A privacy-preserving genetic sequence comparison and pairing method and system

Abstract

The present invention relates to a privacy-protected gene sequence comparison and pairing method and system, belonging to the field of information security technology. It includes: sequence share arrangement: the server specifies two secrets and generates a corresponding secret share set, and then the server arranges the location of the share according to each base and paired base in the gene sequence held; oblivious transfer: the server and the user execute the oblivious transfer protocol, i.e., the OT protocol; the execution process is performed one by one according to the base sequence, the server inputs the arranged share quadruple, and the user obtains one of the quadruple according to the base of the gene sequence held; pattern matching: the user obtains all secret shares and calls the secret recovery algorithm, and recovers the secret result by comparison to obtain the matching result. The present invention utilizes the pairing rules of the base sequence, and the method can simultaneously realize the privacy comparison of the same gene and the paired gene, thereby improving the efficiency of the privacy computing protocol.

Description

Technical Field

[0001] The invention relates to a privacy-protected gene sequence comparison and pairing method and system, belonging to the technical field of information security. Background Art

[0002] Privacy computing usually includes technologies such as secure multi-party computing, federated learning, homomorphic encryption, and trusted execution environment, aiming to achieve secure sharing and efficient use of data while ensuring data privacy and security. Privacy computing has become one of the key technologies in the big data era. Among the many technologies of privacy computing, the privacy comparison protocol is an important basic protocol that can be used to construct secure pattern matching protocols, privacy set operation protocols, and privacy-preserving machine learning basic operators.

[0003] Pattern matching is a basic algorithm in computer science, which can be simply summarized as the problem of finding a specific substring (called a pattern) in a long string (usually called a text). The answer to the pattern matching problem can be the position of the substring, the number of times it appears, etc. When the length of the text and the pattern are equal, the secure pattern matching problem degenerates into a privacy comparison protocol for strings, that is, the two parties involved determine whether the strings held by both parties are the same without leaking their respective data. In the pattern matching problem, the long string represented by the text can be divided into multiple substrings of equal length according to the length of the pattern, and then these substrings are compared one by one with the specific substring. Therefore, the privacy comparison protocol is an important subprotocol for implementing secure pattern matching.

[0004] With the rapid development of genetic technology, gene sequence comparison plays a vital role in biology, modern medicine and other fields. The genetic information of organisms can be analyzed with unprecedented accuracy, which has an important impact on understanding genetic information and finding the root causes of diseases. At the same time, it has brought revolutionary breakthroughs in personalized medicine, drug development and precision agriculture. Genetic information is reflected in the nucleic acid sequence, which is mainly composed of four nitrogenous bases, namely A (adenine), T (thymine), C (cytosine), and G (guanine). These four nitrogenous bases are linearly arranged and paired according to a fixed rule (A is paired with T, and C is paired with G) to form two sequences. However, genetic data contains highly sensitive information, and its leakage may cause serious social ethical issues.

[0005] Given two base sequences, we need to consider whether the two sequences are exactly the same, and whether the bases at the same position in the two sequences are paired. We call the algorithms corresponding to these two situations sequence comparison and sequence pairing, respectively. Obviously, the privacy comparison protocol in the existing security pattern matching protocol can only support privacy-preserving sequence comparison, but cannot support sequence pairing. In order to meet these two requirements, it is necessary to adjust the sequence input and call the protocol again, or to design a privacy-preserving gene sequence pairing protocol separately. Therefore, it is very necessary to design a privacy-preserving method that can complete both sequence comparison and sequence pairing. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention provides a privacy-preserving gene sequence comparison and pairing method;

[0007] The present invention utilizes the pairing rules of base sequences to improve the efficiency of the privacy computing protocol.

[0008] The present invention also provides a privacy-protected gene sequence comparison and pairing system.

