KR20030096658A - A method of identifying drug targets using microbial genomic sequence data - Google Patents

Abstract

A protein sequence similarity comparison program is used for protein databases created from microbial genome information to identify homologs common to each microbial genome and use them to target new targets for antibiotic / antifungal development.

As a method for searching for drug targets, homologs with low sequence similarity can be found and a basis for predicting the function of a gene whose function is not known can be prepared.

Description

A method of identifying drug targets using microbial genomic sequence data}

The present invention relates to a method for searching protein sequence information, and more particularly, to a search method capable of finding homologs having low sequence similarity.

When comparing the genomes of two or more different organisms, the sequence of proteins produced from the genome is compared to predict the function of the two proteins to the extent of similarity. In this case, if the two proteins have similarities, the two proteins become homology and the two proteins are called homologs.

The number of nucleic acids and proteins is growing exponentially as the Genome project of various organisms progresses. To identify the function of the newly discovered protein, several biological experiments must be performed, but the function of the newly discovered protein can be predicted by comparing the similarity of the sequence with the previously confirmed protein.

There are many programs used to compare these sequences, including BLAST and FASTA. Similarity search is performed on a protein sequence database using a protein sequence of unknown function as a query. If a highly similar protein sequence is found and the function of the protein is known experimentally, one can predict that the function used as the query is also unknown.

This method is commonly used because it allows you to easily predict the function of newly discovered proteins without resorting to experiments with protein sequences alone. The problem with this method is that if the sequence similarity is low, even if the two proteins are homologs of the same function, they cannot be found.

Organisms cause changes in nucleic acid sequences over time, and thus change in protein sequences. Although the sequence of proteins is similar among evolutionarily close organisms, the sequence of proteins of evolutionarily distant organisms is very weak. For example, humans are evolutionarily closer to monkeys than bacteria, so human proteins are more similar to monkey proteins. Finding homologs that perform the same function by analyzing the genomes of several organisms can be very useful in many fields. However, current BLAST or FASTA methods have limitations in finding homologs with low sequence similarity.

An object of the present invention is to provide a method for identifying homologs having low similarity in sequence.

Another object of the present invention is to provide a method that can search for a novel target protein of antibiotic / antifungal agent.

It is still another object of the present invention to provide a method that can provide a basis for predicting the function of a protein whose function is unknown.

The target protein search method according to the present invention for achieving this object is a process of selecting a plurality of protein sequence database to know the homology (homology) between protein sequences from a plurality of databases having protein sequence information of the organism (S1) and; Selecting one of the protein sequences included in any one of the protein databases as a query of a sequence similarity search program (S2); Running a similarity search program to compare the similarities with respect to each of the databases except for the protein database of the organism including the query (S3); Among the protein sequences satisfying a predetermined condition among the driving results of the similarity search program, the protein sequence information (“besthit”) having the smallest cut-off value is stored according to the sequence similarity comparison procedure. Generating a computer file (S4); Selecting a protein sequence located first in a file in which the besthit protein sequence information is stored as a query of a similarity search program (S5); Performing steps S5, S3 and S4 a plurality of times in a sequence S5-> S3-> S4 (S6); Between at least three protein sequence information belonging to a different database among the besthit protein sequence information stored in the files generated by the S2 or S4 or S6 process and the query protein sequence information used to generate the besthit protein sequence information. Searching for protein sequence information having a besthit or query relationship with respect to two protein sequences in which any one protein sequence has a relation of (S7); And (S8) extracting homolog information present in different protein databases using the protein sequence information found in step S7.

1 is a flow chart showing a target protein search process according to the present invention,

Figure 2 is an illustration showing a protein sequence database of five organisms,

3 is an illustration of a query and its besthit when using the Blastp program,

4 is an exemplary view illustrating a seed forming process in a search method according to the present invention;

5 is an exemplary view of a final seed forming process of the search method according to the present invention;

6 is an exemplary view illustrating a seed forming process in another case of the searching method according to the present invention;

7 is an exemplary view illustrating a seed forming process in another case of the searching method according to the present invention.

Hereinafter, a method for searching for a target protein according to the present invention will be described with reference to the accompanying drawings. The homolog search method according to the present invention is a method of finding a protein sequence homologous to any one protein sequence in any one of the selected protein databases in another database.

1 is a flow chart showing the progress of the target protein search method according to an embodiment of the present invention.

The implementation of the present invention requires a protein sequence similarity comparison program with the entire protein sequence database of the organism to be compared. Genome projects of various organisms are underway, and in many cases protein sequence databases are available free of charge. For example, the required protein sequence database can be downloaded from the NCBI web page on the Internet. The protein sequence similarity program also uses BLAST provided by NCBI. Among them, BLASTP uses protein sequences as queries to search for homology to selected protein databases. In addition, there are programs for comparing various sequence similarities such as FASTA, but in this description, BLAST (particularly BLASTP) is used as an example.

