implementing embodiments of the present disclosure. The computing device 100 may refer to any type of user terminal or any type of server. Components of the aforementioned computing device 100 are exemplary and some may be excluded or additional components may be included. For example, when the aforementioned computing device 100 includes a user terminal, an output unit (not shown) and an input unit (not shown) may be included within the scope of the computing device 100 .

The computing device 100 in the present disclosure may perform technical features according to embodiments of the present disclosure to be described later. For example, the computing device 100 uses an artificial intelligence-based predictive model using input data corresponding to peptides, MHC, and TCR, and information on CDR3α, the α chain of the TCR corresponding to the input data, and the β chain of the TCR. It can generate a prediction result including information about CDR3β. For example, the computing device 100 may generate a prediction result for a TCR (eg, CDR3, CDR1, and/or CDR2) using peptides and MHC from a sample obtained from a subject. For example, the computing device 100 obtains information on a peptide, MHC, and TCR (e.g., CDR3, CDR1, or CDR2) from a sample obtained from a subject, and based on the information on the peptide, MHC, and TCR, CDR3α amino acid sequence of the TCR and prediction results corresponding to the CDR3β amino acid sequence of the TCR may be generated.

In one embodiment of the present disclosure, the computing device 100 may obtain a result of base sequence analysis (eg, Next Generation Sequencing) from a server or an external entity. In another embodiment, the computing device 100 may perform nucleotide sequence analysis on proteomic data and/or genetic data (eg, DNA or RNA) obtained from a biological sample derived from a subject. As used in the present disclosure, sequencing may be performed by any type of technique capable of analyzing base sequences, and may include, for example, whole genome sequencing, whole exome sequencing, or whole transcriptome sequencing, but is not limited thereto.

As used in the present disclosure, a subject may refer to a subject or individual for obtaining a biological sample containing a peptide, MHC, TCR, and/or a complex thereof.

As used in the present disclosure, the term and sample may be used without limitation as long as they are obtained from an individual or subject for whom MHC type and/or TCR are to be determined, and may be, for example, cells or tissues obtained by biopsy, blood, whole blood, serum, plasma, saliva, cerebrospinal fluid, various secretions, urine and/or feces. Preferably, the sample may be selected from the group consisting of blood, plasma, serum, saliva, nasal fluid, sputum, ascites, vaginal secretion and/or urine, and more preferably blood, plasma or serum. The sample may be pre-treated prior to use in detection or diagnosis. For example, pretreatment methods may include homogenization, filtration, distillation, extraction, concentration, inactivation of interfering components, and/or addition of reagents, and the like. In the present disclosure, a biological sample may be tissue, cell, whole blood, and/or blood, but is not limited thereto.

In one embodiment, the processor 110 may include at least one core, and may include a processor for data analysis and/or processing, such as a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), or a tensor processing unit (TPU) of the computing device 100.

The processor 110 may read a computer program stored in the memory 130 to generate a predictive result including immunogenicity of the PMHC-TCR according to an embodiment of the present disclosure.

According to an embodiment of the present disclosure, the processor 110 may perform an operation for learning a neural network. The processor 110 may perform calculations for neural network learning, such as processing input data for learning in deep learning (DL), extracting features from input data, calculating errors, and updating neural network weights using backpropagation. At least one of the CPU, GPGPU, and TPU of the processor 110 may process learning of the network function. For example, the CPU and GPGPU can process learning of network functions and data classification using network functions. In addition, in one embodiment of the present disclosure, learning of a network function and data classification using a network function may be processed by using processors of a plurality of computing devices together. In addition, a computer program executed in a computing device according to an embodiment of the present disclosure may be a CPU, GPGPU or TPU executable program.

Additionally, the processor 110 may typically handle overall operations of the computing device 100 . For example, the processor 110 processes data, information, or signals input or output through components included in the computing device 100 or drives an application program stored in a storage unit, thereby providing or processing appropriate information or functions to the user.

According to one embodiment of the present disclosure, memory 130 may store any type of information generated or determined by processor 110 and any type of information received by computing device 100 . According to one embodiment of the present disclosure, the memory 130 may be a storage medium that stores computer software that causes the processor 110 to perform operations according to embodiments of the present disclosure. Accordingly, the memory 130 may refer to computer readable media for storing software codes necessary for performing embodiments of the present disclosure, data subject to execution of the codes, and results of execution of the codes.

According to an embodiment of the present disclosure, the memory 130 may refer to any type of storage medium. For example, the memory 130 may include a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (eg SD or XD memory, etc.), RAM (Random Access Memory, RAM), SRAM (Static Random Access Memory), ROM (Read-Only Memory, ROM), E It may include at least one type of storage medium among electrically erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, and optical disk. The computing device 100 may operate in relation to a web storage that performs a storage function of the memory 130 on the Internet. The description of the above memory is only an example, and the memory 130 used in the present disclosure is not limited to the above example.

The communication unit (not shown) in the present disclosure may be configured regardless of its communication aspect, such as wired and wireless, and may be configured with various communication networks such as a personal area network (PAN) and a wide area network (WAN). In addition, the communication unit may operate based on the known World Wide Web (WWW), and may use a wireless transmission technology used for short-range communication such as Infrared Data Association (IrDA) or Bluetooth.

The computing device 100 in the present disclosure may include any type of user terminal and/or any type of server. Accordingly, embodiments of the present disclosure may be performed by a server and/or a user terminal.

A user terminal may include any type of terminal capable of interacting with a server or other computing device. The user terminal may include, for example, a mobile phone, a smart phone, a laptop computer, a personal digital assistant (PDA), a slate PC, a tablet PC, and an ultrabook. A server may include any type of computing system or computing device, such as, for example, microprocessors, mainframe computers, digital processors, portable devices and device controllers, and the like.

In a further embodiment, the aforementioned server may refer to an entity that stores and manages TCR information, immunopeptidome information, peptide sequence information, MHC molecule information, nucleotide sequence information, or gene information. The server may include a storage unit (not shown) for storing immunopeptidome information, peptide sequence information, positional amino acid identifier information, nucleotide sequence information, gene information, or reliability information of a database (e.g., Expasy, VDJdb, TCR3F, huARdb, IMGT, McPAS), and the storage unit may be included in the server or may exist under the management of the server. As another example, the storage unit may exist outside the server and may be implemented in a form capable of communicating with the server. In this case, the storage unit may be managed and controlled by another external server different from the server.

2 illustrates an exemplary structure of an artificial intelligence-based model according to an embodiment of the present disclosure.

Throughout this specification, the terms predictive model, artificial intelligence-based predictive model, artificial intelligence model, artificial intelligence-based model, computational model, neural network, network function, and neural network may be used interchangeably.

A neural network may consist of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network includes one or more nodes. Nodes (or neurons) constituting neural networks may be interconnected by one or more links.

In a neural network, one or more nodes connected through a link may form a relative relationship of an input node and an output node. The concept of an input node and an output node is relative, and any node in an output node relationship with one node may have an input node relationship with another node, and vice versa. As described above, an input node to output node relationship may be created around a link. More than one output node can be connected to one input node through a link, and vice versa.

In a relationship between an input node and an output node connected through one link, the value of data of the output node may be determined based on data input to the input node. Here, a link interconnecting an input node and an output node may have a weight. The weight may be variable, and may be changed by a user or an algorithm in order to perform a function desired by the neural network. For example, when one or more input nodes are connected to one output node by respective links, the output node may determine an output node value based on values input to input nodes connected to the output node and a weight set for a link corresponding to each input node.

As described above, in the neural network, one or more nodes are interconnected through one or more links to form an input node and output node relationship in the neural network. Characteristics of the neural network may be determined according to the number of nodes and links in the neural network, an association between the nodes and links, and a weight value assigned to each link. For example, when there are two neural networks having the same number of nodes and links and different weight values of the links, the two neural networks may be recognized as different from each other.

A neural network may be composed of a set of one or more nodes. A subset of nodes constituting a neural network may constitute a layer. Some of the nodes constituting the neural network may form one layer based on distances from the first input node. For example, a set of nodes having a distance of n from the first input node may constitute n layers. The distance from the first input node may be defined by the minimum number of links that must be passed through to reach the corresponding node from the first input node. However, the definition of such a layer is arbitrary for explanation, and the order of a layer in a neural network may be defined in a method different from the above. For example, a layer of nodes may be defined by a distance from a final output node.

In one embodiment of the present disclosure, a set of neurons or nodes may be defined as a layer.

An initial input node may refer to one or more nodes to which data is directly input without going through a link in relation to other nodes among nodes in the neural network. Alternatively, in a relationship between nodes based on a link in a neural network, it may mean nodes that do not have other input nodes connected by a link. Similarly, the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the neural network. Also, the hidden node may refer to nodes constituting the neural network other than the first input node and the last output node.