[0009] In the present invention, the secret sharing (SS) algorithm: Secret sharing was proposed by Shamir and Blakley in 1979 respectively. The algorithm includes two stages: secret sharing and secret reconstruction: in the secret sharing stage, a secret is divided into multiple shares, and each individual share does not contain enough information to infer or reconstruct the original secret, thereby ensuring the security of the original secret; in the secret reconstruction stage, only when a sufficient number of shares are obtained can the original secret be successfully reconstructed. The minimum number of shares τ required to reconstruct the secret is called the threshold of secret sharing. Oblivious Transfer (OT): The oblivious transfer protocol is one of the basic cryptographic protocols. The protocol includes two participants, namely the sender Sender and the receiver Receiver. In the standard two-choice oblivious transfer protocol, the sender Sender inputs two messages , the receiver inputs the selection bit σ∈{0,1}. After the protocol is completed, the sender has no output and does not know which message the receiver has selected. The receiver gets the output And can't get the message Any information.

[0010] The technical solution of the present invention is:

[0011] A privacy-preserving gene sequence comparison and pairing method, assuming that the user User holds a gene sequence and the server Server holds a gene sequence, allows the user to judge the relationship between the gene sequence held by the user User and the gene sequence held by the server Server without obtaining the gene sequence held by the server Server: if the sequences are identical, the sequences are paired, and if the two sequences are neither identical nor paired; including:

[0012] Sequence share arrangement: The server specifies two secrets and generates a corresponding set of secret shares. The server then arranges the positions of the shares according to each base and paired base in the gene sequence it holds.

[0013] Oblivious transfer: The server and the user execute the oblivious transfer protocol, i.e., the OT protocol. The execution process is performed one by one according to the base sequence. The server inputs the arranged share quadruple, and the user obtains one of the quadruple according to the base of the gene sequence.

[0014] Pattern matching: The user obtains all secret shares and calls the secret recovery algorithm to recover the secret results through comparison and obtain the matching results.

[0015] According to a preferred embodiment of the present invention, the server specifies two secrets and generates a corresponding set of secret shares, and then the server arranges the positions of the shares according to each base and paired base in the gene sequence held by the server; including:

[0016] First, the server randomly generates two The bit string is used as the input of the secret sharing algorithm and is recorded as and ;

[0017] Then, the server calls the secret sharing algorithm to and Generate two sets of secret shares, denoted as and ; According to the properties of the secret sharing algorithm, for , , there exists an equation = ;

[0018] The server sends a bit string to the user ( , );

[0019] for , the server sets the message quad according to the following rules The specific value of each element:

[0020] ① For bases , the server follows Corresponding to The order will Fill in the position corresponding to the message quad ;

[0021] ② For bases Base pairing ,Will Fill in the position corresponding to the message quadruple;

[0022] ③For each message quadruple, the server first randomly generates two The two random bit strings are then regarded as secret shares, expressed as ,Will Fill in There are two unassigned positions left;

[0023] Finally, for the length The server forms a base sequence of all message quads into a The matrix :

[0024] .

[0025] Further preferably, the server Server calls the secret sharing algorithm to and Generate two sets of secret shares; that is, the secret sharing algorithm adopts the form of XOR sharing, that is, when the input and After that, the secret sharing algorithm randomly generates Individual and and Bit strings of equal length are used as secret shares, i.e. and .

[0026] According to the preferred embodiment of the present invention, it is assumed that the server has a The specific implementation process of sequence share arrangement includes:

[0027] First, the server randomly generates three numbers of length The bit string is denoted as , and ;

[0028] Then, the server calls the secret sharing algorithm to and Generate 𝑛-1 secret shares for and The two groups of shares are recorded as ;for , ,when season ,have = ;

[0029] The server sends a bit string to the user ( , );

[0030] for , the server sets the message quad according to the following rules:

[0031] when or When the server sets the message quad according to ①, ② and ③ ;when When the message quadruple Each element is set to ;

[0032] Finally, all tuples are formed into a single array of size The matrix :

[0033] .

[0034] According to the preferred embodiment of the present invention, the server and the user perform an oblivious transfer protocol in which the function is selected from 4, denoted as ; This stage requires calling 𝑛 times , in the 𝑖th call:

[0035] The server acts as the sender and inputs the matrix No. OK ;

[0036] The user acts as the receiver and enters bases ;

[0037] according to Corresponding to In the order of .

