KR102278727B1 - Method for predicting neoantigen using a peptide sequence and hla class ii allele sequence and computer program - Google Patents

Abstract

According to the present invention, disclosed is a method for predicting a neoantigen by using a peptide sequences and an HLA class II allele sequence. According to this, not only a coupling force among a peptide sequence included in a cancer tissue, HLA class II allele alpha, a beta complex, and alpha and beta single sequences, but also immunity of the peptide sequence are measured. A neoantigen in a cancer tissue can be determined based on the measured immunity.

Description

Method and computer program for predicting neoantigens using target cancer tissue and cell-free DNA-derived peptide sequences and HLA class II allele sequences {METHOD FOR PREDICTING NEOANTIGEN USING A PEPTIDE SEQUENCE AND HLA CLASS II ALLELE SEQUENCE AND COMPUTER PROGRAM}

Embodiments of the present disclosure relate to methods and computer programs for predicting neoantigens using peptide sequences and HLA class II allele sequences derived from target cancer tissue and cell-free DNA.

Due to the development of new anticancer drugs, first-generation chemotherapy and second-generation targeted anti-cancer drugs, and recently, third-generation immuno-oncology drugs are in the spotlight. In particular, in the case of the third-generation immuno-oncology drug, unlike the previous anti-cancer drugs, it is a treatment strategy that utilizes the patient's own immune system, so side effects are significantly lower. However, despite these advantages, there is a limit in which a treatment strategy using immunotherapy can be established only for patients who show the expression of marker genes such as PD-L1 and microsatellite instability (MSI-H). Due to these limitations, it is necessary to establish a strategy to treat patients who have difficulty in administering existing anticancer drugs, and one of the alternatives is a cancer vaccine using a new antigen.

In addition, these cancer vaccines are not limited to cancer tissues and can be grafted onto cell-free DNA. Cell-free DNA is present in body fluids such as blood, plasma, saliva, lymph, cerebrospinal fluid, synovial fluid, cystic fluid, ascites, pleural effusion, amniotic fluid, chorionic villi sample, placental sample, tracheal lavage fluid, interstitial fluid, or ocular fluid. Exosomes present in the blood also contain small amounts of protein, RNA, and DNA, so it is possible to predict neoantigens for cancer vaccines from DNA in the exosomes.

Each patient's cancer tissue contains mutations that are not found in normal tissues, and the key strategy for cancer vaccines is to use peptides derived from these mutations as neoantigens so that the patient's immune system can recognize and attack the new antigens. to be. First, a stable binding between the mutant-derived peptide and the patient-specific HLA allele is a prerequisite for this process. In particular, in the case of HLA class II alleles, more diverse genes (DM, DO, DP, DQ, DR) than HLA class I alleles (A, B, C) exist, and because they form a complex of alpha and beta, the binding force is predicted. There are some more difficult things to do. The second is to check whether the peptide derived from the mutation has immunogenicity that stimulates the patient's immune system well. In particular, in order to reflect immunogenicity as much as possible, it is necessary to simulate all stages of immunogenicity and extract key features. However, in this process, missing or missing steps may occur, which may act as a limiting point in future immunogenicity prediction. have.

Accordingly, the present technology tried to implement a strategy for preventing omission of immune processes and extracting key features based on the combination data of the known immunogenic peptide and MHC (HLA in the case of humans) class II sequence. In addition, i)Immunogenicity score ii) derived from each model after modeling the HLA class II alpha, beta complex and the peptide, the HLA class II alpha or beta sequence and the peptide, and the immunogenicity of the peptide sequence itself. Based on the avidity score between the alpha and beta complex and the peptide iii) the avidity score between the alpha or beta sequence and the peptide, iv) the neoantigen prediction score between the alpha and beta complex and the peptide v) the neoantigen prediction score between the alpha or beta sequence and the peptide We tried to finally derive neoantigens applicable to patients through various ensemble models of machine learning. The above-mentioned background art is technical information possessed by the inventor for derivation of the present invention or acquired in the process of derivation of the present invention, and cannot necessarily be said to be a known technique disclosed to the general public prior to filing of the present invention.

The present invention is in accordance with the above-mentioned necessity, wherein a mutation-derived peptide sequence present in a patient's cancer tissue binds to a patient-specific HLA class II alpha, beta complex and a single chain, while undergoing a series of immune processes to finally achieve immunogenicity The purpose of this study is to predict the expression of , and based on this, determine a neoantigen that can be used for a cancer vaccine customized for cancer patients.