The neural network according to an embodiment of the present disclosure may have the same number of nodes of the input layer as the number of nodes of the output layer, and may be a neural network in which the number of nodes decreases and then increases again as the number of nodes progresses from the input layer to the hidden layer. In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes of the input layer may be less than the number of nodes of the output layer and the number of nodes decreases as the number of nodes progresses from the input layer to the hidden layer. In addition, the neural network according to another embodiment of the present disclosure may have more nodes in the input layer than nodes in the output layer, and the number of nodes increases as the number of nodes increases from the input layer to the hidden layer. A neural network according to another embodiment of the present disclosure may be a neural network in the form of a combination of the aforementioned neural networks.

A deep neural network (DNN) may refer to a neural network including a plurality of hidden layers in addition to an input layer and an output layer. Deep neural networks can reveal latent structures in data. That is, the structure of a photograph, text, video, audio, protein sequence structure, gene sequence structure, structure of a peptide sequence, structure of an amino acid sequence, potential structure of music (e.g., which object is in a photograph, content and emotion of text, content and emotion of voice, etc.), binding affinity between TCR and pMHC, and/or binding affinity between peptide and MHC. The deep neural network may include a convolutional neural network (CNN), a recurrent neural network (RNN), an auto encoder, a Generative Adversarial Network (GAN), a Restricted Boltzmann Machine (RBM), a deep belief network (DBN), a Q network, a U network, a Siamese network, and the like. The description of the deep neural network described above is only an example, and the present disclosure is not limited thereto.

As an example, the AI-based predictive model of the present disclosure may include a Recurrent Neural Network (RNN), a Long Short Term Memory (LSTM) network, a Bidirectional Long Short Term Memory (BiLSTM) network, Bidirectional Encoder Representations from Transformers (BERT), spanBERT, Gated Recurrent Unit (GRU), or Bidirectional Gated Recurrent Unit (BiGRU).

The artificial intelligence-based predictive model of the present disclosure may be represented by a network structure of any of the foregoing structures including an input layer, a hidden layer, and an output layer.

Neural networks that can be used in the artificial intelligence-based model of the present disclosure can be trained in at least one of supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. Learning of the neural network may be a process of applying knowledge for the neural network to perform a specific operation to the neural network. As an example, the predictive model according to an embodiment of the present disclosure applies a mask to at least some of the frames or units constituting the amino acid sequences, and then matches the masked frames or units. It can be learned by a semi-supervised learning method.

A neural network can be trained in a way that minimizes output errors. In neural network learning, iteratively inputs training data to the neural network, calculates the output of the neural network for the training data and the error of the target, and backpropagates the error of the neural network from the output layer of the neural network to the input layer in order to reduce the error, thereby updating the weight of each node of the neural network. In the case of supervised learning, each learning data is labeled with the correct answer (ie, labeled learning data), and in the case of unsupervised learning, the correct answer may not be labeled in each learning data. That is, for example, learning data in the case of supervised learning related to data classification may be data in which each learning data is labeled with a category. Labeled training data is input to a neural network, and an error may be calculated by comparing an output (category) of the neural network and a label of the training data. As another example, in the case of unsupervised learning for data classification, an error may be calculated by comparing input learning data with a neural network output. The calculated error is back-propagated in a reverse direction (ie, from the output layer to the input layer) in the neural network, and connection weights of each node of each layer of the neural network may be updated according to the back-propagation. The amount of change in the connection weight of each updated node may be determined according to a learning rate. The neural network's computation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently according to the number of iterations of the learning cycle of the neural network. For example, a high learning rate may be used in the early stage of neural network training to increase efficiency by allowing the neural network to quickly obtain a certain level of performance, and a low learning rate may be used in the late stage to increase accuracy.

In learning of a neural network, in general, training data may be a subset of real data (i.e., data to be processed using the trained neural network), and therefore, a learning cycle may exist in which errors for the training data decrease but errors for the actual data increase. Overfitting is a phenomenon in which errors for actual data increase due to excessive learning on training data. For example, a phenomenon in which a neural network that has learned a cat by showing a yellow cat does not recognize that it is a cat when it sees a cat other than yellow may be a type of overfitting. Overfitting can act as a cause of increasing the error of machine learning algorithms. Various optimization methods can be used to prevent such overfitting. To prevent overfitting, methods such as increasing training data, regularization, inactivating some nodes of the network during learning, dropout, and using a batch normalization layer may be applied.

A computer readable medium storing a data structure according to an embodiment of the present disclosure is disclosed. The above-described data structure may be stored in a storage unit in the present disclosure, executed by a processor, and transmitted and received by a communication unit.

Data structure can refer to the organization, management, and storage of data that enables efficient access and modification of data. Data structure may refer to the organization of data to solve a specific problem (eg, data retrieval, data storage, data modification in the shortest time). A data structure may be defined as a physical or logical relationship between data elements designed to support a specific data processing function. A logical relationship between data elements may include a connection relationship between user-defined data elements. A physical relationship between data elements may include an actual relationship between data elements physically stored in a computer-readable storage medium (eg, a persistent storage device). The data structure may specifically include a set of data, a relationship between data, and a function or command applicable to the data. Through an effectively designed data structure, a computing device can perform calculations while using minimal resources of the computing device. Specifically, the computing device can increase the efficiency of operation, reading, insertion, deletion, comparison, exchange, and search through an effectively designed data structure.

The data structure can be divided into a linear data structure and a non-linear data structure according to the shape of the data structure. A linear data structure may be a structure in which only one data is connected after one data. Linear data structures may include lists, stacks, queues, and decks. A list may refer to a series of data sets in which order exists internally. The list may include a linked list. A linked list may be a data structure in which data are connected in such a way that each data is connected in a single line with a pointer. In a linked list, a pointer can contain information about connection to the next or previous data. A linked list can be expressed as a singly linked list, a doubly linked list, or a circular linked list depending on the form. A stack can be a data enumeration structure that allows limited access to data. A stack can be a linear data structure in which data can be processed (eg, inserted or deleted) at only one end of the data structure. The data stored in the stack may be a LIFO-Last in First Out (Last in First Out) data structure. A queue is a data listing structure that allows limited access to data, and unlike a stack, it can be a data structure (FIFO-First in First Out) in which data stored later comes out later. A deck can be a data structure that can handle data from either end of the data structure.

The nonlinear data structure may be a structure in which a plurality of data are connected after one data. The non-linear data structure may include a graph data structure. A graph data structure can be defined as a vertex and an edge, and an edge can include a line connecting two different vertices. A graph data structure may include a tree data structure. The tree data structure may be a data structure in which one path connects two different vertices among a plurality of vertices included in the tree. That is, it may be a data structure that does not form a loop in a graph data structure.

Throughout this specification, the terms predictive model, artificial intelligence-based model, computational model, neural network, network function, and neural network may be used interchangeably. Hereinafter, a neural network is unified and described. The data structure may include a neural network. And the data structure including the neural network may be stored in a computer readable medium. The data structure including the neural network may also include pre-processed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, an activation function associated with each node or layer of the neural network, a loss function for learning the neural network, and the like. A data structure including a neural network may include any of the components described above. That is, the data structure including the neural network may include all or any combination of data preprocessed for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation functions associated with each node or layer of the neural network, loss functions for learning the neural network, and the like. In addition to the foregoing configurations, the data structure comprising the neural network may include any other information that determines the characteristics of the neural network. In addition, the data structure may include all types of data used or generated in the computational process of the neural network, but is not limited to the above. A computer readable medium may include a computer readable recording medium and/or a computer readable transmission medium. A neural network may consist of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network includes one or more nodes.

The data structure may include data input to the neural network. A data structure including data input to the neural network may be stored in a computer readable medium. Data input to the neural network may include training data input during a neural network learning process and/or input data input to a neural network that has been trained. Data input to the neural network may include preprocessed data and/or data to be preprocessed. Pre-processing may include a data processing process for inputting data to a neural network. Accordingly, the data structure may include data subject to pre-processing and data generated by pre-processing. The foregoing data structure is only an example, and the present disclosure is not limited thereto.

The data structure may include weights of the neural network (weights and parameters may be used interchangeably in this specification). And the data structure including the weight of the neural network may be stored in a computer readable medium. A neural network may include a plurality of weights. The weight may be variable, and may be changed by a user or an algorithm in order to perform a function desired by the neural network. For example, when one or more input nodes are connected to one output node by respective links, the output node may determine a data value output from the output node based on values input to input nodes connected to the output node and a weight set for a link corresponding to each input node. The foregoing data structure is only an example, and the present disclosure is not limited thereto.

As a non-limiting example, the weights may include weights that are varied during neural network training and/or weights for which neural network training has been completed. The variable weight in the neural network learning process may include a weight at the time the learning cycle starts and/or a variable weight during the learning cycle. The weights for which neural network learning has been completed may include weights for which learning cycles have been completed. Accordingly, the data structure including the weights of the neural network may include a data structure including weights that are variable during the neural network learning process and/or weights for which neural network learning is completed. Therefore, it is assumed that the above-described weights and/or combinations of weights are included in the data structure including the weights of the neural network. The foregoing data structure is only an example, and the present disclosure is not limited thereto.