[0038] Preferably, according to the present invention, the user obtains all secret shares and calls the secret recovery algorithm, recovers the secret result by comparison, and obtains the matching result; including:

[0039] Users according to 𝑛 times Output, calculation = , and Respectively and Compare:

[0040] if , then the gene sequence and Same; User User holds the gene sequence , the server holds the gene sequence ;

[0041] if , then the gene sequence and The paired sequence same;

[0042] In other cases, gene sequences and gene sequences Not identical and not matching.

[0043] A privacy-preserving gene sequence comparison and matching system, comprising:

[0044] The sequence share arrangement module is configured as follows: the server specifies two secrets and generates a corresponding set of secret shares, and then the server arranges the location of the shares according to each base and paired base in the gene sequence held;

[0045] The oblivious transfer module is configured as follows: the server and the user execute the oblivious transfer protocol, i.e., the OT protocol; the execution process is performed one by one according to the base sequence, the server inputs the arranged share quadruple, and the user obtains one of the quadruple according to the base of the gene sequence held;

[0046] The pattern matching module is configured as follows: the user obtains all secret shares and calls the secret recovery algorithm, recovers the secret result by comparison, and obtains the matching result.

[0047] The beneficial effects of the present invention are:

[0048] 1. The present invention provides a privacy-preserving gene sequence comparison and pairing method and system. By utilizing the pairing rules of base sequences, the method can simultaneously realize the privacy comparison of the same gene and paired genes, thereby improving the efficiency of the privacy computing protocol.

[0049] 2. The present invention provides a privacy comparison method for gene sequences with wildcards, which can achieve privacy comparison of double-matching gene sequences even when individual bases in the gene sequence are meaningless.

[0050] 3. Based on XOR secret sharing and oblivious transmission, the server cannot obtain any information about the genetic sequence held by the user, and when the match is unsuccessful, the user does not know the genetic sequence information held by the server, thus ensuring the security of the data of both parties. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] FIG. 1 is a schematic diagram of the process flow of the privacy-preserving gene sequence comparison and pairing method in Example 2. DETAILED DESCRIPTION

[0052] The present invention will be further defined below in conjunction with the accompanying drawings and embodiments, but is not limited thereto.

[0053] Example 1

[0054] A privacy-preserving gene sequence comparison and pairing method, assuming that the user User holds the gene sequence , the server holds the gene sequence , allowing users to access the gene sequence without obtaining the server In the case of and the server holds the gene sequence The relationship between: if the sequences are identical, the sequences are paired; if the two sequences are neither identical nor paired; Figure 1 As shown, including:

[0055] Sequence share arrangement: The server specifies two secrets and generates a corresponding set of secret shares. The server then arranges the positions of the shares according to each base and paired base in the gene sequence it holds.

[0056] Oblivious transfer: The server and the user execute the oblivious transfer protocol, i.e., the OT protocol. The execution process is performed one by one according to the base sequence. The server inputs the arranged share quadruple, and the user obtains one of the quadruple according to the base of the gene sequence.

[0057] Pattern matching: The user obtains all secret shares and calls the secret recovery algorithm to recover the secret results through comparison and obtain the matching results.

[0058] Example 2

[0059] The privacy-preserving gene sequence comparison and pairing method described in Example 1 is different in that:

[0060] An example of the solution of the present invention is shown in FIG1 , which uses a gene sequence of length 7 for illustration. The server specifies two secrets and generates a corresponding set of secret shares. In Figure 1, the two sets of secret shares are distinguished by superscripts 1 and 2, respectively. The corresponding calculation relationships are represented by solid arrows and dotted arrows, respectively, and two random shares are generated for the position corresponding to each base. Subsequently, the server arranges the positions of the shares according to the conditions of each base and the paired bases in the gene sequence. Only when the user obtains all real shares with superscripts 1 or 2, can he recover one of the two secrets according to the rules of the solid arrows or dotted arrows when arranging the shares, which means that the gene sequence matches successfully.

