JP3353263B2 - Gene motif extraction processing apparatus and processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a gene motif extraction processing apparatus and method, and more particularly to a gene motif for extracting a motif that is a conserved site between sequences by comparing a plurality of pieces of given gene sequence information. The present invention relates to an extraction processing device and a processing method. With the progress of genetic engineering in recent years, D
Gene sequence information databases represented by NA sequences and amino acid sequences are rapidly increasing. Attempts to elucidate all the gene sequences of specific organisms, such as the Human Genome Project, have been made worldwide, and it is expected that the gene sequence information will increase rapidly in the future.

[0002] Among these gene sequences, sequence information has been clarified, but many of their functions and structures are unknown. The functions and structures of such genes are
As an effective method for predicting from the sequence information, there is a search for a motif having a regularity characteristic of the sequence. For that purpose, a technique for extracting many motifs from known sequences is required.

[0003]

2. Description of the Related Art Hitherto, a motif that specifies a gene function in a gene sequence and has a characteristic regularity in the sequence has been determined based on experiments and reports in the literature. PROSII is a database that registers such motifs.
TE is known. Incidentally, it is generally known that functionally important sites (sites) in a gene sequence are hard to change. By utilizing this fact, a motif can be extracted as a conserved region by comparing a plurality of gene sequences. However, a technique for extracting a motif from a comparison of gene sequences has not been established.

[0004]

Determining a motif artificially by experiment or the like is a difficult task. Therefore,
If a motif can be mechanically extracted from the comparison of gene sequences, it is considered that a lot of information effective for elucidation of gene function and the like can be obtained. However, when a method of simply comparing each site of a plurality of gene sequences and examining the similarity of each site is adopted, there are the following problems.

[0005] In other words, when a plurality of gene sequence information to be extracted is biased toward a specific kind of organism, the regularity to be extracted is biased. For example, human gene sequence information, monkey gene sequence information, horse gene sequence information, ...
If you try to extract a motif from the similarity of each site in a sequence information group that has a lot of gene sequence information of higher organisms such as However, it is not always possible to identify a storage area that has not changed much in, and there is a possibility that the identification of a storage area to be extracted as a motif may be erroneous. The same applies to the reverse case.

An object of the present invention is to solve the above-mentioned problems and to mechanically (automatically) extract a motif based on a plurality of pieces of gene sequence information.

[0007]

FIG. 1 is a diagram showing a configuration example of the present invention. In FIG. 1, reference numeral 10 denotes a processing device including a CPU, a memory, and the like. The sequence information input means 11
This is a means for inputting alignment data of a plurality of gene sequences from which a motif is to be extracted. Phylogenetic tree creation means 12
Is a means for creating an evolutionary phylogenetic tree 13 based on the degree of difference between gene sequences based on alignment data of a plurality of gene sequences input by the sequence information input means 11. The phylogenetic tree 13 may be created in advance using, for example, paleontological information.

The array weight calculator 14 calculates the weight of each array from the length of the branch of the phylogenetic tree 13. The score calculating means 15 calculates, for each site in the array, a score indicating the degree of similarity of the array element appearing in that site, based on the weight of the array and the similarity of the element determined in advance according to the type of the array element. This is means for calculating based on the gender-based score table 16.

The feature information extracting means 17 is a means for extracting a portion having a characteristic regularity in the gene sequence as a motif based on the calculated score, and outputting the motif to an output device 18 such as a display or a printer. In particular, when the value of the score calculated by the score calculating means 15 exceeds a predetermined threshold value or a set threshold value, the characteristic information extracting means 17 extracts that part as a motif part. Further, the feature information extraction means 17 calculates the appearance rate of the motif region with a predetermined continuous region width or a set continuous region width, and when the value exceeds a predetermined random level or a set random level, A continuous region is defined as a motif region, and an adjacent motif region is defined as one motif region. The calculation results and the like are output to the output device 18.
The output device 18 plots a motif appearance rate, a random level value, and the like, and outputs a graph display.