The method for predicting a neoantigen using the peptide sequence and the allele sequence of the HLA class II alpha and beta chain according to the embodiments of the present invention includes a peptide sequence derived from a target cancer tissue or cell-free DNA and an HLA class II alpha, beta chain. receiving an allele sequence of obtaining T cell activity data from the peptide sequence, inputting the T cell activity data into an immunity prediction model, and outputting a first predicted value for predicting immunity of the peptide sequence; Binding data is obtained from the allele sequence of the HLA class II alpha, beta complex, and the binding data is input to a binding prediction model to bind the peptide sequence and the allele sequence of the HLA class II alpha, beta chain complex outputting a second prediction value for predicting Binding data is obtained from the allele sequence of the HLA class II alpha or beta chain, and the binding data is input to a binding prediction model to predict the binding property of the peptide sequence and the HLA class II alpha or beta chain allele sequence outputting a third predicted value; and outputting a fourth predicted value for predicting a neoantigen for the target cell using the T cell activity data and the first and second predicted values; and outputting a fifth predicted value for predicting neoantigen information on the target cell using the T cell activity data and the first and third predicted values.

The immune predictive model (first predictive value), the HLA class II alpha, beta complex and peptide binding predictive model (second predictive value), the HLA class II alpha or beta complex and peptide binding predictive model (third predictive value), At least one of the HLA class II alpha and beta complex and peptide neoantigen prediction model (fourth predictive value) and the HLA class II alpha or beta complex and peptide neoantigen predictive model (fifth predictive value) is selected from a plurality of target cancer tissues. It can be trained by an ensemble algorithm of machine learning based on a training data set comprising the present peptide sequences and HLA class II allele sequences.

The subject cancer tissue may comprise cells engineered to express a single MHC class I or class II allele.

The target cancer tissue may include human cells obtained from or derived from a plurality of patients.

The target cancer tissue may include fresh or frozen tumor cells obtained from a plurality of patients.

The target cancer tissue may include fresh or frozen tissue cells obtained from a plurality of patients.

The target cancer tissue may include blood free DNA obtained from a plurality of patients.

The target cancer tissue may include DNA in exosomes obtained from a plurality of patients.

The target cancer tissue may comprise a peptide identified using T-cell analysis.

The training data set includes data related to proteomic sequences related to the target cancer tissue, data related to MHC class II alpha and beta peptide sequences related to the target cancer tissue, and peptides related to the target cancer tissue and HLA class II alpha, beta complexes. Alternatively, it may include at least one of binding data between alpha and beta alleles, data related to the transcriptome related to the target cancer tissue, and data related to the genome related to the target cancer tissue.

The immunity prediction model may be a model trained by inputting T cell activity data from the peptide sequence as an input and outputting the immunity of the peptide sequence.

The binding prediction model is the HLA class II alpha, beta complex or alpha, beta single allele sequence and binding data from the peptide sequence as inputs, the peptide sequence and the HLA class II alpha, beta complex or alpha, beta single allele It may be a model trained as an output of the binding properties of a gene sequence.

The neoantigen prediction model is derived from the peptide sequence and T cell activity data from the HLA class II alpha, beta complex or alpha, beta single allele sequence and the HLA class II allele alpha, beta complex or alpha, beta single sequence and the peptide sequence. It may be a trained model with the binding data of the as input and the neoantigen predicted value between the peptide sequence and the HLA class II allele sequence as output.

A computer program according to an embodiment of the present invention may be stored in a medium to execute any one of the methods according to an embodiment of the present invention using a computer.

In addition to this, another method for implementing the present invention, another system, and a computer readable recording medium for recording a computer program for executing the method are further provided.

Other aspects, features and advantages other than those described above will become apparent from the following drawings, claims, and detailed description of the invention.

According to one embodiment of the present invention made as described above, the binding force between the peptide sequence and the HLA class II allele alpha, beta complex or alpha, beta single sequence included in the cancer tissue as well as the immunity of the peptide sequence is measured and , it is possible to determine neoantigens in cancer tissues based on the measured immunity.

1 is a block diagram of a neoantigen determination apparatus 100 according to embodiments of the present invention.
2 is an exemplary diagram of type information of major HLA class II contained in cells of Koreans.
3 is an exemplary diagram of a blosum matrix showing the similarity of characteristics between amino acids.
4 is a diagram showing the structure of HLA allele class II.
5 is a training data set for the HLA class II alpha beta chain complex.
6 is a training data set for HLA class II alpha or beta single allele sequences.
7, 8 and 9 are diagrams illustrating examples of implementation of the immunity predicting unit 121' and the binding predicting unit 122' according to embodiments of the present invention.
10 to 12 are diagrams of implementation examples of the neoantigen determination system according to embodiments of the present invention.
13 is a block diagram of the learning server 10 for learning an immunity prediction model, a binding prediction model, an immune resistance prediction model, and the like.

Hereinafter, the configuration and operation of the present invention will be described in detail with reference to the embodiments of the present invention shown in the accompanying drawings.

Since the present invention can apply various transformations and can have various embodiments, specific embodiments are illustrated in the drawings and described in detail in the detailed description. Effects and features of the present invention, and a method of achieving them, will become apparent with reference to the embodiments described below in detail in conjunction with the drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various forms.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, and when described with reference to the drawings, the same or corresponding components are given the same reference numerals, and the overlapping description thereof will be omitted. .