[0061] Gene sequences are written in uppercase letters Indicates that Indicates gene sequence No. The length of the gene sequence held by both parties is ; Safety parameters are ;

[0062] The basic components used include the oblivious transfer protocol and the secret sharing algorithm;

[0063] The basic components used include the oblivious transfer protocol and the secret sharing algorithm;

[0064] The gene sequence input by the server is represented as , whose paired sequence is ;

[0065] The gene sequence entered by the user is .

[0066] The server specifies two secrets and generates a corresponding set of secret shares. The server then arranges the positions of the shares according to each base and paired base in the gene sequence it holds. Specifically, the server follows Corresponding to The shares are placed in the order of the bases, that is, when the base is A, the share is placed in position 1, when it is T, it is placed in position 2, when it is C, it is placed in position 3, and when it is G, it is placed in position 4. Including:

[0067] First, the server randomly generates two The bit string is used as the input of the secret sharing algorithm and is recorded as and ;

[0068] Then, the server calls the secret sharing algorithm to and Generate two sets of secret shares, denoted as and ; According to the properties of the secret sharing algorithm, for , , there exists an equation = ;

[0069] The server sends a bit string to the user ( , );

[0070] for , the server sets the message quad according to the following rules The specific value of each element:

[0071] ① For bases , the server follows Corresponding to The order will Fill in the position corresponding to the message quad ; That is: base When it is A, the share Put it in position 1 and make ; When T, Put it in position 2 and make ; When it is C, Put it in position 3 and make , When G, Put it in position 4 and ;

[0072] ② For bases Base pairing ,Will Fill in the position corresponding to the message quadruple; base When it is A, the share Put it in position 1 and make ; Base When T, Put it in position 2 and make , base When it is C, Put it in position 3 and make ; Base When G, Put it in position 4 and make ;

[0073] ③ Base and paired bases Message quad The two elements in are assigned values, and the server now processes There are two unassigned elements in the message. For each message quadruple, the server first randomly generates two The two random bit strings are then regarded as secret shares, expressed as ,Will Fill in There are two unassigned positions left;

[0074] Finally, for the length The server forms a base sequence of all message quads into a The matrix :

[0075] .

[0076] Example 3

[0077] The privacy-preserving gene sequence comparison and pairing method according to Embodiment 1 or 2 is different in that:

[0078] Some bases in a gene sequence are meaningless, that is, they have no genetic expression function. If such meaningless bases exist in the gene sequence held by the server, it is hoped that the algorithm of the present invention can ignore the base pairing situation, that is, the user's gene sequence can still be compared or matched except for specific positions. The mutated base in is treated as a wildcard *, which can match Any one of them.

[0079] The server calls the secret sharing algorithm to and Generate two sets of secret shares; that is, the secret sharing algorithm adopts the form of XOR sharing, that is, when the input and After that, the secret sharing algorithm randomly generates Individual and and The bit strings of equal length are used as secret shares, i.e. and .

[0080] Assume that the server has a The specific implementation process of sequence share arrangement includes:

[0081] First, the server randomly generates three numbers of length The bit string is denoted as , and ;

[0082] Then, the server calls the secret sharing algorithm to and Generate 𝑛-1 secret shares for and The two groups of shares are recorded as ; Obviously, for , ,when season ,have = ;

[0083] The server sends a bit string to the user ( , );

[0084] for , the server sets the message quad according to the following rules:

[0085] when or When the server sets the message quad according to ①, ② and ③ ;when When the message quadruple Each element is set to ;

[0086] Finally, all tuples are formed into a single array of size The matrix :

[0087] .

[0088] The server and the user perform an oblivious transfer protocol in which the function is 1 out of 4, denoted as ; Therefore, the algorithm time consumption is evaluated by inadvertent transmission expansion. This stage requires calling 𝑛 times , in the 𝑖th call:

[0089] The server acts as the sender and inputs the matrix No. OK ;

[0090] The user acts as the receiver and enters bases ;

[0091] according to Corresponding to In the order of .

[0092] The user obtains all secret shares and calls the secret recovery algorithm to recover the secret result through comparison and obtain the matching result; including:

[0093] Users according to 𝑛 times Output, calculation = , and Respectively and Compare:

[0094] if , then the gene sequence and Same; User User holds the gene sequence , the server holds the gene sequence ;

[0095] if , then the gene sequence and The paired sequence same;

[0096] In other cases, gene sequences and gene sequences Not identical and not matching.