[0010]

The gene motif extraction processing apparatus of the present invention uses multiple alignment data as input data, and outputs amino acids highly conserved in the sequence constituting the alignment data at each site as motifs.
However, a phylogenetic tree 13 is created based on the alignment data in order to correct the bias in the frequency of appearance of amino acids due to the presence of sequences closely related to the evolution, and each gene sequence is determined based on the branch length and shape of the phylogenetic tree 13. Is weighted. Further, in order to allow the appearance of amino acids having similar properties, a score at each site is calculated using the score table 16 calculated based on the similarity of amino acids. A higher score determined here indicates that the amino acid is highly conserved at that site.

Further, in order to extract a motif part, an operation of setting a score threshold is performed. In addition, a part showing a score exceeding the threshold set here is extracted as a motif part. Then, in order to limit the motif region, an operation of setting a region width and a random level is performed. If the appearance rate of the motif region within the region width set here exceeds a random level, the region is regarded as a motif region. Adjacent motif regions are regarded as one motif region.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described with reference to the drawings, using gene sequence information represented by an amino acid sequence as an example. FIG. 2 is a processing flowchart of the embodiment of the present invention. The processes (a) to (k) in the following description correspond to the processes (a) to (k) shown in FIG.

Consider alignment data consisting of five gene sequences A to E shown in FIG. One letter of the alphabet corresponds to one amino acid, and * in the sequence length indicates a gap. (A) Alignment data input The sequence information input means 11 inputs the alignment data of the sequences A to E shown in FIG. When there are a large number of sequences having very similar sequence information, a bias occurs when amino acids representing each site are determined from their appearance frequencies. Therefore, in the following processing, a phylogenetic tree 13 is created from the input alignment data, and weighting for each gene sequence is calculated based on the branch length and shape of the phylogenetic tree 13. The bias is corrected by weighting each array using the calculation result.

(B) Creation of a phylogenetic tree The creation of a phylogenetic tree 13 by the phylogenetic tree creating means 12 uses, for example, the UPG (Unweighted Pair-Group Clustering) method. Other creation methods may be used. In the present embodiment, specifically, a phylogenetic tree is created as follows. First, the degree of difference between gene sequences is determined based on the alignment data. The degree of difference is calculated as the number of amino acid substitutions between pairs of two sequences. The following formula (1), which is generally used when calculating the number of amino acid substitutions, is used for the calculation.

K = −log (1-p) (1) where K is the number of amino acid substitutions, and p is the ratio of sites having different amino acids between the two sequences. Also, a portion including a gap is excluded from the calculation. According to equation (1), a set of all two arrays, ie, (A, B), (A,
C), ... (A, E), (B, C), ..., (C,
D), (C, E) and (D, E) are calculated for the difference. Also, the degree of difference is calculated as V _AB , V _{AC ,.}
, The smallest one of the degrees of difference V _AB , V _AC ,..., V _DE is selected, and the set is connected. In this example, the arrays D and E are linked. This difference is defined as the length of the branch.

Next, array D and array E are grouped into one group, and the degree of difference between these arrays and each of the other arrays is similarly calculated by equation (1). For example, the degree of difference V _{(DE) A} between the arrays D and E and the array A is obtained by V _{(DE) A} = ( _VAD + _VAE ) / 2. Similarly, V _{(DE) B} and V _{(DE) C} are calculated, and have the smallest value from those obtained by removing V _DE from V _AB , V _{AC ,.} Choose one. In this example, the dissimilarity V _AB between array A and array B is the smallest, and these are grouped second. Hereinafter, grouping and calculation of the degree of difference are similarly performed, and a phylogenetic tree 13 is created based on the results.

FIG. 4A shows an example of a phylogenetic tree 13 constructed based on the alignment data shown in FIG. 3, and the numbers in parentheses in the figure indicate the length of each branch in the phylogenetic tree 13. Represents. (C) Calculation of Weight for Each Gene Sequence Next, the sequence weight calculation means 14 calculates the phylogenetic tree 1
Weight is given to each branch based on the length of each branch of No. 3. The weight given to each branch is obtained by dividing the length of the branch by the number of arrays branched from the branch.