The data structure including the weights of the neural network may be stored in a computer readable storage medium (eg, a memory or a hard disk) after going through a serialization process. Serialization can be the process of converting a data structure into a form that can be stored on the same or another computing device and later reconstructed and used. A computing device may serialize data structures to transmit and receive data over a network. The data structure including the weights of the serialized neural network may be reconstructed on the same computing device or another computing device through deserialization. The data structure including the weights of the neural network is not limited to serialization. Furthermore, the data structure including the weights of the neural network may include a data structure (e.g., B-Tree, R-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree in a nonlinear data structure) to increase computational efficiency while minimizing resource use of a computing device. The foregoing is only an example, and the present disclosure is not limited thereto.

The data structure may include hyperparameters of the neural network. Also, the data structure including the hyperparameters of the neural network may be stored in a computer readable medium. A hyperparameter may be a variable variable by a user. Hyperparameters may include, for example, a learning rate, a cost function, the number of repetitions of learning cycles, weight initialization (eg, setting a range of weight values to be targeted for weight initialization), and the number of hidden units (eg, the number of hidden layers and the number of nodes in the hidden layer). The foregoing data structure is only an example, and the present disclosure is not limited thereto.

A transformer may be considered as a network function for a predictive model according to an embodiment of the present disclosure. As an example, the predictive model may be operated based on a transformer. Such a predictive model may be operated using, for example, a recurrent neural network to which an attention algorithm is applied or a transformer to which an attention algorithm is applied.

In one embodiment, a transformer may consist of an encoder that encodes the embedded data and a decoder that decodes the encoded data. The transformer may have a structure that receives a series of data and outputs a series of data of different types through encoding and decoding steps. In one embodiment, the series of data can be processed into a form operable by a transformer. A process of processing a series of data into a form in which a transformer can operate may include an embedding process. Expressions such as data token, embedding vector, and embedding token may refer to embedded data in a form that can be processed by a transformer.

In order for the transformer to encode and decode a series of data, encoders and decoders within the transformer may be processed using an attention algorithm. The attention algorithm may refer to an algorithm that calculates an attention value by calculating the similarity of one or more keys for a given query, reflecting the given similarity to each key and corresponding value, and then weighting the values reflecting the similarity.

Depending on how to set the query, key, and value, various types of attention algorithms can be classified. For example, when attention is obtained by setting the same query, key, and value, this may mean a self-attention algorithm. In order to process a series of input data in parallel, when the dimensions of an embedding vector are reduced and individual attention heads are obtained for each divided embedding vector to obtain attention, this may mean a multi-head attention algorithm.

In one embodiment, a transformer may consist of modules that perform a plurality of multi-head self-attention algorithms or multi-head encoder-decoder algorithms. In one embodiment, the transformer may also include additional elements other than the attention algorithm, such as an embedding layer, a normalization layer, and a softmax layer. Methods for constructing transformers using the attention algorithm may include methods disclosed in Vaswani et al., Attention Is All You Need, 2017 NIPS, which is incorporated herein by reference.

Transformers can be applied to various data domains such as embedded natural language, embedded sequence information, segmented image data, and audio waveforms to convert a series of input data into a series of output data. In order to convert data having various data domains into a series of data that can be input to the transformer, the transformer can embed the data. Transformers can process additional data representing relative positional or phase relationships between a set of input data. Alternatively, a series of input data may be embedded by additionally reflecting vectors representing a relative positional relationship or phase relationship between input data to the series of input data. In one example, the relative positional relationship between a series of input data may include, but is not limited to, word order in a natural language sentence, relative positional relationship of each segmented image, temporal sequence of segmented audio waveforms, and the like. A process of adding information representing a relative positional relationship or phase relationship between a series of input data may be referred to as positional encoding.

One example of how to embed data and transform it into a transformer is disclosed in Dosovitskiy, et al., AN IMAGE IS WORTH 16X16 WORDS: TRANSFORMERS FOR IMAGE RECOGNITION AT SCALE, which document is incorporated herein by reference.

3 illustrates an exemplary method for obtaining predictive results related to CDR3α and CDR3β of a TCR from input data using an artificial intelligence-based predictive model according to an embodiment of the present disclosure.

Regarding "whether TCR α chain and TCR β chain can be combined" in the present disclosure, "whether CDR3α and CDR3β correspond to each other in a precedent-or-following relationship" will be used as an example. Therefore, the description of whether a precedent relationship or whether a combination is possible used below is an example for expressing a description of whether a precedence relationship is applicable or whether a combination is possible, and the scope of rights should not be construed as being limited through the example of “whether CDR3α and CDR3β correspond to precedence relationship”. As such, in the present disclosure, “whether combinable” of TCR α chain and TCR β chain and “whether a precedence relationship” between CDR3α and CDR3β can be used interchangeably.

3 illustrates an exemplary method for obtaining predictive results related to CDR3α and CDR3β of a TCR from an input data set using an artificial intelligence-based predictive model according to an embodiment of the present disclosure. In the present disclosure, the step of generating a predictive result of a predictive model may include content for training the predictive model and content for actually obtaining an inference (eg, prediction including information on CDR3α and CDR3β) through the learned predictive model.

In the present disclosure, the operation of the predictive model may be described as one of a process of learning a predictive model or a process of performing inference using the predictive model (eg, a process of generating a prediction result). This is written for the convenience of description, and even if it is described as one of the processes of performing learning or inference, it should be interpreted as intended to cover the other process of learning or inference.

In one embodiment, the steps shown in FIG. 3 may be performed by computing device 100 . In a further embodiment, the steps in FIG. 3 may be implemented by a plurality of entities, such that some of the steps shown in FIG. 3 are performed in a user terminal and others are performed in a server.

In one embodiment of the present disclosure, the computing device 100 may obtain first data corresponding to CDR3α of TCR and second data corresponding to CDR3β of TCR (S310).

In a further embodiment, the first data may include an amino acid sequence corresponding to CDR3α of the TCR, and the second data may include an amino acid sequence corresponding to CDR3 of the TCR.

In one embodiment of the present disclosure, information on CDR3α included in the first data and information on CDR3β included in the second data may be output through an artificial intelligence-based or rule-based generation model different from the first model. For example, the information on CDR3α may be generated using a second model having pMHC as an input and CDR3α corresponding to the pMHC as an output. As another example, the information on CDR3β may be generated using a third model having pMHC as an input and CDR3β corresponding to the pMHC as an output.

In one embodiment, amino acid sequence information of peptides, amino acid sequence information of MHC, and/or amino acid sequence information of CDR3, CDR1, and CDR2 of TCR may be obtained through sequencing of a proteomic sample obtained from a subject.

In one embodiment, amino acid sequence information of peptides, amino acid sequence information of MHC, and/or amino acid sequence information of CDR3, CDR1, and CDR2 of TCR may be obtained from a public DB (eg, VDJdb, Expasy, etc.).

In one embodiment, the computing device 100 may use a mass spectrometer (eg, mass spectrometer, LC-MS/MS) that performs mass spectrometry in analyzing the amino acid sequence of TCR, peptide, or MHC. In one embodiment, the computing device 100 may use a Sanger sequencing method, Edman Degradation, or PMF analysis (Peptide Mass Fingerprinting) to analyze the amino acid sequence of a peptide, MHC or TCR.

In one embodiment of the present disclosure, the computing device 100 receives the first data and the second data using the artificial intelligence-based first model, and determines whether the first data including the CDR3α and the second data including the CDR3β correspond to a precedence relationship (S320).

In one embodiment, the first data and the second data may be CDR3α of TCR and CDR3β of TCR that interact with a specific peptide-MHC conjugate (pMHC). In the case of data identified as CDR3α and CDR3β to be combined as a result of input to the first model, it may be regarded as True Positive data. For example, when CDR3α and CDR3β identified as being combined are determined to actually bind to the pMHC even in a molecular biological experiment, the first data including the CDR3α and the second data including the CDR3β are input to the first model and the output result that the CDR3α and the CDR3β correspond to a precedence relationship can be stored in the database as True Positive.

In an additional embodiment, in order to secure true negative data, the first data and the second data may be CDR3α of TCR and CDR3β of TCR that are not combined with each other. Specifically, when the CDR3α and CDR3β identified as not being combined with each other are determined to be not actually combined with each other even in molecular biological experiments, the first data or the second data is input to the decoder and outputted. The result that CDR3α and CDR3β do not correspond to a precedence relationship can be stored in the database as True Positive. Such True Positive, False Positive, True Negative, or False Negative (true positive, false positive, true negative, or false negative) data may be stored and/or used as a dataset for training a predictive model.

In a further embodiment, the coincidence/inconsistency between the molecular biological test result for whether CDR3α and CDR3β are combined and the prediction result for whether CDR3α and CDR3β are combined output from the first model do not follow logical and temporal precedence. For example, if the combination of CDR3α and CDR3β is first determined not to be combined experimentally by molecular biology, and then predicted to not be combined with the CDR3α and CDR3β in the first model, the combination of CDR3α and CDR3β can be stored as True Negative data. As another example, if it is first predicted that the combination of CDR3α and CDR3β is combined in the first model, and then it is determined that the combination of CDR3α and CDR3β is not combined through molecular biological experiments, the combination of CDR3α and CDR3β can be stored as False Positive data.