Among a plurality of databases having protein sequence information of various organisms, a plurality of protein sequence databases whose homology is desired to be known are selected as shown in FIG. 2. Figure 2 is a schematic diagram showing the protein sequence database of five organisms (S1 process).

First, a file (eg: mge1.doc) for querying protein sequence information A1 of a randomly selected database A is generated. Here, the file is for processing electronic information in a form that can be read using a computer, and can also be processed using a storage device such as a RAM (S2 process).

BLASTP is run to select one of the protein sequences included in any one of the searched protein databases as a query to compare the similarities to the rest of the databases except the protein database of the organism containing the query. That is, any protein sequence (in this example, A1 of organism A is selected as an arbitrary value) of any database is used as a query for the remaining protein databases (B, C, D, E). To compare the similarities. In this case, a gapped local alignment mode (gapped blastp) is used and the e-value is <0.0001 (S3 process).

Similarity is compared between protein sequences B1 to BN in database B to indicate that the e-value is the smallest (hereinafter referred to as "besthit") among the protein sequences of which the e-value is less than 10-4. Assuming the protein sequence is B1, store protein sequence B1 in the first location of the result file (eg: mge1.out).

Thereafter, similarity is compared with protein sequences C1 to CN of another database C in the same manner as described above. In this case, if the protein sequence corresponding to besthit is C2, the protein sequence is stored after the protein sequence B1 is located. Thereafter, the same process is performed on the databases D and E. At this time, the besthit is determined only when there is a protein sequence satisfying a predetermined condition, that is, e-value <0.0001. The Besthit data file (mge1.out) is stored in FASTA format. The FASTA format consists of a header describing the sequence of the protein and the characteristics of the protein.

Table 1 below shows the relationship between the query and the beshit. This example is for illustrative purposes only and does not represent actual data.

TABLE 1

3 is an exemplary diagram illustrating a BLAST search result. In other words, when a sequence similarity search is performed on the remaining protein databases B, C, D, and E using one of the protein sequences of organism A as a query, the besthit is B1, C2, D1, and E2. (S4 course)

The second BLAST search uses the protein sequence (ie, B1) stored in the first location of the file where besthit is stored (mge1.out) (step S5).

A sequence similarity comparison is made for all databases (A, C, D, E) to be compared, except for database (B), including protein sequence (B1) used as a query. In this case, it is assumed that A2, C2, D2, and E2 become besthits. In other words, assume that the protein sequence A2 in the database A, the protein sequence C2 in the database C, the protein sequence D2 in the database D, and the protein sequence E2 in the database E satisfy the besthit condition for the query protein sequence B1.

Similarly, save the besthit that satisfies the condition to a file. At this time, the besthit (A2) retrieved from the database (A) containing the protein sequence already used as a query in the database in order of adjacent to the database used as a query in the prior search process is stored at the end. That is, they are stored in the order of (C2, D2, E2, A2).

[Table 2] below shows the relationship between query and beshit. As shown in [Table 1] above, it is expressed in a format for explanation and is not actual data.

TABLE 2

The similarity comparison is then performed for each protein sequence in the database (A, B, D, E) except for C, using the first protein sequence of the besthit when B1 is a query, that is, C2 as a query. do.

Table 3 below repeats these processes many times and ends the sequence similarity search when the protein sequence belonging to the database containing the protein sequence used as the initial query is placed in the first position. The files created according to the description so far and the contents thereof are briefly expressed as shown in [Table 3].

TABLE 3

File name data (QUERY) File name data (BESTHIT) mge1.doc A1 mge1.out B1, C2, D1, E2 mge2.doc B1 mge2.out C2, D2, E2, A2 mge3.doc C2 ⇒ mge3.out D1, E2, A1, B1 mge4.doc D1 mge4.out E2, A1, B1, C2 mge5.doc E2 mge5.out A1, B1, C2, D2

Next, two files are randomly selected from the files having the besthit protein sequence information (mge1.out, mge2.out… mge5.out), and the files with the query information (mge1.doc, mge2.doc…). mge5.doc) to identify relationships between at least three pieces of protein sequence information belonging to different databases.

The relationship between each piece of information appears in various forms. Therefore, in order to confirm the relationship between each piece of information, it exists in one of the two selected files.

It is necessary to determine whether any one besthit protein sequence is used as a query in another selected file, and to confirm the existence of a besthit protein sequence stored in both files in common.