A description will be given by taking the phylogenetic tree 13 of FIG. 4A as an example.
Branch 1 has a length of 0.158, and three sequences of sequence A, sequence B, and sequence C branch from branch 1. Therefore, branch 1
Is 0.158 / 3 and is calculated as 0.053. Similarly, when the weights of all the branches are obtained, the result is as follows. Weight of branch 1 = 0.158 / 3 = 0.053 Weight of branch 2 = 0.903 / 2 = 0.452 Weight of branch 3 = 0.367 / 2 = 0.184 Weight of branch 4 = 0.745 Weight of branch 5 = weight of branch 6 = 0.378 / 2 = 0.189 Weight of branch 7 = weight of branch 8 = 0.000 The weight of each array is calculated based on the weight of each branch thus obtained. .

The weight given to each array is obtained as the total weight of the branches that pass when going back from the root of the phylogenetic tree 13. In the example of the phylogenetic tree 13 in FIG. Sequence A passes through branch 1, branch 3, and branch 5 on the phylogenetic tree 13. The weight given to each branch is 0.053,
0.184 and 0.378. Therefore, the weight of array A is determined to be 0.615 in total. Similarly, the weights for all the arrays are calculated. Further, the sum of the weights of all the arrays is obtained, the weight of each array is divided by the sum, and the sum is standardized so that the sum of the weights becomes 1. FIG.
(B) shows each sequence A determined from the phylogenetic tree 13 in FIG. 4 (A).
ＥE indicate weights.

(D) Weight calculation for each amino acid Next, the sequence weight calculation means 14 calculates the weight of each amino acid at each site based on the standardized weight of each of the sequences A to E. Then, the weight of the amino acid appearing for each site is determined as the sum of the weights of all the sequences in which the amino acid appears.

The alignment data shown in FIG. 3 and FIG.
A description will be given based on the weight of the array in (B). At the first position, amino acids of sequence A (Q (glutamine), sequence B (L (leucine)), sequence C (E (glutamic acid)), and sequences D and E (S (serine)) appear. Therefore, at the first position, the weight of amino acid Q
Similarly, the amino acid L is given a weight of 0.210, the amino acid E is given a weight of 0.272, the amino acid S is given a weight of 0.308 by the sum of the weights of the sequence D and the sequence E, respectively. . Other amino acids have a weight of 0 at the first position.

Similarly, at the second site, sequence A,
Amino acid V (valine) appears in B and C, and amino acid A (alanine) appears in sequences D and E. The weight of amino acid V at this site is the sum of the weights of sequences A, B, and C, and is 0.210 + 0.210 + 0.272 = 0.592. The weight of amino acid A at the second site is 0.154 + 0.154 = 0.308.

Similarly, the weight of each amino acid is calculated at all the sites. FIG. 5 shows the amino acid weights from the first site to the tenth site calculated as described above. FIG. 5 shows the calculation result up to the fourth digit after the decimal point. FIG. 6 is a flowchart showing one embodiment of the processing (d) shown in FIG. 2 performed by the array weight calculator 14.

In FIG. 6, a step 31 initializes a part position counter of the array weight calculating means 14, ie, in the CPU of the processing device 10. A step 32 initializes a gene sequence number counter of the sequence weight calculation means 14, that is, in the CPU of the processing device 10. Step 33 is
The amino acid of the current gene sequence at the current site position is stored in the current amino acid type storage area of the sequence weight calculation means 14, that is, in the memory of the processing device 10. In the following description, the current site position, the current gene sequence, the current amino acid type, and the like respectively indicate the position of the current site of interest, the current gene sequence of interest, the current amino acid type, and the like. Step 34
The weight of the current gene sequence is added to the score stored in the score storage area of the current amino acid type at the current site position in the sequence weight calculation means 14, ie, in the memory of the processing device 10, and stored. A step 35 adds 1 to the gene sequence number counter.