In an additional embodiment, the computing device 100 may use and/or store as true (e.g., true positive or true negative) data the results of molecular biological experiments on whether CDR3α and CDR3β are combined, regardless of the prediction result through the first model on whether or not CDR3α and CDR3β are combined.

In one embodiment, the computing device 100 evaluates the performance of the predictive model using True Positive, False Positive, True Negative, or False Negative data on whether or not the combination of CDR3α and CDR3β is determined according to the prediction model and molecular biological experiment results. Specifically, as the number of input data increases, the performance evaluation accuracy of the predictive model may improve. For example, if the performance of a predictive model that is repeatedly trained with various training datasets improves, the ratio of outputting false positive and false negative results may decrease compared to true positive and true negative results.

In one embodiment, molecular biological experiments on the combination of CDR3α and CDR3β may be performed through IFN-γ detection, FACS, flow cytometry, LC-MS / MS analysis after exposure of pMHC to TCR, etc., but is not limited thereto.

In one embodiment, the computing device 100 may pre-process the identifiers of amino acids constituting the sequence included in the first data related to CDR3α, the identifiers of amino acids included in the second data related to CDR3β, the identifiers of amino acids constituting the sequence corresponding to the peptide, and/or the identifiers of amino acids constituting the sequence corresponding to MHC. In a further embodiment, the first data or the second data related to the TCR may include amino acid sequences constituting at least one of the complementary determining regions (eg, CDR3α, CDR3β) constituting the T cell receptor. In a further example, preprocessing for learning or inference of a predictive model (e.g., the first model, the second model, or the third model in the present disclosure) is performed using peptides and MHC (e.g., HLA, mouse MHC) as input data. Guide information generated through an external prediction model that operates. Specifically, the guide information generated through the external prediction model may include HLA type (eg, HLA-B14: 02, HLA-A02: 03, etc.), TCR amino acid sequence length information, TCR V (D) J type information and the like. For example, the amino acid sequence length information of the TCR together with the V/J type information may enable a prediction model for a portion of CDR3 having a large amino acid diversity to set a wide range of amino acid prediction result diversity of output data.

In an additional embodiment, the TCR information corresponding to the first data and/or the second data may include a V/D/J type obtained from a public data DB for T cell receptors. In a further embodiment, the TCR information corresponding to the first data and / or the second data may include a known TCR amino acid sequence (eg, CDR1, CDR2, CDR3) extracted from a public database for T cell receptors. In a further embodiment, the TCR information corresponding to the first data and/or the second data may include a proteomic sequence obtained from an organism-derived subject and experimentally sequenced.

In one embodiment, at least one of the first data and the second data may be obtained from data output from an external separate prediction model different from the first model.

In a further embodiment, the external separate prediction model different from the first model may be a rule-based or artificial intelligence-based model that takes pMHC as input data and uses CDR3 of a TCR that binds to pMHC as output data. In one embodiment of the present disclosure, the peptide included in pMHC is an antigen-peptide, and may be a peptide found in a specific tissue (eg, lesion tissue, tumor tissue). In a further embodiment, the peptide included in pMHC may be a peptide presented on MHC molecules of somatic cells or antigen presenting cells. In a further embodiment, the peptide included in pMHC may be a peptide extracted from a public data DB. In a further embodiment, the peptide included in pMHC may be a peptide having a sequence consisting of 6 to 11 amino acids, which is assumed to be an arbitrary antigen-peptide, but is not limited thereto.

In a further embodiment, the MHC included in the pMHC may include an HLA type obtained from a public DB for the major histocompatibility complex. In a further embodiment, the MHC included in the pMHC may include a known amino acid sequence extracted from a public DB for the major histocompatibility complex. In a further embodiment, the MHC included in the pMHC may include a proteomic sequence obtained from an organism-derived subject and experimentally sequenced.

In one embodiment, the first data or the second data may include input data to which Blosum encoding or one-hot encoding is applied to an amino acid sequence.

In one embodiment, the first data or the second data may further include at least one of a first characteristic representing polarity between amino acids, a second characteristic representing the size of amino acids, a third characteristic representing hydrophobicity or hydrophilicity of amino acids, a fourth characteristic representing whether an amino acid has a charge, or a fifth characteristic representing aromatic or aliphatic amino acids.

In one embodiment, data related to MHC or TCR may include Multiple Sequence Alignment (MSA) based on evolutionary biological genome data.

In one embodiment, correct answer data for whether CDR3α and CDR3β are combined can be determined by molecular biology experiments. In a further example, the molecular biological experiment, for example, reacts the TCR and pMHC in which the constitutive CDR3α and CDR3β are combined (e.g., coculture of peptide and lymphocytes, coculture of peptide and TCR-T (TCR engineered T cell), introduction of peptide proteome into somatic cells and exposure to lymphocytes, etc.). and CDR3β can be determined.

In an additional embodiment, answer data on whether the CDR3α of the TCR and the CDR3β of the TCR are combined (eg, paired to form a TCR) may be obtained from a public DB (eg, VDJdb, Expasy).

A further description of a method for predicting whether or not the combination of CDR3α and CDR3β of the first model (eg, whether or not a precedent relationship corresponds) will be described later with reference to FIG. 4 .

In one embodiment of the present disclosure, the computing device 100 may store a combination of the first data and the second data in the CDR set candidate list when a result is output that the first data and the second data correspond to a precedence relationship (S330).

In one embodiment, when the first model predicts that the CDRα included in the first data and the CDRβ included in the second data correspond to a precedence relationship, the computing device 100 may determine that the CDRα and CDRβ may be combined with each other to form a TCR.

Additional descriptions of the negative dataset including random combinations of CDRα and CDRβ and the positive dataset including real combinations of CDRα and CDRβ used for learning and inference of the first model and learning of the first model will be described later with reference to FIGS. 4 to 7.

4 exemplarily illustrates a method for generating a prediction result including combination information for CDR3α of TCR and CDR3β of TCR, according to an embodiment of the present disclosure.

FIG. 4 illustratively illustrates a method for learning or inferring a prediction model that generates a prediction result including whether or not a precedence relationship between CDR3α and CDR3β is applied through a preprocessed input data set according to an embodiment of the present disclosure.

In one embodiment, FIG. 4 shows an input data set 420 consisting of data related to TCR obtained by preprocessing raw data 41a and 41b including amino acid sequences of CDR3α and CDR3β of TCR, and a prediction model 4300 driven using the input data set 420 and a prediction result 440 (eg, a combination of CDR3α and CDR3β or CDR3α and Combination of CDR3β is not possible). As an example, the model shown in FIG. 4 may correspond to the predictive model 4300 .

In one embodiment, the computing device 100 may output the preceding prediction result 440 of the TCR α chain and the TCR β chain through the artificial intelligence-based prediction model 4300 .

In the process of learning the predictive model 4300 in FIG. 4 , reference number 440 may be used as label data including a first label indicating that CDR3α and CDR3β can be combined or a second label indicating that CDR3α and CDR3β cannot be combined.

According to one embodiment of the present disclosure, whether or not the combination of the TCR α chain and the TCR β chain inferred by the prediction model 4300 is possible is information corresponding to CDR3α of TCR (41a) and / or information corresponding to CDR3β of TCR. It can be generated based on (41b).

In one embodiment, in the process of receiving information necessary for the predictive model 4300 to infer the immunogenicity of pMHC-TCR, the computing device 100 may input information 41a corresponding to CDR3α of TCR and/or information 41b corresponding to CDR3β of TCR in the form of a series of continuous data 410. In an embodiment, in the process of learning and/or reasoning, first data corresponding to CDR3α of TCR and second data corresponding to CDR3β of TCR may be preprocessed and used as the input data set 420 .

In a further example, preprocessing for learning or inference of the prediction model may be performed through guide information generated through an external prediction model operating with peptides and MHC (eg, HLA, mouse MHC) as input data. Specifically, the guide information generated through the external prediction model may include HLA type (eg, HLA-B14: 02, HLA-A02: 03, etc.), TCR amino acid sequence length information, TCR V (D) J type information and the like. For example, the amino acid sequence length information of the TCR together with the V/J type information may enable a prediction model for a portion of CDR3 having a large amino acid diversity to set a wide range of amino acid prediction result diversity of output data.

In one embodiment, the contiguous data 410 of information 41a corresponding to CDR3α of TCR and/or information 41b corresponding to CDR3β of TCR may include a unique classification token [cls] in front of information 41a corresponding to CDR3α of TCR. In one embodiment, as shown in FIG. 4, an input data set 420 in which a series of consecutive data 410 is preprocessed may include a delimiter token [sep] between information 41a corresponding to CDR3α of TCR and information 41b corresponding to CDR3β of TCR.