First, assume that optionally mge1.out and mge2.out are selected. Determining whether or not any besthit protein sequence in the mge1.out file is used as a query for the selected mge2.out, and confirming the existence of the besthit protein sequence in common in both files. do. In this example, B1 is used as a query for mge2.out, and C2 and E2 are the besthits in common. At this time, A1, B1, and C2 become a relationship between a query and besthit. Similarly, A1, B1, and E2 also have a relationship between query and besthit. Thus, two protein sequences in which one protein sequence differs in relation between at least three protein sequence information belonging to different databases among the besthit protein sequence information and the query protein sequence information used to generate the besthit protein sequence information. Seeds are sequence information of proteins that satisfy a besthit or query relationship. This seed is stored in a separate file (A.doc).

4 is a schematic of seed formation. That is, B1, C2, D1, and E2 are besthit when A1 is a query, and A2, C2, D2, and E2 are besthit when B1 is a query. At this time, A1, B1, and C2 satisfy the seed formation conditions mentioned above. Similarly, A1, B1, and E2 also satisfy the seed formation conditions. 4 is for explaining the seed formed by the present invention is actually extracted by the operation of the program stored in the computer that meets the seed condition is stored in the file (A.doc).

Then find the seed in the same way for the rest of the files. That is, when the number of all output files (mge1.out, mge2.out… mge5.out) is “N”, seed formation is performed by comparing the two files as many times as NC2. Figure 5 illustrates the seed formation performed for all of these files.

In this example, there is an example in which a relation between an arbitrary besthit protein sequence stored in one of two selected files is used as a query of another selected file, but it may not be the case. As an example, there may be a protein sequence that is present in both files even when the relationship between any besthit protein sequence stored in one file of two selected files is not established as a query of another selected file. For example, according to another embodiment, the result shown in [Table 4] below is shown.

TABLE 4

File name data (QUERY) File name data (BESTHIT) mge1.doc A1 mge1.out B1, C1, E2 mge2.doc B1 mge2.out C2, D2 mge3.doc C2 ⇒ mge3.out D2, A1, B1 mge4.doc D2 mge4.out E2, B2 mge5.doc E2 mge5.out A1, D2

As described above, two arbitrary files are selected to check the correlation of the stored data. Among them, compare mge1.out and mge4.out. None of the besthits in mge1.out correspond to the query in mge4.doc for creating mge4.out. At this time, there may be besthit protein sequence information commonly present in both files (mge1.out and mge4.out). That is, E2 is in common. In this case, one or more S3-> S4-> S5 processes are performed by selecting each query protein sequence information (A1, D2) and one of these queries (A1) used to generate the two selected files. Thus, the selected query protein sequence information (B1, C2) and the common E2 until the other query D2 is selected satisfy the seed formation conditions. Thus, A1, B1, C2, D2, and E2 are seeded (see FIG. 6).

According to another embodiment of the present invention, there may be a case where a result value as shown in Table 5 below is provided. In this case, no relationship between any besthit protein sequence stored in one of the two selected files is used as a query for the other selected file.

There is no protein sequence present in both files.

TABLE 5

File name data (QUERY) File name data (BESTHIT) mge1.doc A1 mge1.out B1, C1, E1 mge2.doc B1 mge2.out C2, D1 mge3.doc C2 ⇒ mge3.out D2, A2, B2 mge4.doc D2 mge4.out E2 mge5.doc E2 mge5.out A1

Compare the files mge1.out and mge5.out corresponding to the above situation, in which case there is no common protein between the two files storing the besthit protein sequence information, but one query protein sequence (A1) and the other If the first besthit protein sequence (A1) in the file generated by the query protein (E2) of the same is the same, each of the query protein sequence information (A1, E2) and one of them (A1) by selecting the S3-> S4-> S5 process is performed one or more times to include all selected query protein sequence information (B1, C2, D2) in the result until the other query (E2) is selected. In this case, A1, E2, B1, C2, and D2 form seeds. This is shown in Figure 7 (S7 process).

When the seed found by the seed formation process is stored in one file (A.doc), the besthit protein sequence information therein may be called a homolog. At this time, the stored seed may have duplicate besthits, so that there is no overlap. Thereafter, there may be two or more besthits derived from the same microbial genome, so if the microbial name is duplicated, clean it up. For example, if there are 26 microbial names, it means that homologs are found in 26 microorganisms.

What has been described above relates to a method of searching for a target protein according to the present invention, that is, to find a protein sequence homologous to any one protein sequence in any one of the selected protein databases in another database.

Through the method of searching for a target protein using microbial genome information according to the present invention, it is possible to find homologous genetic factors common to each microorganism, and the protein encoded by these genes is expected to play an essential function for the survival of the organism. It can be a good target protein for drug development. This method can be used to efficiently find new target proteins for antibiotics / antifungals. It also has the effect of predicting the function of a protein whose function is unknown.