In step 36, it is determined whether or not the value of the gene sequence number counter is larger than the number of genes. If the determination result is NO, the process returns to step 34. On the other hand, if the decision result in the step 36 is YES, a step 37 performs a standardization process so that the sum of the scores of each amino acid at the current site position becomes 1. Next, a step 38 calculates a score for each part, and a step 39 adds 1 to the part position. A step 40 decides whether or not the value of the part position counter is larger than the alignment data length. If the decision result in the step 40 is YES, the process ends. On the other hand, if the decision result in the step 40 is NO, the process returns to the step 32.

(E) Calculation of score at each site Depending on the sequence, substitution to an amino acid having similar properties may occur, but even in such a case, it is often functionally conserved. Therefore, in order to extract such a site as a motif, the score calculation unit 15 calculates a score of each site based on a score table based on physicochemical similarity between amino acids.

Score table 16a based on amino acid similarity
Is determined in advance based on the physical and chemical properties of each amino acid.Based on the substitution frequency of each amino acid set and the distance indicating the degree of difference in properties, each amino acid set is It is a table with the values given by For example,
The score of the set of glycine (G) and other amino acids is given the following values. However, in this case, each score is multiplied by 100 for convenience.

[0028] Since various types of such score tables are known, the description here will be limited to this level.

The score of each site calculated in consideration of the value in the score table 16a indicates that the larger the value, the more conservative the amino acid is at that site. For example,
The score of the first part in FIG. 5 is obtained by the following equation (2). S ₁ = D (S, S) × S (S) × S (S) + D (S, L) × S (S) × S (L) + D (S, E) × S (S) × S (E ) + D (S, Q) × S (S) × S (Q) + D (L, S) × S (L) × S (S) + D (L, L) × S (L) × S (L) + D (L, E) × S (L) × S (E) + D (L, Q) × S (L) × S (Q) + D (E, S) × S (E) × S (S) + D (E , L) × S (E) × S (L) + D (E, E) × S (E) × S (E) + D (E, Q) × S (E) × S (Q) + D (Q, S ) × S (Q) × S (S) + D (Q, L) × S (Q) × S (L) + D (Q, E) × S (Q) × S (E) + D (Q, Q) × S (S) × S (Q) Formula (2) Here, S ₁ is obtained from the score of the first site, and D (amino acid 1, amino acid 2) is obtained from the score table 16a for amino acid 1 and amino acid 2. Was Acetonide, is S (amino acid) is the weight of the amino acid at that site (Figure 5).

FIG. 7 is a diagram showing the operation of the score calculating means 15 shown in FIG.
6 is a flowchart illustrating an example of a process (e) illustrated in FIG. In FIG. 7, step 51 is performed by the score calculation unit 15.
That is, the score storage area of the current site position in the memory of the processing device 10 is initialized. A step 52 initializes an amino acid type counter of the score calculation means 15, that is, in the memory of the processing device 10. A step 53 initializes the comparison amino acid type counter of the score calculation means 15, that is, in the memory of the processing device 10. In step 54, the similarity between the current amino acid species and the current comparative amino acid species is obtained with reference to the amino acid similarity score table 16a. Step 55 is as follows: S _i = S _i + D (A ₁ , A ₂ ) × S (A ₁ ) × S
(A ₂ ) Here, _Si is the current position (i-th position)
, A ₁ is the current amino acid type, A ₂ is the current comparative amino acid type, D (A ₁ , A ₂ ) is the score of the similarity between the current amino acid type and the current comparative amino acid type, and S (A ₁ ) is the current site position. , S (A ₂ ) indicates the score of the current comparative amino acid species at the current site position.

Step 56 changes the comparative amino acid species to the next comparative amino acid species, and step 57 determines whether or not all amino acid species have been compared. Step 5
If the decision result in the step 7 is NO, the process returns to the step 53. On the other hand, if the decision result in the step 57 is YES,
In step 58, the amino acid type is changed to the next amino acid type. A step 59 decides whether or not scores have been calculated for all amino acid species, and the process ends if the decision result in the step 59 is YES. On the other hand, if the decision result in the step 59 is NO, the process returns to the step 52.