In one embodiment, as shown in FIG. 4 , the first data 41a corresponding to CDR3α of the TCR may be pre-processed into E ₁ to E _N . In a further embodiment, the second data 41b corresponding to CDR3β of the TCR may be pre-processed into E _1' to E _N' , as shown in FIG. 4 . In a further embodiment, the first data or the second data related to the TCR may include amino acid sequences constituting at least one of the T cell receptor-forming complementary determining regions (eg, CDR3α, CDR3β).

In one embodiment of the preprocessing process, the CDR amino acid sequences 41a and 41b of the TCR may be segmented, for example, based on guide information (eg, information through an external model) or randomly. For example, when the second data related to CDR3β of the TCR includes the amino acid sequence CASSPGTGGALAEQFFGPG, the CDR3β can be segmented as CAS/SPG/T/GGA/LA/EQ/FFGPG or CASS/PGT/GG/A/LAEQ/FFGPG.

In one embodiment, the computing device 100 pre-processes raw data to generate units E ₁ to E _N constituting CDR3α 41a and units E _1′ to E _N constituting CDR3β 41b. In an additional embodiment, the amino acid sequence included in the first data and the second data may be obtained based on pMHC (40) formed by binding MHC and a peptide, but is not limited thereto.

For example, when CDR3α included in the first data 41a and CDR3β included in the second data 41b are a pair of α and β chains of a TCR that bind to a specific pMHC (40), the first model 4300 receiving the first data and the second data must determine that the corresponding CDR3α and CDR3β are combined to achieve the prediction purpose of the present disclosure. .

As another example, when CDR3α included in the first data and CDR3β included in the second data are paired with the α and β chains of the TCR that do not bind to a specific pMHC (40), the first model 4300 receiving the first data 41a and the second data 41b must determine that the corresponding CDR3α and CDR3β are not combined to achieve the prediction purpose of the present disclosure. can see

In one embodiment of the present disclosure, information 402 on CDR3α included in the first data 41a and information 403 on CDR3β included in the second data 41b may be output through an artificial intelligence-based or rule-based generation model (e.g., the second model 4020 and the third model 4030) different from the first model. In one embodiment, information 402 on CDR3α may be generated using a second model 4020 that takes pMHC 40 as an input and outputs CDR3α corresponding thereto. In one embodiment, information 403 on CDR3β may be generated using a third model 4030 having pMHC 40 as an input and a corresponding CDR3β as an output.

In one embodiment, the raw data of the first data 41a and the second data 41b input to the first model generates a result of CDR3α and CDR3β corresponding to the corresponding CDR3α, or CDR3β and CDR3β. It may be obtained by an external model different from the first model that generates a result of CDR3α corresponding to the corresponding CDR3β. Additional description of the external model different from the first model will be described later with reference to FIG. 5 .

In one embodiment, computing device 100 may embed preprocessed input data sets 420 for CDR3α and CDR3β. In a further embodiment, the computing device 100 may positional embed the location information of the preprocessed input data set 420 for CDR3α and CDR3β, the [cls] token, and the [sep] token.

In one embodiment, token embedding and position embedding may be included in the process of learning or inferring the predictive model. The learning method of the first model 4300 may include embedding tokenized or segmented amino acid identifiers into the input data set 420 of CDR3α and CDR3β generated by preprocessing the raw data 410 . In one embodiment according to the present disclosure, the learning method of the first model 4300, the second model 4020, or the third model 4030 may include embedding positional data reflecting order information in the amino acid sequence of units constituting the CDR3α and CDR3β sequences of the TCR. For example, if the number of amino acids in the CDR3α sequence of the TCR is 9 and the number of amino acids in the CDR3β sequence of the TCR is 13, the first model 4300 can be trained with 22 (e.g., sum of 9 and 13) position embedding vectors.

In one embodiment according to the present disclosure, positional embedding may mean an operation for reflecting positional information about amino acids constituting pMHC and/or TCR. For example, positional embedding may be performed in a separate embedding layer for reflecting sequence information for amino acids.

As a further example, E ₁ to E _N , E _1' to E _N' may represent a token comprising one or more amino acids. For example, a token constituting a part of a TCR sequence may be a sequence of a plurality of amino acids, such as CASS, CAS, ASS, CA, AS, SS, C, A, S, or a single amino acid.

In one embodiment, whether a precedence relationship between CDR3α and CDR3β may be predicted based on the NSP task 450 of the natural language process. For example, the first model 4300 may use a process of determining a relationship between two sentences of Next Sentence Prediction (NSP) that predicts whether two sentences are connected sentences in a natural language processing process. In one embodiment, the NSP task 450 may be followed by "Those airports including ICN, JFK, FRA, CDG, BER, MAN and HKG are located in Incheon, New York, Frankfurt, Paris, Berlin, Manchester and Hongkong." (IsNext) process can be used.

In an additional embodiment, the NSP task 450 may use a process that determines that “Those airports including NRT, KIX, PEK, PVG and PUS are located in Tokyo, Osaka, Beijing, Shanghai and Busan.” as the next sentence is NotNext when one sentence is “Several airports are connected by Korean Air (KE), United (UA), Air Berlin (AB), Delta Airlines (DL), and Cathay Pacific Airways (CX).”

In one embodiment according to the present disclosure, NSP may be used as a method of identifying whether or not there is a precedence relationship between sentences, and is not limited to a technique used in a natural language processing process of a language processing model. Exemplarily, the first model of the present disclosure is not limited to tokens of a language model containing components of a sentence, word order, context, relationship between words, etc., and species of amino acids, electrochemical and biochemical characteristics, relationships between neighboring amino acid species, and three-dimensional compound structures. This means that the machine learning method used by the tokens of the language model can be borrowed.

In one embodiment, the first model 4300 through the NSP task 450 using the above process may determine that CDR3α having the CASSATGSQNTLYFGPG amino acid sequence may be followed by CDR3β having the amino acid sequence CASTDTSQNTLYFGAG (IsNext). For example, the first model 4300 may receive the CDR3α and CDR3β data as described above and output a prediction result that CDR3α having the CASSATGSQNTLYFGPG amino acid sequence and CDR3β having the CASTDTSQNTLYFGAG amino acid sequence are combined with each other.

In a further embodiment, the first model 4300 through the NSP task 450 using the above process can determine that CDR3α having the CASSIRSSYEQYFEG amino acid sequence cannot follow CDR3β having the CASTDTSQNTLYFGAG amino acid sequence (NotNext). For example, the first model 4300 receives the CDR3α and CDR3β data as described above and outputs a prediction result that CDR3α having the CASSIRSSYEQYFEG amino acid sequence and CDR3β having the CASTDTSQNTLYFGAG amino acid sequence cannot be combined with each other.

In such an embodiment, the computing device 100 generates information of the pMHC antigen-specific TCR through the output of various CDR3 sequences, and more effectively in silico , in vivo or in vitro Combinable TCR α and β chains (eg, CDR3α and CDR3β) amino acid sequence information can be obtained.

5 exemplarily illustrates a method of generating data including information of CDR3α of TCR or CDR3β of TCR according to an embodiment of the present disclosure.

In an embodiment according to the present disclosure, in order to acquire raw data of the first data 41a and the second data 41b input to the first model 4300, a CDR3α and a result of CDR3β corresponding to the corresponding CDR3α are generated, or an external model different from the first model (e.g., the second model 4020), a third Model 4030) can be used. In an additional embodiment, information on CDR3α and CDR3β, which may be generated by external models such as the second model 4020 and the third model 4030, may be obtained by the predictive model 500 as shown in FIG. 5.

In one embodiment, FIG. 5 shows a peptide obtained by preprocessing raw data 511 including peptides constituting pMHC and amino acid sequences of MHC, a first input data set 501 composed of data related to MHC, and a prediction model 500 including an encoder driven using the first input data set 501 and second output data 55a (e.g., CDR3α combined with input or output CDR3β) The amino acid sequence of) is shown. As an example, the model shown in FIG. 5 may correspond to the predictive model 500 .

In a further embodiment, the predictive model may include a decoder driven using a second input data set 502 comprising data related to CDR3α and data related to CDR3β. In one embodiment, the CDR3α sequence prediction result 55a of the prediction model including the encoder and the decoder may be output before the CDR3β result, eg, CVRYLCAIENTF[eos].

In a further embodiment, the EOS token of CVRYLCAIENTF[eos] as an exemplary output result may include information of CDR3, peptides constituting pMHC, and information of MHC. In addition, as shown in FIG. 5, the second input data set 502 including CDR3α and CDR3β input to the decoder is preprocessed to include [EOS] tokens between the CDR3α and CDR3β data. In a further embodiment, these EOS tokens may be the subject of prediction model 500 training. Specifically, CDR3 information of the EOS token, peptides constituting pMHC, and MHC information may be updated through a learning and/or inference process.

According to an embodiment of the present disclosure, the computing device 100 may output a CDR3α/CDR3β prediction result 550, which is a combination of TCR α chain and β chain, through an artificial intelligence-based prediction model 500.