Claims (12)

Selecting a plurality of protein sequence databases (S1) from which a plurality of databases having protein sequence information of an organism want to know homology between protein sequences (S1); Selecting one of the protein sequences included in any one of the protein databases as a query of a sequence similarity search program (S2); Running a similarity search program to compare the similarities with respect to each of the databases except for the protein database of the organism including the query (S3); Protein sequence information (“besthit”) having the smallest cut-off value among the protein sequences satisfying a predetermined condition among the driving results of the similarity search program is stored according to the sequence similarity comparison procedure. Generating a computer file (S4); Selecting a protein sequence located first in a file in which the besthit protein sequence information is stored as a query of a similarity search program (S5); Performing steps S5, S3 and S4 a plurality of times in a sequence S5-> S3-> S4 (S6); Between at least three protein sequence information belonging to a different database among the besthit protein sequence information stored in the files generated by the S2 or S4 or S6 process and the query protein sequence information used to generate the besthit protein sequence information. (S7) finding these protein sequence information when the relation of any one protein sequence satisfies a besthit or query relation for two different protein sequences; A method of searching for a target protein using a similarity comparison of microbial genome sequence information comprising a step (S8) of extracting homolog information present in different protein databases using the protein sequence information found in the step S7. The method of claim 1, wherein the S7 process is: Identifying whether any one besthit protein sequence stored in the selected besthit protein sequence file is used as a query in another selected besthit protein sequence file; A method of searching for a target protein using a comparison of similarity of microbial genomic sequence information, comprising processing the protein sequence data comprising the step of confirming the existence of a besthit protein sequence stored in common in two files. The method of claim 2, wherein any one stored in the file selected in step S7 When the besthit protein sequence is used as a query for another selected file, the seed includes the protein sequence information common to both files and each query protein sequence information used to generate the selected two files. A target protein search method using similarity comparison of microbial genome sequence information. The seed of claim 2, wherein when the besthit protein sequence stored in one file selected in step S7 is not established as a query for another selected file, the seed includes protein sequence information common to both files. Target protein search method using a similarity comparison of microbial genomic sequence information, characterized in that. 5. The method of claim 4, wherein each of the query protein sequence information used to generate the two selected files and one of them is selected as a query, and at least one S3-> S4-> S5 process is performed to perform the other query. A method for searching for a target protein using a similarity comparison of microbial genome sequence information, comprising seeding of all selected query protein sequence information until being selected. 5. The method of claim 4, wherein if one query protein sequence is the same as the first besthit protein sequence in a file generated by the other query protein, each of the query protein sequence information and one of the queries is selected to be S3. A method of searching for a target protein using a similarity comparison of microbial genomic sequence information, characterized in that the seed includes all the selected protein sequence information in the seed until one or more other queries are selected. . The method (S8) of claim 1, wherein the homolog extraction is performed: Storing the protein sequence information searched by the S7 process into a group, but overlapping the protein sequence information in one file in a calculation method of a union (U) leaving only one; A method of searching for a target protein using similarity comparison of microbial genomic sequence information, characterized in that the step of calculating the number of microorganisms that are the source of protein sequence information stored in one file. The method according to any one of claims 1 to 7, If the sequence similarity search program is BLAST, e-value <0.01 value target protein search method using a similarity comparison of the microbial genomic sequence information, characterized in that for adopting. The method according to any one of claims 1 to 7, Target protein search method using a similarity comparison of the microbial genomic sequence information, characterized in that when using the FASTA as a sequence similarity search program Z-SCORE> 90 is adopted. Selecting a plurality of protein sequence databases (S1) from which a plurality of databases having protein sequence information of an organism want to know homology between protein sequences (S1); Selecting one of the protein sequences included in any one of the protein databases as a query of a sequence similarity search program (S2); Comparing the similarities with respect to each of the databases except the protein database of the organism including the query (S3); Protein sequence information (“besthit”) having the smallest cut-off value among the protein sequences satisfying a predetermined condition among the driving results of the similarity search program is stored according to the sequence similarity comparison procedure. Generating a computer file (S4); Selecting a protein sequence located first in a file in which the besthit protein sequence information is stored as a query of a similarity search program (S5); Performing steps S5, S3 and S4 a plurality of times in a sequence S5-> S3-> S4 (S6); The relationship between the besthit protein sequence information stored in the files generated by the steps S2 to S6 and at least three protein sequence information belonging to different databases among the query protein sequence information used to generate the besthit protein sequence information. Searching for protein sequence information in which any one protein sequence has a besthit or query relationship for two different protein sequences (S7); A series of instructions for executing an operation consisting of a process of extracting homolog information present in different protein databases (S8) using the protein sequence information found in the S7 process. A computer-readable recording medium storing a program comprising a. A computer system comprising the recording medium of claim 10. 12. The recording medium of claim 11, wherein the recording medium is: Computer system, characterized in that selected from the group of recording media consisting of a hard disk (optical memory), optical memory (Random Access Memory), ROM (Read Only Memory) and Flash Memory (Flash Memory) .