In the following embodiments, terms such as first, second, etc. are used for the purpose of distinguishing one component from another, not in a limiting sense.

In the following examples, the singular expression includes the plural expression unless the context clearly dictates otherwise.

In the following embodiments, terms such as include or have means that the features or components described in the specification are present, and the possibility of adding one or more other features or components is not excluded in advance.

In the drawings, the size of the components may be exaggerated or reduced for convenience of description. For example, since the size and thickness of each component shown in the drawings are arbitrarily indicated for convenience of description, the present invention is not necessarily limited to the illustrated bar.

Where certain embodiments are otherwise feasible, a specific process sequence may be performed different from the described sequence. For example, two processes described in succession may be performed substantially simultaneously, or may be performed in an order opposite to the order described.

Here, the target cancer tissue refers to a tissue to be tested. For example, the target cancer tissue may be a cancer tissue in which an antigen capable of eliciting an immune response is to be detected. Preferably, the target cancer tissue may be a tumor cell or an aggregate of cancer cells.

Here, the mutation means that the sequence of the nucleotide sequences A (adenine), T (thymine), G (guanine), and C (cytosine) of the gene containing the genetic information in each organism is altered to be different from the original genetic information of the species. means all phenomena. These mutations cause small-scale or large-scale structural changes. Small-scale mutations include point mutations in which a single nucleotide sequence is converted, and mutations in which nucleotide sequences are additionally inserted or deleted. Mutations that occur on a large scale and affect structure include gene duplication, gene deletion, chromosomal inversion, interstitial deletion, chromosomal translocation, and loss of heterozygosity.

Mutations are largely divided into germline mutations and somatic mutations according to the type of cell that occurs. Somatic mutation is a gene mutation that occurs in somatic cells, also called somatic mutation or somatic mutation, and may be caused by mutations in genes or chromosomal abnormalities.

Due to the occurrence of such a mutation, the function of the protein produced by the gene may be changed, and a specific function may be lost or activated with another function. Since these changes in protein function cause or accelerate cancer development, these mutations may be directly or indirectly related to cancer development and progression.

As described above, the nucleotide sequence in DNA containing the genetic information of living things consists of A, T, G, and C. When these nucleotide sequences are gathered three in a row, it becomes a code to form one specific amino acid, and these codes are When combined, it can be converted into a single protein. Amino acids are alanine (Ala), cysteine (Cys), aspartic acid (Asp), glutamic acid (Glu), phenylalanine (Phe), glycine (Gly), histidine (His), isoleucine (Ile), lysine (Lys), leucine (Leu), methionine (Met), asparagine (Asn), parolysine (Ply), proline (Pro), glutamine (Gln), arginine (Arg), serine (Ser), threonine (Thr), selenocysteine (Sec) ), valine (Val), tryptophan (Trp), and tyrosine (Tyr).

The peptide may refer to a peptide or polypeptide composed of amino acid sequences. An immune system exists to remove foreign substances that are not derived from genetic information within each species, and in particular, among externally-derived peptides, immunogenic peptides that can induce an immune response exist. Mutations that occur differently from the original genetic information in the course of cancer development also generate these immunogenic peptides, which can bind to HLA class II proteins through a series of processes in the immune system. Furthermore, the immunogenic peptide may have a mutant amino acid sequence, and the length of the amino acid may be 25 or less, but is not limited thereto, and may have various lengths.

A neoantigen refers to a peptide that induces an immune response. That is, the neoantigen may be an immunogenic peptide. Neoantigens may be induced by tumor cell-specific mutations and may be expressed as epitopes of tumor cells. Hereinafter, for clarity of description, the immunogenic peptide is named as a neoantigen.

Here, T cell activity data is data that measures the immune response that occurs when stimulated by binding of a specific peptide sequence to a specific HLA class II allele. Cells detected by immunogenicity measurement experimental methodology including multimer/tetramer and ELISPOT It is obtained from data such as my cytokine expression value and the expression value of immune cell-specific activation markers, and the result is “Positive”, “Positive-High”, “Positive-Low”, “Positive-Intermediate”, “Negative” can be classified.

The neoantigen determining device according to the embodiments of the present invention analyzes the peptide sequence of the target cancer tissue and the HLA class II alpha and beta allele sequence of the patient to generate a specific peptide of the target cancer tissue to be used for the treatment of the target cancer tissue. antigen can be determined. 4 is a schematic diagram showing the characteristics of HLA classes I and II, and in the case of HLA class II, unlike I, it is composed of a complex of alpha and beta chains. Among the peptides contained in the target cancer tissue of a patient having a specific HLA class II allele, a neoantigen suitable as an antigen can be determined. An antibody acting on the determined neoantigen can be searched for and used for treatment of the target cancer tissue of the patient. In particular, the types and distributions of HLA class II, which appear frequently in Koreans, are shown in FIG. 2 .