(F) Output of Calculation Result The score calculation means 15 calculated the score for the alignment data shown in FIG. (Site 01-05) 0.5183 0.7744 0.5677 0.8198 0.4881 (site 06-10) 0.9328 0.4940 0.8683 0.3165 0.3580 (site 11-15) 0.9311 0.3834 0.4072 0.3611 0.6114 (site 16-20) 0.6937 0.5976 0.5699 0.5574 0.5010 (site 21-25) 0.3880 0.6168 0.5530 0.5739 0.6296 (parts 26 to 30) 0.7718 0.3473 0.3772 0.6956 1.0000 (parts 31 to 35) 0.9841 0.9646 1.0000 0.9149 0.8891 (parts 36 to 40) 1.0000 0.6916 0.7864 0.7804 0.7903 (parts 41 to 45) 0.5830 0.6021 0.7753 0.5654 0.6976 ( (Parts 46 to 50) 0.9037 0.6428 0.8303 0.9542 0.7105 (g) Threshold value setting The characteristic information extracting means 17 determines a threshold value for a score, and extracts a part having a score exceeding the threshold value as a motif. Therefore, the threshold value is set by a value specified in advance by the user or as a default value.

(H) Using a portion exceeding the threshold as a motif When the threshold of the output score is th, the feature information extracting means 17
A part satisfying the following condition is extracted as a motif candidate. S> th The motifs extracted for the alignment data shown in FIG. 3 with the score threshold th set to 0.90 were as follows.

"30 D [LI] [IM] L [LIF]
[KRH] L ”Here,“ 30 ”means that the position of the first amino acid of the motif in the alignment data in FIG. 3 is 30. [] Indicates that a plurality of amino acids in [] appear at that site. That is, the extracted motif sites are the sixth (F or Y), the eleventh (F or Y), the thirty (D),
31st (L or I), 32nd (I or M), 3
3rd (L), 34th (L or I or F), 36
(L), 46th (I or V) and 49th (L
Or I or M).

FIG. 8 is a diagram showing motif sites extracted according to the embodiment of the present invention. By the way, conventionally, a motif which is a sequence pattern characteristic of a gene sequence could be extracted to some extent by manual operation using a motif as a site, but it was difficult to identify the motif as a region. However, when a functional region is identified or an ancestral gene is estimated, it is very important to identify a motif as a region. Therefore, in the present invention, a method for identifying a motif sequence extracted as a site as a region will be described in more detail.

(I), (j), (k) Setting of area width and random level, calculation of appearance rate of motif part, output of motif area FIG. 9 is shown in FIG. It is a flowchart which shows one Example of a process (i), (j), (k). The present embodiment generally includes three processes. First
In the process (1), an arbitrary region width is set, and the appearance rate of the motif portion within the region width is obtained. In the second process, a random level for determining whether or not the appearance rate of the motif portion within the set region width is sufficiently high is determined. A motif site within the region width is identified as one motif region. In the third process, when the identified motif regions are continuous, they are combined into one motif region.

That is, more specifically, the following processes S1 to S1
S6 is repeated. S1: Extract a motif site. S2: Set the initial area width, expansion width, and maximum expansion width. Further, the region width for obtaining the random level of the appearance rate of the motif part is set to the maximum expansion width. However, if the maximum extension width exceeds half the alignment data length, the area width of the random level is set to a value not exceeding half the alignment data length. S3: Calculate and plot the appearance rate of the motif site in each of the initial region width and the random level region width. S4: If the appearance rate of the motif part in the initial area width exceeds the appearance rate of the motif part in the random level area width, the initial area width is regarded as the “motif area”. S5: When both the appearance rates of the adjacent motif regions having the initial region width are “motif regions”, these are combined and regarded as one “motif region”. S6: Steps S4 and S5 are repeated over the entire length of the alignment data.