According to an embodiment of the present disclosure, the CDR3α / CDR3β prediction result, which is a combination of the TCR α chain and β chain inferred by the prediction model 500, is an embedding layer 510, a first set of hidden layers 520, and a second set of hidden layers 530 and projection layers 540.

In one embodiment, the computing device 100 is input to the embedding layer 510, the first input data set 501 based on the raw data 511 of pMHC and the second input data set 502 based on the raw data of CDR3 are pre-processed data.

In one embodiment, the computing device 100 may process data input from the embedding layer 510 to the first set of hidden layers 520 . As a further example, as shown in FIG. 5 , in the first set of hidden layers 520 of the predictive model 500, the sub hidden layers receiving the first input data 501 and the sub hidden layers receiving the data 502 for CDR3 may be residually connected. As a further example, in the second set of hidden layers 530 of the predictive model, sub hidden layers receiving the first input data 501 and sub hidden layers receiving the data 502 for CDR3 may be residually connected to each other.

According to one embodiment of the present disclosure, a predictive model may include an encoder and a decoder. In one embodiment, the first input data set 501 including pMHC data is input to the encoder, the second input data set 502 including CDR3α data and CDR3β data is input to the decoder, and the CDR3α data and CDR3β data.

In a further embodiment, the second input data set 502 related to CDR3 input to the decoder may include an EOS token between data related to CDR3α and data related to CDR3β. In a further embodiment, the second input data set 502 associated with CDR3 input to the decoder may include an EOS token at the end of data associated with CDR3β.

In one embodiment according to the present disclosure, the first EOS token present between CDR3α data and CDR3β data included in the second input data set related to CDR3 input to the decoder may include information on pMHC and CDR3α. In another embodiment, the second EOS token present at the end of CDR3β data included in the second input data set 502 related to CDR3 input to the decoder may include a second EOS token including information on pMHC and CDR3β.

In a further embodiment, the second EOS token including information on pMHC and CDR3β may be used to obtain second output data 55a including information on the amino acid sequence of CDR3α corresponding to the pMHC and/or CDR3β.

In one embodiment, the predictive model converts the V vector 541 used for processing in the projection layer 540 into a vector consisting of probability values between 0 and 1 by passing, for example, the last hidden state of the LSTM through the Dense Layer. It can be used to calculate the output probability of amino acid identifiers or EOS tokens. As an example, the predictive model may determine one class (e.g., amino acids R, H, K, D, E, S, T, N, Q, C, U, G, P, A, I, L, M, F, W, Y, V or [eos]) after applying the V vector 541 in the projection layer 540. In a further example, when the prediction model predicts that the 8th amino acid of the second output data 55a is Y (Tyrosine) with a probability of 0.50, W (Tryptophan) with a probability of 0.35, and F (Phenylalanine) with a probability of 0.15 for CDR3α to which the V vector 541 is applied to the input data in the projection layer 540, the 8th amino acid of the CDR3α is Y can create

In a further example, when the prediction model 500 predicts that the probability that the 9th amino acid of the second output data 55a is S (Serine) is 0.61, the probability that it is N (Asparagine) is 0.20, and the probability that it is T (Threonine) is 0.19 for CDR3α to which the V vector 541 is applied to the input data in the projection layer 540, the 9th amino acid of the CDR3α is predicted to be T. can create

In one embodiment, the predictive model 500 may output data at positions corresponding to positions in the sequence of amino acids constituting CDR3 included in the second input data set 502 . In a further embodiment, when an amino acid is included in the first position of the second input data set 502, the prediction model 500 may output, as a prediction result, an amino acid derived as a prediction result of a second rank, which is a lower order, instead of the EOS token, which is the first rank prediction, if an EOS token is derived as a first rank prediction result at a second position corresponding to the first position in the output data.

For example, when an amino acid is included in a first position (eg, N-terminal 4th) of the second input data set 502, the predictive model 500 excludes [eos] at a second position (eg, N-terminal 4th) corresponding to the first position in the output data as a prediction result of the first rank (eg, probability 0.52), and then excludes it and second rank (eg, probability 0.4 Amino acids (eg, Q: glutamine) derived as a result of the prediction of 8) may be used as output data. Through this, the computing device 100 can prevent an end-of-sequence result from being generated prior to the length of the biologically valid CDR3, thereby generating more accurate TCR information.

In one embodiment, the prediction model 500, when an EOS token is derived as a first-rank prediction result at a position before exceeding a predetermined critical length (e.g., the length of 5 amino acids), the first-rank prediction EOS token, instead of an amino acid derived as a second-rank prediction result that is more subordinate to the prediction result.

As an example, the predictive model 500 may exclude [eos] at the N-terminal 5th position of CDR3β before exceeding a predetermined threshold length 5 as a prediction result of the first rank (eg, probability 0.54), and exclude it and use an amino acid (eg, I: isoleucine) derived as a prediction result of the second rank (eg, probability 0.41) subordinate to [eos] as output data. Through this, the computing device 100 can prevent an end-of-sequence result from being generated prior to the length of the biologically valid CDR3, thereby generating more accurate TCR information.

In one embodiment, MHC and peptides capable of binding or each of MHC and peptides may be obtained through a public database (eg, IEDB, HLA atlas Ligand, SysteMHC Altas). In an additional embodiment, bindable MHC and peptides (ie, pMHC) may be determined by a separate artificial intelligence or rule-based model that takes MHC and peptides as inputs and outputs whether bindable or not.

In an additional embodiment, information on a TCR that can be combined with pMHC of the first input data set (e.g., length information, V/J type, etc.) takes the pMHC as an input and generates TCR information corresponding to the pMHC. It may be determined by a separate artificial intelligence or rule-based model.

In an additional embodiment, accurate performance evaluation is possible through the CDR3 prediction results of the TCRs of the sub-prediction models referred to for each epoch performed multiple times in the plurality of sub-prediction models included in the prediction model 500, and learning and/or inference efficiency and/or accuracy of the predictive model can be improved.

According to an embodiment of the present disclosure, the computing device 100 may generate a CDR3β candidate list included in the first output data generated using the first EOS token. In an additional embodiment, the computing device 100 may generate a CDR3α candidate list included in the second output data generated using the second EOS token. In a further embodiment, the CDR3 set candidate list including the results generated by the model (e.g., the first model 4300 and the prediction model 500) outputting a combination of CDR3α and CDR3β combined in the present disclosure may refer to a list including candidates of TCR constituent CDR3 (CDR3α and CDR3β) that can be used for TCR-T, for example. For example, a candidate list of CDR3α or CDR3β for subject 1 may be generated, and a TCR including a constitutive CDR3 selected from the candidate list may be used for treatment of subject 1.

In an additional embodiment, specifically, the computing device 100 combines CDR3α and CDR3β in each of the CDR3α and CDR3β candidate lists included in the first output data and the second output data to derive a complementarity determining region of the TCR corresponding to a specific peptide and a specific MHC. Accordingly, when a target antigen (eg, an antigen specifically found in a tumor of a solid cancer) is identified, the prediction model can derive the most effective TCR corresponding to the target antigen.

Through the output data 55a as illustrated in FIG. 5, the computing device 100 can generate antigen-specific TCR information by outputting information on the amino acid sequence of CDR3α, which is the TCRα-chain, which is combined with the TCRβ-chain in silico , in vivo , or in vitro .

In one embodiment, when peptides included in the first input data set 501 and MHC form a bond, the prediction model 500 receiving the first input data set 501 including the peptides and the MHC forming the bond forms information (55a) including the amino acid sequence of the CDR3α that can be combined with the CDR3β when the information of CDR3α and CDR3β to be combined is included in the second input data set 502 ), it can be seen that the prediction purpose of the present disclosure has been achieved. In another example, when the peptides and MHCs included in the first input data set 501 form a bond, the prediction model 500 receiving the first input data set 501 including the peptides and the MHCs forming the bond forms an input. It can be seen that the prediction purpose of the present disclosure has been achieved only when information 55a having a similar amino acid sequence of at least 5 amino acids among amino acids is generated.

6 illustratively illustrates first positive data, first negative data, and second negative data including information related to a combination of CDR3α of TCR and CDR3β of TCR, according to an embodiment of the present disclosure.

In one embodiment of the present disclosure, as shown in FIG. 6 , the computing device 100 randomly combines a first dataset corresponding to CDR3α or a second dataset corresponding to CDR3β of TCR to generate a first negative dataset.

In one embodiment, the negative data may include combinations of unknown CDR3α and CDR3β. Specifically, for example, assuming that 46 CDR3α amino acid sequences and 61 CDR3β amino acid sequences are stored in a storage medium, the computing device 100 databases random combinations of the 46 CDR3α amino acid sequences and 61 CDR3β amino acid sequences (e.g., 46 × 61 = 2806) as possible unknown combinations of the TCR α chain and the TCR β chain (first 1 voice dataset).