1 is a block diagram of an apparatus 100 for determining a neoantigen according to embodiments of the present invention.

The neoantigen determination apparatus 100 is an apparatus for determining neoantigens for treating a disease existing in a cancer tissue based on genomic data in the cancer tissue.

The genome data input unit 110 may receive a peptide sequence extracted from cancer tissue and HLA class II alpha and beta allele sequences. The peptide sequence may be for one or more peptides included in cancer tissue. A peptide sequence can be represented as a two-dimensional matrix to include sequences for peptides. HLA class II alpha and beta allele sequences can be expressed as an embedding vector in a specific size through a word embedding technique in which 1 to k amino acid units are set as one word, but is not limited thereto, and as shown in FIG. It can be expressed in various formats, such as vectorizing the similarity between amino acids through a blosum50 or blosum62 matrix. The HLA allele sequence is divided into class I and class II, and as shown in FIG. 4 , the HLA class II allele sequence of Class II may have an alpha, beta chain structure.

The genome data input unit 110 is based on the peptide sequence and the HLA class II allele sequence, the binding data between the peptides and the alpha chain data of the peptides and the HLA class II allele, and the peptides and the HLA class II Binding data between the beta chain data of alleles can be calculated separately.

The genome data input unit 110 may measure T cell activity data for peptides and calculate T cell activity data for peptides of cancer tissue using a table or database in which the measured data is recorded.

Here, the HLA class II allele may be input by dividing the HLA class II allele sequence in units of 1 to 1 kmer and expressing it as a virtual word set regardless of the entire sequence or pseudo sequence. The HLA class II allele can be generated as a training data set including alpha chain data and beta chain data as shown in FIG. 5 . The training data set can be created by separating the alpha chain data and beta chain data with a delimiter. The training data set may be input to the associativity prediction units 1221 and 1222 .

The HLA class II allele can be generated as a training data set having either alpha chain data or beta chain data alone, as shown in FIG. 4 . The training data set can independently generate alpha chain data or beta chain data. The training data set may be input to the associativity prediction unit 122 .

The peptide and the HLA class II allele may generate a training data set for the immunity prediction unit 121 as shown in FIGS. 5 and 6 .

The genomic data input unit 110 measures the binding data related to the binding force for all binding relationships between the peptides and the HLA class II alleles, and uses a table or database in which the measured binding data is recorded. Binding data between peptides and HLA class II alleles can be calculated. By measuring the binding data related to the binding force for all binding relationships between the alpha chain data and the beta chain data of the peptides and the HLA class II alleles, and using a table or database in which the measured binding data is recorded, the target cancer tissue Binding data between the peptides and the alpha chain data of the HLA class II allele and the binding data between the peptides and the beta chain data of the HLA class II allele can be calculated.

The immunity prediction unit 121 may input peptides and HLA class II allele sequences as T cell activity data and output predicted values corresponding to immunity to the peptides. The immunity predictor 121 may output first predicted values corresponding to immunity to peptides by using the T cell activity data and a model learned as immunity to the peptides. A peptide may be plural or singular. The T cell activity data may include those for the peptides sequence and/or those for the HLAs sequence.

The binding predictor 1221 may receive binding data for the complex of the peptides and the HLA class II alpha-beta allele as input and output second predicted values corresponding to binding to the peptides.

The binding predictor 1222 may input binding data on a binding relationship between the peptides and the HLA class II alpha or beta single allele and output third predicted values corresponding to binding to the peptides.

The complex neoantigen predictor 131 may output a fourth predictive value, which is a neoantigen predictor of a target cancer tissue, by using the complex neoantigen predictive model learned based on the first predictive value and the second predictive value.

The single neoantigen predictor 132 may output a fifth predictive value, which is a neoantigen predictor of a target cancer tissue, by using a single neoantigen predictive model learned based on the first predictive value and the third predictive value.

The neoantigen determining unit 140 has immunity and binding properties that can be used for treatment through a machine learning ensemble model based on one first to fifth predicted values determined for each peptide among the first to fifth predicted values for each peptide. It is possible to predict whether it is a neoantigen or not.

Through this, the neoantigen determining apparatus 100 may output whether the neoantigen can be utilized for treatment in consideration of immune characteristics as well as binding characteristics of tumor cells or cancer cells included in the target cancer tissue. Also, the neoantigen determining apparatus 100 may output whether the neoantigen is a neoantigen in consideration of T cell activity data for peptides of a target cancer tissue.