[0097] when Length is , Length is When you need to execute indivual Example. The test is mainly based on the libOTe library, and the time overhead corresponding to the XOR operation is ignored. Table 1 shows the test results of the present invention under different settings. All running times are the average values ​​of 20 experiments. The settings are , .

[0098] Table 1 Operation schedule (unit );

[0099]

[0100] Example 4

[0101] A privacy-preserving gene sequence comparison and matching system, comprising:

[0102] The sequence share arrangement module is configured as follows: the server specifies two secrets and generates a corresponding set of secret shares, and then the server arranges the location of the shares according to each base and paired base in the gene sequence held;

[0103] The oblivious transfer module is configured as follows: the server and the user execute the oblivious transfer protocol, i.e., the OT protocol; the execution process is performed one by one according to the base sequence, the server inputs the arranged share quadruple, and the user obtains one of the quadruple according to the base of the gene sequence held;

[0104] The pattern matching module is configured as follows: the user obtains all secret shares and calls the secret recovery algorithm, recovers the secret result by comparison, and obtains the matching result.

Claims

1. A privacy-preserving gene sequence comparison and pairing method, characterized in that: Assume that the user User holds a gene sequence and the server Server holds a gene sequence. The user can judge the relationship between the gene sequence held by the user User and the gene sequence held by the server Server without obtaining the gene sequence held by the server Server: if the sequences are the same, the sequences are paired; if the two sequences are neither the same nor paired; include: Sequence share arrangement: The server specifies two secrets and generates a corresponding set of secret shares. The server then arranges the positions of the shares according to each base and paired base in the gene sequence it holds; including: First, the server randomly generates two bit strings of length λ as the input of the secret sharing algorithm, denoted as D1 and D2; Subsequently, the server calls the secret sharing algorithm to generate two sets of secret shares for D1 and D2, respectively, denoted as and According to the properties of the secret sharing algorithm, for i∈{1,2,…,n},j∈{1,2}, there exists the equation The server sends a bit string (D1, D2) to the user; For 1≤i≤n, the server sets the message quadruple (t i,1 ,t i,2 ,t i,3 ,t i,4 )The specific value of each element: ① For base X s [i], the server follows the order (A, T, C, G) corresponding to (1, 2, 3, 4) respectively. Fill in the position t corresponding to the message quadruple i,j ; ②For base X s [i]'s paired base X s′ [i], will Fill in the position corresponding to the message quadruple; ③For each message quadruple, the server first generates two random bit strings with the same length as D1 locally, and then regards these two random bit strings as secret shares, expressed as and Will and Fill in (t i,1 ,t i,2 ,t i,3 ,t i,4 ) leaves two unassigned positions; Oblivious transfer: The server and the user execute the oblivious transfer protocol, i.e., the OT protocol. The execution process is performed one by one according to the base sequence. The server inputs the arranged share quadruple, and the user obtains one of the quadruple according to the base of the gene sequence. Pattern matching: The user obtains all secret shares and calls the secret recovery algorithm to recover the secret results through comparison and obtain the matching results; Assume that there is a meaningless base at the kth position on the server; the specific implementation process of sequence share arrangement includes: First, the server randomly generates three bit strings of length λ, denoted as D1 and D2; Then, the server calls the secret sharing algorithm to and Generate n-1 secret shares. For i∈{1,2,…,n} and i≠k, the two sets of shares are recorded as and For i∈{1,2,…,n},j∈{1,2}, when i=k, let have The server sends a bit string (D1, D2) to the user; For 1≤i≤n, the server sets the message quad according to the following rules: When 1 ≤ i < k or k < i ≤ n, the server sets the message quadruple (t i,1 , t i,2 , t i,3 , t i,4 ) according to ①, ②, and ③; when i = k, each element of the message quadruple (t i,1 , t i,2 , t i,3 , t i,4 ) is set to Finally, for a base sequence of length n, the server forms all message quads into a matrix T of size n*4:

2. A privacy-preserving gene sequence comparison and pairing method according to claim 1, characterized in that: The server calls the secret sharing algorithm to generate two sets of secret shares for D1 and D2 respectively; that is, the secret sharing algorithm adopts the form of XOR sharing, that is, when D1 and D2 are input, the secret sharing algorithm randomly generates n bit strings of the same length as D1 and D2 as secret shares, that is, and 3. A privacy-preserving gene sequence comparison and pairing method according to claim 1, characterized in that: The server and the user perform an oblivious transfer protocol in which the function is 1 out of 4, denoted as This stage needs to be called n times In the i-th call: The server acts as the sender and inputs the i-th row (t i,1 ,t i,2 ,t i,3 ,t i,4 ); The user acts as the receiver and inputs the i-th base X u [i]; According to the order of (A, T, C, G) corresponding to (1, 2, 3, 4), the user obtains the output 4. A privacy-preserving gene sequence comparison and pairing method according to any one of claims 1 to 3, characterized in that: The user obtains all secret shares and calls the secret recovery algorithm to recover the secret result through comparison and obtain the matching result; including: User according to n times Output, calculation And D u Compare with D1 and D2 respectively: If D u =D1, then the gene sequence X u With X s Same; User User holds the gene sequence X u The server holds the gene sequence X s ; If D u =D2, then the gene sequence X u With X s The paired sequence X s′ same; In other cases, gene sequence X u and gene sequence X s Not identical and not matching.

5. A privacy-preserving gene sequence comparison and matching system, characterized in that: include: The sequence share arrangement module is configured as follows: the server specifies two secrets and generates a corresponding secret share set, and then the server arranges the location of the share according to each base and paired base in the gene sequence held; including: First, the server randomly generates two bit strings of length λ as the input of the secret sharing algorithm, denoted as D1 and D2; Subsequently, the server calls the secret sharing algorithm to generate two sets of secret shares for D1 and D2, respectively, denoted as and According to the properties of the secret sharing algorithm, for i∈{1,2,…,n},j∈{1,2}, there exists the equation The server sends a bit string (D1, D2) to the user; For 1≤i≤n, the server sets the message quadruple (t i,1 ,t i,2 ,t i,3 ,t i,4 )The specific value of each element: ① For base X s [i], the server follows the order (A, T, C, G) corresponding to (1, 2, 3, 4) respectively. Fill in the position t corresponding to the message quadruple i,j ; ②For base X s [i]'s paired base X s′ [i], will Fill in the position corresponding to the message quadruple; ③For each message quadruple, the server first generates two random bit strings with the same length as D1 locally, and then regards these two random bit strings as secret shares, expressed as and Will and Fill in (t i,1 ,t i,2 ,t i,3 ,t i,4 ) leaves two unassigned positions; The oblivious transfer module is configured as follows: the server and the user execute the oblivious transfer protocol, i.e., the OT protocol; the execution process is performed one by one according to the base sequence, the server inputs the arranged share quadruple, and the user obtains one of the quadruple according to the base of the gene sequence held; The pattern matching module is configured as follows: the user obtains all secret shares and calls the secret recovery algorithm, recovers the secret result by comparison, and obtains the matching result; Assume that there is a meaningless base at the kth position on the server; the specific implementation process of sequence share arrangement includes: First, the server randomly generates three bit strings of length λ, denoted as D1 and D2; Then, the server calls the secret sharing algorithm to and Generate n-1 secret shares. For i∈{1,2,…,n} and i≠k, the two sets of shares are recorded as and For i∈{1,2,…,n},j∈{1,2}, when i=k, let have The server sends a bit string (D1, D2) to the user; For 1≤i≤n, the server sets the message quad according to the following rules: When 1 ≤ i < k or k < i ≤ n, the server sets the message quadruple (t i,1 , t i,2 , t i,3 , t i,4 ) according to ①, ② and ③; when i = k, each element of the message quadruple (t i,1 , t i,2 , t i,3 , t i,4 ) is set to Finally, for a base sequence of length n, the server forms all message quads into a matrix T of size n*4:

Images