In one embodiment, the computing device 100 may identify CDR3α and CDR3β identified as binding to each other in the randomly combined first voice dataset. In a further embodiment, whether the CDR3α and the CDR3β bind to each other can be confirmed through molecular biological experiments. For example, as a result of experimentally analyzing the amino acid sequence of the TCR obtained from the subject, the computing device 100 may identify a pair (eg, combination) of CDR3α and CDR3β that binds to each other. In another embodiment, whether the CDR3α and CDR3β bind to each other can be confirmed through public DB data.

In an additional embodiment, the computing device 100 may generate a second audio dataset by excluding CDR3α and CDR3β identified as binding to each other through the molecular biological experiment or a public DB from the first audio dataset.

In one embodiment, the computing device 100 may include a combination of CDR3α and CDR3β identified as binding to each other through a molecular biological experiment or through a public DB in the first positive dataset. Specifically, the combination of CDR3α and CDR3β included in the first positive dataset may be composed of combinations of TCR α chain and TCR β chain that are known to bind to each other. For example, combinations of CDR3α and CDR3β known on public DBs may be included in the first positive dataset. As another example, combinations of CDR3α and CDR3β found experimentally in TCRs obtained from individual subjects may be included in the first positive dataset.

In an additional embodiment, when it is confirmed that CDR3α and CDR3β with unknown binding, classified as negative data, bind to each other, the computing device 100 may transfer the combination of the corresponding CDR3α and CDR3β to the first positive dataset. In an additional embodiment, if it is confirmed that CDR3α and CDR3β with unknown binding, classified as voice data, do not bind to each other, the computing device 100 may store the combination of the corresponding CDR3α and CDR3β as a second voice dataset.

In an embodiment according to the present disclosure, the first model 4300 may be trained based on the first positive dataset, the first negative dataset, and/or the second negative dataset generated in the above method. In one embodiment, the first model 4300 trained using the combination of CDR3α and CDR3β included in the first positive dataset as training data may learn a set of CDR3α and CDR3β that is a precedent relationship. In a further embodiment, the first model 4300 trained using the combination of CDR3α and CDR3β included in the second speech dataset as training data may learn a set of CDR3α and CDR3β that does not have a precedence relationship.

In a further embodiment, when the predicted result of the first model 4300 for the combination of CDR3α and CDR3β included in the first negative dataset is consistent with the result of molecular biology experiment, the computing device 100 may transfer the combination of CDR3α and CDR3β to the second negative dataset or the first positive dataset. For example, if the combination of CDR3α and CDR3β predicted by the first model 4300 in a precedent relationship is identified as binding through an actual molecular biological experiment, the computing device 100 may transfer the combination of CDR3α and CDR3β predicted in the precedence relationship to the first positive dataset. As another example, if a combination of CDR3α and CDR3β predicted by the first model 4300 in a precedent relationship is identified as not binding through a molecular biological experiment, the computing device 100 may transfer the combination of CDR3α and CDR3β incorrectly predicted to correspond to the precedence relationship to the second audio dataset.

In one embodiment according to the present disclosure, the negative dataset or the positive dataset updated in the above manner by adding the latest known information of the supervised / semi-supervised learning data can improve the learning efficiency and / or accuracy of the predictive model for the combination of the TCR α chain and the TCR β chain.

7 illustratively illustrates storing a combination of CDR3α and CDR3β in a CDR set candidate list, according to an embodiment of the present disclosure.

In one embodiment, the computing device 100 may obtain the list 702 of CDR3αs generated through the second model 4020 . In one embodiment, the computing device 100 may obtain the list 703 of CDR3β generated through the third model 4030 .

In an additional embodiment, the CDR3α/CDR3β combination prediction model 710 (eg, the first model 4300) may predict whether the input CDR3α and CDR3β correspond to a precedence relationship. For example, as a result of determining whether the CDR3α and CDR3β input to the CDR3α / CDR3β combination prediction model 710 correspond to a precedence relationship, if it is determined that CDR3α and CDR3β do not correspond to a precedence relationship, the combination of CDR3α and CDR3β may be output as not binding. CDR3α and CDR3β combinations that do not have a precedence relationship can be classified as NotNext output data 724 as shown in FIG. 7 .

As another example, as a result of determining whether the CDR3α and CDR3β input to the CDR3α / CDR3β combination prediction model 710 correspond to a precedence relationship, when it is determined that CDR3α and CDR3β correspond to a precedence relationship, the combination of CDR3α and CDR3β may be output as being combined. The combination of CDR3α and CDR3β, which is in a precedent relationship, can be classified as IsNext output data 722 as shown in FIG. 7 .

In one embodiment, combinations of CDR3α and CDR3β included in the IsNext output data 722 predicted by the CDR3α/CDR3β combination prediction model 710 may be stored 720 in a CDR set candidate list. In a further embodiment, combinations of a plurality of CDR3α and CDR3β finally stored in the CDR set candidate list (720) may be used as information on TCRs expressed in TCRs and/or TCR-Ts. In a further embodiment, combinations of a plurality of CDR3α and CDR3β finally stored in the CDR set candidate list (720) may be used as information of a target antigen-specific TCR.

In one embodiment, the generated results stored in the CDR set candidate list 720 from a model (e.g., the first model 4300, the predictive model 500) outputting a combination of CDR3α and CDR3β combined in the present disclosure may mean a list including candidates of TCR constituent CDR3 (CDR3α and CDR3β) that can be used for TCR-T, for example. For example, a CDR3α or CDR3β candidate list for subject 1 may be generated, and a TCR including CDR3α and CDR3β selected from the candidate list may be used for treatment of subject 1.

In an additional embodiment, the computing device 100 may derive a complementarity determining region of a TCR corresponding to a specific peptide and a specific MHC from among the combinations of CDR3α and CDR3β stored in the CDR set candidate list (720). By storing (720) only combinations of CDR3α and CDR3β corresponding to a precedence relationship in the CDR set candidate list, the computing device 100 can select combinations of CDR3α and CDR3β suitable for the purpose of treating a subject (e.g., a lesion, a test subject, etc.) from the CDR set candidate list. For example, when a target antigen (eg, an antigen specifically found in a tumor of solid cancer) is identified, the computing device 100 may derive a TCR corresponding to the target antigen and including the most effective CDR3α and CDR3β.

8 is a schematic diagram of a computing environment according to one embodiment of the present disclosure.

A component, module or unit in this disclosure includes routines, procedures, programs, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Further, those skilled in the art will fully appreciate that the methods presented in this disclosure may be practiced with other computer system configurations, including single-processor or multiprocessor computing devices, minicomputers, mainframe computers, as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, and the like, each of which may be operative in connection with one or more associated devices.

Embodiments described in this disclosure may also be practiced in distributed computing environments where certain tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

A computing device typically includes a variety of computer readable media. Computer readable media can be any medium that can be accessed by a computer, and such computer readable media includes volatile and nonvolatile media, transitory and non-transitory media, removable and non-removable media. By way of example, and not limitation, computer readable media may include computer readable storage media and computer readable transmission media.

Computer readable storage media includes volatile and nonvolatile media, transitory and non-transitory media, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer readable storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage device, magnetic cassette, magnetic tape, magnetic disk storage device or other magnetic storage device, or any other medium that can be accessed by a computer and used to store desired information.

A computer readable transmission medium typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery medium. The term modulated data signal means a signal that has one or more of its characteristics set or changed so as to encode information within the signal. By way of example, and not limitation, computer readable transmission media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Combinations of any of the above are also intended to be included within the scope of computer readable transmission media.

An exemplary environment 2000 implementing various aspects of the present invention is shown comprising a computer 2002, which includes a processing unit 2004, a system memory 2006 and a system bus 2008. Computer 200 herein may be used interchangeably with a computing device. System bus 2008 couples system components, including but not limited to system memory 2006, to processing unit 2004. Processing unit 2004 may be any of a variety of commercially available processors. Dual processor and other multiprocessor architectures may also be used as the processing unit 2004.

System bus 2008 may be any of several types of bus structures that may additionally be interconnected to a memory bus, a peripheral bus, and a local bus using any of a variety of commercial bus architectures. System memory 2006 includes read only memory (ROM) 2010 and random access memory (RAM) 2012 . A basic input/output system (BIOS) is stored in non-volatile memory 2010, such as ROM, EPROM, or EEPROM, and includes basic routines that help transfer information between components within the computer 2002, such as during startup. RAM 2012 may also include high-speed RAM, such as static RAM, for caching data.

Computer 2002 also includes internal hard disk drives (HDDs) 2014 (e.g., EIDE, SATA), magnetic floppy disk drives (FDDs) 2016 (e.g., for reading from or writing to removable diskettes 2018), SSDs, and optical disk drives 2020 (e.g., for reading CD-ROM disks 2022 or reading from or writing to other high capacity optical media such as DVDs). intended to do). Hard disk drive 2014, magnetic disk drive 2016, and optical disk drive 2020 may be connected to system bus 2008 by hard disk drive interface 2024, magnetic disk drive interface 2026, and optical drive interface 2028, respectively. The interface 2024 for external drive implementation includes, for example, at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.