The neoantigen determination apparatus 100 may be implemented including at least one of a communication unit, an input unit, and an output unit (not shown), but is not limited thereto. The neoantigen determining apparatus 100 may output data such as whether the neoantigen is recognized through the output unit. The neoantigen determination apparatus 100 may receive a data output input through the input unit. The neoantigen determination device 100 may have a communication unit and communicate with external devices. At least one of the genome data input unit 110 , the immunity prediction unit 121 , the binding prediction unit 122 , the neoantigen prediction unit 130 , and the neoantigen determination unit 140 of the neoantigen determination apparatus 100 is software Alternatively, it may be implemented in hardware. At least one of the immunity predictor 121 , the binding predictor 122 , the neoantigen predictor 130 , and the neoantigen determiner 140 may be implemented as one component.

The cancer tissue of interest may be a cell engineered to express a single HLA class I or class II allele. The target cancer tissue may be human cells obtained from or derived from a plurality of patients. The cancer tissue of interest may include fresh or frozen tumor cells obtained from a plurality of patients. The target cancer tissue may include fresh or frozen tissue cells obtained from a plurality of patients. The cancer tissue of interest may comprise peptide(s) identified using T-cell analysis.

The neoantigen determining apparatus 100 learns the algorithms of the immunity predicting unit 121, the binding predicting unit 122, the neoantigen predicting unit 130, and the neoantigen determining unit 140 based on the plurality of target cancer tissues. can do it The neoantigen determining device 100 includes data related to proteomic sequences of target cancer tissues, data related to HLA class II alleles, binding data between peptides and HLA class II alleles, data related to transcripts related to the target cancer tissue, and target cancer tissues. At least one algorithm among the immunity predictor 121, the binding predictor 122, the neoantigen predictor 130, and the neoantigen determiner 140 is to be trained using data related to the cancer tissue-related genome. can

At least one of the immunity predicting unit 121, the binding predicting unit 122, the neoantigen predicting unit 130, and the neoantigen determining unit 140 is not independently constructed for each length of the peptides, and irrespective of the length of the peptides. An algorithm (model) can be built by recognizing it as a single word. At least one of the immunity predicting unit 121, the binding predicting unit 122, the neoantigen predicting unit 130, and the neoantigen determining unit 140 may implement peptides as one word using a word embedding technique. .

At least one of the immunity predictor 121 , the binding predictor 122 , the neoantigen predictor 130 , and the neoantigen determiner 140 may be vectorized based on a similarity matrix between amino acids such as blosum.

Training data for at least one of the immunity predicting unit 121, the binding predicting unit 122, the neoantigen predicting unit 130, and the neoantigen determining unit 140 may also be input irrespective of the length of the peptide. have. The neoantigen determiner 140 may use an algorithm learned using an ensemble model of machine learning.

The neoantigen determination apparatus 100 may build a deep learning and machine learning ensemble model that classifies each positive (Y)/negative (N) based on data. The neoantigen determination apparatus 100 sets up an additional neural network by fixing weights for the immunity prediction unit 121 , the binding prediction unit 122 , the neoantigen prediction unit 130 , and the neoantigen determination unit 140 . Available.

Through this, using the immune data between the HLA class II allele and the peptide in the T cell activity data, the immunity predictor 121, the binding predictor 122, the neoantigen predictor 130, and the neoantigen determiner 140 ) may be implemented.

The immunity prediction unit 121 may apply a word embedding technique to each amino acid of the peptides. The immunity prediction unit 121 may apply CNN to a vector of peptides obtained by applying a word embedding technique and extract feature values. Here, the feature values may be acquired through learning in various layers such as CNN. The immunity prediction unit 121 may generate an algorithm through a process of training positive or negative for immunity of each peptide by applying a Gated Recurrent Unit (GRU) to the extracted feature values for the vector of peptides.

The binding predictor 122 generates vectors by applying the word embedding technique to both the alpha chain data, the beta chain data and the peptide of the HLA class II allele, and targets the vector of the HLA class II allele and the vector of the peptide. Feature values can be extracted by applying CNN. The binding predictor 122 applies the feature values to two neural networks to generate an encoder of an HLA class II allele and an encoder of a peptide, and uses the encoder of the HLA class II allele and an encoder of a peptide to generate positive for binding Alternatively, an algorithm can be generated through the process of training the voice.

The binding predictor 122 generates amino acid similarity matrix-based vectors such as blosum for both alpha chain data, beta chain data and peptide of the HLA class II allele, and targets the vector of the HLA class II allele and the vector of the peptide. By applying CNN, we can extract feature values. The binding predictor 122 applies the feature values to two neural networks to generate an encoder of an HLA class II allele and an encoder of a peptide, and uses the encoder of the HLA class II allele and an encoder of a peptide to generate positive for binding Alternatively, an algorithm can be generated through the process of training the voice.

The training data used to generate the algorithm of the neoantigen prediction unit 130 may be divided into positive and negative for immunity. The neoantigen prediction unit 130 may output a predicted value of the neoantigen having immunogenicity while binding to the HLA allele.

The immunity predicting unit 121 of the neoantigen determining apparatus 100 may output a first predicted value for immunity by receiving T cell activity data as an input.

The binding predictor 1221 may output a second predicted value for binding between the HLA class II alpha-beta complex and the peptide by inputting the T cell activity data and the binding data.

The binding predictor 1222 may output a third predicted value for binding between the HLA class II alpha or beta single allele and the peptide by inputting the T cell activity data and the binding data as inputs.

The neoantigen prediction unit 131 may output a fourth predicted value for the immunogenicity between the HLA class II alpha-beta complex and the peptide by inputting the T cell activity data and the binding data as inputs.

The neoantigen prediction unit 132 may output a fifth predicted value for the immunogenicity between the HLA class II alpha or beta single allele and the peptide by inputting the T cell activity data and the binding data as inputs.

120 and 130 of the neoantigen determination apparatus 100 may further include a predictor for predicting various factors other than the immunity predictor, the binding predictor, and the neoantigen predictor.

140 of the neoantigen determining apparatus 100 may receive the first to fifth predicted values and T cell activity data as inputs, and output whether the neoantigen can be used for treatment as either Y or N.

7 is a view for explaining a prediction model of the neoantigen determination apparatus 100 for the HLA class II alpha beta complex.

According to an embodiment of the present invention, an HLA class II allele sequence and a peptide sequence extracted from a target cancer tissue may be used as input data, and a neoantigen predicted value may be returned as output data.

In this case, in the case of the HLA class II alpha beta complex, the neoantigen determining device 100 as described in FIG. 7 includes an immunity prediction model (121'-a), a binding prediction model (122'-a), and a neoantigen Using the prediction model 123'-a, the neoantigen predicted value may be output. At this time, the first predicted value output through the immunity prediction model (121'-a), the second predicted value output through the binding prediction model (122'-a), and output through the neoantigen prediction model (123'-a) derived third predicted value.

8 is a diagram for explaining a prediction model of the neoantigen determination apparatus 100 for HLA class II alpha or beta monoliths.

According to an embodiment of the present invention, an HLA class II allele sequence and a peptide sequence extracted from a target cancer tissue may be used as input data, and a neoantigen predicted value may be returned as output data.

At this time, in the case of HLA class II alpha or beta monolith, the neoantigen determining device 100 is an immunity prediction model (121'-b), a binding prediction model (122'-b), and a neoantigen prediction model (123'-) b) can be used to output the neoantigen predicted value. At this time, the first predicted value output through the immunity prediction model (121'-b), the fourth predicted value output through the binding prediction model (122'-b), and output through the neoantigen prediction model (123'-b) A fifth predicted value is derived.

9 is a view for explaining a prediction model of the neoantigen determining device 100 for determining whether a neoantigen based on the HLA class II allele of each patient and the peptide information in the cancer tissue.

According to an embodiment of the present invention, the first to fifth predicted values may be used as input data, and whether or not a neoantigen is present may be returned as output data.

At this time, the machine learning ensemble model is learned based on the first to fifth predicted values, and the machine learning methodology used at this time uses Random Forest, Ridge, Kernel Ridge, Gradient Boosting Regression, etc., but is not limited thereto. The first to fifth predicted values may be used as input data, and whether or not the neoantigen is present may be returned as output data.

7, the immunity prediction unit 121 converts the peptide sequence into embedding wording and may be implemented as a CNN layer, a GRU layer, an NN layer, and an NN layer (121'-a).

The binding predictor 122 receives the peptide sequence and HLA class II alpha beta complex data as inputs, and calculates binding data between the peptide sequence and the HLA class II alpha chain complex, and based on the binding data, the peptide sequence and the HLA class II allele Binding between gene sequences can be output.

Binding prediction unit 122 embedding wording processing (Embedding) the peptide sequence, HLA class II alpha chain data, and HLA class II beta chain data, CNN, GRU layer (122'-a) the embedding wording processing data Through this, binding between the peptide sequence, HLA class II alpha chain data and HLA class II beta chain data can be predicted.

The binding predictor 122 vectorizes the peptide sequence, HLA class II alpha chain data, and HLA class II beta chain data based on a blosum matrix expressing the similarity between amino acids, and then CNN, GRU layer (122'-a) It is possible to predict the binding between the peptide sequence, HLA class II alpha chain data, and HLA class II beta chain data.

As shown in FIG. 8 , the HLA allele may be class II single chain data, and may be implemented as 121'-b, 122'-b.

The binding predictor 122 of the neoantigen determining device 100 predicts binding with a pair model for alpha and beta chain data of the HLA class II allele, or binds to the HLA class II allele with a single model gender can be predicted.

The neoantigen determination apparatus 100 receives the first predicted value for immunity, the second and third predicted values for binding, and the fourth predicted value and the fifth predicted value that are neoantigen predicted values, and the peptide to the target cancer tissue is the neoantigen. Whether or not it is recognized can be printed.