These drives and their associated computer readable media provide non-volatile storage of data, data structures, computer executable instructions, and the like. In the case of computer 2002, drives and media correspond to storing any data in a suitable digital format. Although the description of computer-readable storage media above refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art will appreciate that other types of computer-readable storage media, such as zip drives, magnetic cassettes, flash memory cards, cartridges, and the like, may also be used in the exemplary operating environment and any such media may include computer-executable instructions for performing the methods of the present invention.

A number of program modules may be stored on the drive and RAM 2012, including an operating system 2030, one or more application programs 2032, other program modules 2034, and program data 2036. All or portions of the operating system, applications, modules and/or data may also be cached in RAM 2012. It will be appreciated that the present invention may be implemented in a variety of commercially available operating systems or combinations of operating systems.

A user may enter commands and information into the computer 2002 through one or more wired/wireless input devices, such as a keyboard 2038 and a pointing device such as a mouse 2040. Other input devices (not shown) may include a microphone, IR remote control, joystick, game pad, stylus pen, touch screen, and the like. These and other input devices are often connected to processing unit 2004 through input device interface 2042, which is connected to system bus 2008, but may be connected by other interfaces such as parallel ports, IEEE 1394 serial ports, game ports, USB ports, IR interfaces, and the like.

A monitor 2044 or other type of display device is also connected to the system bus 2008 through an interface such as a video adapter 2046. In addition to the monitor 2044, computers typically include other peripheral output devices (not shown) such as speakers, printers, and the like.

Computer 2002 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) 2048 via wired and/or wireless communications. Remote computer(s) 2048 may be a workstation, server computer, router, personal computer, portable computer, microprocessor-based entertainment device, peer device, or other common network node, and generally includes many or all of the components described for computer 2002, although for simplicity, only memory storage device 2050 is shown. The logical connections shown include wired/wireless connections to a local area network (LAN) 2052 and/or a larger network, such as a wide area network (WAN) 2054 . Such LAN and WAN networking environments are common in offices and corporations and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to worldwide computer networks, such as the Internet.

When used in a LAN networking environment, computer 2002 is connected to local network 2052 through wired and/or wireless communication network interfaces or adapters 2056. Adapter 2056 may facilitate wired or wireless communications to LAN 2052, which also includes a wireless access point installed therein to communicate with wireless adapter 2056. When used in a WAN networking environment, computer 2002 may include a modem 2058, be coupled to a communication server on WAN 2054, or have other means of establishing communications over WAN 2054, such as over the Internet. Modem 2058, which can be internal or external and wired or wireless, is connected to system bus 2008 through serial port interface 2042. In a networked environment, program modules described for computer 2002, or portions thereof, may be stored in remote memory/storage device 2050. It will be appreciated that the network connections shown are exemplary and other means of establishing a communication link between computers may be used.

Computer 1602 is operative to communicate with any wireless device or entity that is deployed and operating in wireless communication, e.g., printers, scanners, desktop and/or portable computers, portable data assistants (PDAs), communications satellites, any equipment or location associated with wireless detectable tags, and telephones. This includes at least Wi-Fi and Bluetooth wireless technologies. Thus, the communication may be a predefined structure as in conventional networks or simply an ad hoc communication between at least two devices.

It is to be understood that the specific order or hierarchy of steps in the processes presented is an example of example approaches. Based upon design priorities, it is to be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of the present disclosure. The method claims of this disclosure present elements of the various steps in a sample order, but are not meant to be limited to the specific order or hierarchy presented.

Claims (11)

As a method for generating a prediction result using artificial intelligence technology performed by a computing device,
Acquiring first data corresponding to CDR3α of T cell receptor (TCR) and second data corresponding to CDR3β of TCR, wherein the first data includes an amino acid sequence corresponding to CDR3α, and the second data includes an amino acid sequence corresponding to CDR3β;
Using an artificial intelligence-based first model, receiving the first data and the second data, and determining whether the first data including the CDR3α and the second data including the CDR3β correspond to a precedence relationship; and
storing a combination of the first data and the second data in a CDR set candidate list when it is output that the first data and the second data correspond to the precedence relationship;
Including,
The first model is:
generating a first negative dataset by randomly combining a first dataset corresponding to CDR3α or a second dataset corresponding to CDR3β of TCR;
identifying CDR3α and CDR3β identified as binding to each other in the randomly combined first voice dataset;
generating a second audio dataset by excluding the combination of CDR3α and CDR3β identified as binding from the first audio dataset; and
Incorporating the combination of CDR3α and CDR3β identified as binding into a first positive dataset;
Corresponding to the pre-learned model based on
method. According to claim 1,
The first data corresponding to CDR3α of the TCR is generated by a second model different from the first model, and the second model is an artificial intelligence-based or rule-based model pretrained to receive a peptide-MHC conjugate (pMHC) and output the first data including CDR3α corresponding to the pMHC, and
The second data corresponding to CDR3β of the TCR is generated by a third model different from the first model, and the third model is an artificial intelligence-based or rule-based model that has been pretrained to output the second data including the CDR3β corresponding to the pMHC as an input,
method. According to claim 1,
The first data corresponding to CDR3α of the TCR and the second data corresponding to CDR3β of the TCR are generated by a fourth model different from the first model,
The fourth model includes an encoder and a decoder, and
In the fourth model, a first input dataset including amino acid sequences corresponding to peptides and MHC is input to the encoder, amino acid sequences corresponding to CDR3α and CDR3β related to the peptide and MHC are input to the decoder, and the first data and a prediction result including the first data is output from the decoder.
method. delete According to claim 1,
Combinations of CDR3α and CDR3β included in the second voice dataset are labeled as correct answer data that do not correspond to a precedence relationship in the learning process of the first model, and
The combinations of CDR3α and CDR3β included in the first positive dataset are labeled as correct answer data that correspond to a precedence relationship in the learning process of the first model,
method. According to claim 1,
The first model is:
Among the combinations of CDR3α and CDR3β included in the second negative dataset, identifying at least one similar combination with a CDR3α and CDR3β combination included in the first positive dataset and a similarity equal to or greater than a predetermined threshold; and
including the identified similar combinations in the first positive dataset;
In addition to being pre-learned based on,
method. According to claim 1,
The first model includes a language model,
method. According to claim 7,
The language model is
Recurrent Neural Network (RNN), Long Short Term Memory (LSTM) network, Bidirectional Long Short Term Memory (BiLSTM) network, Diffusion model, Generative Pre-trained Transformer (GPT), Bidirectional Encoder Representations from Transformers (BERT), spanBERT, Gated Recurrent Unit (GRU), or Bidirectional Gated Recurrent Unit (BiGRU),
method. According to claim 1,
The CDR set candidate list refers to a combination of TCR α chain and β included in TCR capable of binding to pMHC, and
The combination of the first data and the second data stored in the CDR set candidate list is used for TCR-T generation,
method. A computer program stored on a computer-readable storage medium, which, when executed by a computing device, causes the computing device to perform operations for generating a prediction result using artificial intelligence technology, the operations comprising:
Obtaining first data corresponding to CDR3α of TCR and second data corresponding to CDR3β of TCR, wherein the first data includes an amino acid sequence corresponding to CDR3α, and the second data includes an amino acid sequence corresponding to CDR3β;
Using an artificial intelligence-based first model, receiving the first data and the second data, and determining whether the first data including the CDR3α and the second data including the CDR3β correspond to a precedence relationship; and
storing a combination of the first data and the second data in a CDR set candidate list when a result indicating that the first data and the second data correspond to the precedence relationship is output;
Including,
The first model is:
generating a first negative dataset by randomly combining a first dataset corresponding to CDR3α or a second dataset corresponding to CDR3β of TCR;
identifying CDR3α and CDR3β identified as binding to each other in the randomly combined first voice dataset;
generating a second audio dataset by excluding the combination of CDR3α and CDR3β identified as binding from the first audio dataset; and
Incorporating the combination of CDR3α and CDR3β identified as binding into a first positive dataset;
Corresponding to the pre-learned model based on
A computer program stored on a computer readable storage medium. As a computing device,
at least one processor; and
Memory;
Including,
The at least one processor is:
Obtaining first data corresponding to CDR3α of TCR and second data corresponding to CDR3β of TCR, wherein the first data includes an amino acid sequence corresponding to CDR3α, and the second data includes an amino acid sequence corresponding to CDR3β;
Using an artificial intelligence-based first model, receiving the first data and the second data, and determining whether the first data including the CDR3α and the second data including the CDR3β correspond to a precedence relationship; and
storing a combination of the first data and the second data in a CDR set candidate list when a result indicating that the first data and the second data correspond to the precedence relationship is output;
to perform,
The first model is:
generating a first negative dataset by randomly combining the first dataset corresponding to CDR3α or the second dataset corresponding to CDR3β of TCR;
In the randomly combined first voice dataset, CDR3α and CDR3β identified as binding to each other are identified;
generating a second audio dataset by excluding the combination of CDR3α and CDR3β identified as binding from the first audio dataset; and
incorporating the combination of CDR3α and CDR3β identified as binding into a first positive dataset;
Corresponding to the pre-learned model based on
computing device.