10 to 12 are diagrams of implementation examples of the neoantigen determination system according to embodiments of the present invention.

As shown in FIG. 10 , the neoantigen determination apparatus 100 may receive genomic data on cancer tissue from the external electronic device 200 . The neoantigen determination apparatus 100 may transmit information on whether the output peptide for cancer tissue is a neoantigen to the electronic device 200 .

The electronic device 200 may be a computing device including one or more processors storing genomic data on cancer tissue. The electronic device 200 may be a device that outputs genome data of cancer tissue. The electronic device 200 may be electrically connected to the neoantigen determining device 100 or may be connected through a wired/wireless network to transmit/receive data.

The electronic device 200 may acquire and store genome data of cancer tissues of a plurality of samples several times. The neoantigen determination apparatus 100 may sequentially output whether the genome data received from the electronic device 200 are neoantigens or the like.

11, the neoantigen determination device 100 receives data from a plurality of electronic devices (201, 202, ..., 20n), and a plurality of electronic devices (201, 202, ..., 20n) output data can be sent.

The neoantigen determination apparatus 100 may receive genome data from a plurality of electronic devices 201, 202, ..., 20n. The plurality of electronic devices 201, 202, ..., 20n may be managed by one or more subjects.

12, the neoantigen determination apparatus 100 may output output data through the output unit of one or more terminal devices (301, 302, ..., 30n). The output data may be output through the output unit of the neoantigen determination apparatus 100 . The output data may be output through an output unit of one or more of the terminal devices 301 , 302 , ..., 30n. The neoantigen determination device 100 may request payment for a predetermined cost to one or more terminal devices 301, 302, ..., 30n as data related to the neoantigen is transmitted. The one or more terminal devices 301, 302, ..., 30n may request neoantigen-related information on peptides and HLA class II alleles included in cancer tissue. In response to the request, output data may be output.

13 is a block diagram of the learning server 10 for learning an immunity prediction model, a binding prediction model, a neoantigen prediction model, and the like.

The learning server 10 may include a data input unit 11 , a first learning unit 12 , a second learning unit 13 , a third learning unit 14 , and a fourth learning unit 15 .

The first learning unit 12 is generated by learning the immunity prediction model, and learns the T cell activity data of the peptide sequence or the HLA class II allele sequence and the immunity of the peptide sequence as a training data set. As shown in 121 of FIG. 1, the immunity prediction model trained by the first learning unit 12 processes the peptide sequence by word embedding technique, and inputs the processed peptide sequence to the layers of CNN, GRU, and NN. will learn

The second learning unit 13 is generated by learning the binding prediction model, and training data on binding between the peptide sequence and the HLA class II allele sequence by inputting the binding data of the peptide sequence or the HLA class II allele sequence. learn as a set. As shown in 122 of FIG. 1, the binding prediction model learned by the second learning unit 13 processes the peptide sequence and the HLA class II allele sequence by word embedding technique, respectively, and the processed peptide sequence is CNN , learned by input into the layers of the GRU, and HLA class II allele sequences were applied to CNN, CNN. It learns by inputting it into the GRU layer. The binding prediction model trains another model (NN1) with the predicted value for binding to the peptide sequence and the predicted value for binding to the HLA class II allele sequence, and finally the predicted value for the neoantigen to the target cancer tissue. It can be learned to output.

The third learning unit 14 is generated by learning the neoantigen prediction model, and when the peptide sequence and the HLA allele sequence are given by inputting the peptide sequence and the HLA class II allele sequence, whether the corresponding peptide is a neoantigen possibility is trained as a training data set. As shown in 130 of FIG. 1, the new antigen prediction model trained by the third learning unit 14 processes the peptide sequence and the HLA class II allele sequence by word embedding technique, respectively, and the processed peptide sequence is CNN , it learns by inputting it into the layers of the GRU, and learning by inputting the HLA class II allele sequence into the layers of CNN, CNN, and GRU. The neoantigen prediction model may be trained to learn another model (NN2) with a given HLA class II allele sequence and a peptide sequence-based neoantigen predicted value, and finally to output a predicted value for the neoantigen.

The learning server 10 may transmit the learning models generated by the first to third learning units 12 , 13 , and 14 to the neoantigen determination apparatus 100 . Through this, the algorithms of the immunity predictor 121 , the binding predictor 122 , and the neoantigen predictor 130 of the neoantigen determination apparatus 100 may be periodically updated (updated).

The device described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA). , a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions, may be implemented using one or more general purpose or special purpose computers. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that can include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

The software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or device, to be interpreted by or to provide instructions or data to the processing device. , or may be permanently or temporarily embody in a transmitted signal wave. The software may be distributed over networked computer systems, and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

As described above, although the embodiments have been described with reference to the limited embodiments and drawings, various modifications and variations are possible from the above description by those skilled in the art. For example, the described techniques are performed in a different order than the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (12)

receiving, by the neoantigen determining device, a peptide sequence extracted from a target cancer tissue and an HLA class II allele sequence as inputs;
obtaining, by the neoantigen determining device, T cell activity data from the peptide sequence, inputting the T cell activity data into an immunity prediction model, and outputting a first predicted value for predicting immunity of the peptide sequence;
The neoantigen determining device inputs the alpha-beta complex sequence data and the peptide sequence data of the HLA class II allele sequence to the binding prediction model to predict the binding property of the peptide sequence and the HLA class II allele sequence. outputting a predicted value;
The neoantigen determination device inputs the alpha or beta single sequence data and the peptide sequence data of the HLA class II allele sequence to the binding prediction model to predict the binding of the peptide sequence and the HLA class II allele sequence 3 outputting a predicted value;
outputting, by the neoantigen determining device, a fourth predicted value for predicting whether the peptide sequence of the target cancer tissue is a neoantigen using the T cell activity data and the first predicted value and the second predicted value;
outputting, by the neoantigen determining device, a fifth predicted value for predicting whether the peptide sequence of the target cancer tissue is a neoantigen using the T cell activity data and the first predicted value and the third predicted value;
Generating neoantigen information including whether the peptide sequence of the target cancer tissue is a neoantigen using the T cell activity data and the first to fifth predicted values by the neoantigen determining device A method for predicting neoantigens using a peptide sequence derived from cancer tissue and cell-free DNA and an HLA class II allele sequence. According to claim 1,
In the step of generating neoantigen information on the target cancer tissue,
A target cancer tissue and cell-free DNA-derived peptide sequence and HLA class II allele sequence for generating neoantigen information on a peptide of the target cancer tissue using the T cell activity data and the first to fifth predicted values A method for predicting neoantigens using According to claim 1,
At least one of the immunity prediction model and the binding prediction model includes a peptide sequence present in a plurality of target cancer tissues, alpha chain data of HLA class II allele sequence, and beta chain data of HLA class II allele sequence A method for predicting neoantigens using peptide sequences and HLA class II allele sequences from target cancer tissue and cell-free DNA, trained by a machine learning algorithm based on a training data set. 3. The method of claim 2,
The target cancer tissue is
A method for predicting neoantigens using peptide sequences and HLA class II allele sequences from subject cancer tissue and cell-free DNA, including cells engineered to express a single MHC class I or class II allele. 4. The method of claim 3,
The target cancer tissue is
A method for predicting neoantigens using peptide sequences and HLA class II allele sequences derived from target cancer tissue and cell-free DNA, including human cells obtained from or derived from a plurality of patients. 4. The method of claim 3,
The target cancer tissue is
A method for predicting neoantigens using peptide sequences and HLA class II allele sequences derived from target cancer tissue and cell-free DNA, comprising fresh or frozen tumor cells obtained from a plurality of patients. 4. The method of claim 3,
The target cancer tissue is
A method for predicting neoantigens using a peptide sequence derived from a target cancer tissue and cell-free DNA and an HLA class II allele sequence, comprising fresh or frozen tissue cells obtained from a plurality of patients. 4. The method of claim 3,
The target cancer tissue is
A method for predicting neoantigens using peptide sequences derived from subject cancer tissue and cell-free DNA and HLA class II allele sequences, including peptides identified using T-cell analysis. 4. The method of claim 3,
The training data set is
data related to the proteomic sequence related to the target cancer tissue, data related to the MHC peptide sequence related to the target cancer tissue, binding data between the peptide related to the target cancer tissue and class II alpha chain data of the HLA class II allele, the target At least one of binding data between a peptide associated with cancer tissue and class II beta chain data of an HLA class II allele, data associated with a transcript associated with the target cancer tissue, and data related to a genome associated with the target cancer tissue A method for predicting a neoantigen using a target cancer tissue and cell-free DNA-derived peptide sequence and an HLA class II allele sequence. According to claim 1,
The immunity prediction model is
Predicting neoantigens using T cell activity data from a peptide sequence as an input, and a peptide sequence derived from the target cancer tissue and cell-free DNA and HLA class II allele sequence, which is a trained model as an output of the immunity of the peptide sequence. How to. According to claim 1,
The binding prediction model is
First binding data between the alpha chain data of the HLA class II allele sequence and the peptide sequence, the second binding data between the beta chain data of the HLA class II allele sequence and the peptide sequence, and the alpha beta chain of the HLA class II allele sequence A neoantigen using a target cancer tissue and cell-free DNA-derived peptide sequence and HLA class II allele sequence, which is a model trained to output the binding properties of the peptide sequence and the HLA class II allele sequence as an input to the complex How to predict. A computer program stored in a computer-readable storage medium for executing the method of any one of claims 1 to 11 using a computer.

Images