To explain the motif region identification processing with reference to FIG. 9, step 61 sets a region width for obtaining the motif site appearance rate. A step 62 initializes a part position counter of the characteristic information extracting means 17, that is, in the CPU of the processing device 10. In step 63, the appearance rate of the motif part within the set area width is calculated from the following equation with the current part position as the center.

(Occurrence rate of motif part) = (Number of motif parts in area width) / (area width) Step 64 plots the appearance rate of the motif part on a graph. Is calculated and plotted on a graph. Step 66 is
One is added to the part position, and a step 67 determines whether or not the value of the part position counter is larger than the alignment data length. If the decision result in the step 67 is NO, the process returns to the step 63.

On the other hand, if the decision result in the step 67 is YES, a motif area flag of the feature information extracting means 17, that is, in the CPU of the processing device 10, is initialized in a step 68. A step 69 initializes a part position counter. Step 70 determines whether or not the motif appearance rate at the site position is higher than the random level. Step 7
If the determination result of 0 is YES, the process proceeds to step 71; otherwise, the process proceeds to step 75.

A step 71 decides whether or not the motif area flag is set (set). If the decision result in the step 71 is YES, a step 72 sets the area centered on the current position as the current motif area. In addition, it extends. On the other hand, if the decision result in the step S71 is NO, a step 73 sets a motif area flag, and the step 74 sets a motif area flag.
The part position where the motif part first appears in the area width centered on the current part position is defined as the start part of the current motif area. After performing Step 72 or 74, the process proceeds to Step 78.

A step 75 decides whether or not the motif area flag is set. If the decision result in the step 75 is NO, the process proceeds to a step 78. On the other hand, if the decision result in the step 75 is YES, a step 76 initializes the motif area flag, and the step 77 outputs the current motif area. After performing step 76, the process proceeds to step 7
Proceed to 8.

In step 78, 1 is added to the position of the part.
A step 79 decides whether or not the value of the part position counter is larger than the alignment data length.
If it is ES, the process ends. On the other hand, if the decision result in the step 79 is NO, the process returns to the step 70.
An experimental result when the above-described motif region identification processing is performed will be described below.

In the experiment, a motif region was identified using FLAA7A-1 as a probe and alignment data as a target. 10 to 12 show that the probe name is FLA
A7A-1, the number of homologues is 53, the initial region width is 21, the maximum extension width is 101, the alignment data length is 97, the region width for obtaining the random level is 41,
The experiment result when the setting value at the time of motif part extraction is 0.90 is shown. FIG. 10 shows the proportion of the motif region in the set region width, where “o” indicates a plot with the initial value of the motif region width, and “...” Indicates a random level plot. If the plots overlap, the plot with the higher ratio is preferentially plotted. FIG.
Indicates an extracted motif site. Further, FIG.
The motif region obtained by the motif region identification processing is shown. In FIG. 12, “:” indicates the start position and end position of the motif, “[]” indicates that a plurality of amino acids appear at the motif site, and “−” indicates an arbitrary amino acid or gap at the site. (That is, a site that is not a motif site).

FIGS. 13 to 15 show ECODH in the same manner.
Fig. 4 shows other experimental results obtained using FOLG-1 as a probe. 13 to 15 show that the probe name is ECOD.
The experimental results when HFOLG-1, homologue, the number is 11, the initial region width is 11, the alignment data length is 179, the region width for obtaining the random level is 81, and the setting value when extracting the motif site are 0.90 are shown. Show. 13 and 14 show plots of the same alignment data divided, and FIG. 15 shows a motif region obtained by the motif region identification process. FIG.
3 and FIG. 14 show the proportion of the motif site in the set region width in association with the alignment data, where “o” is a plot of the initial value of the motif region width,
That is, the appearance rate of the motif portion in the set region width is shown, and "..." indicates a plot of the random level. That is, when the appearance rate of the motif site is lower than the random level, the motif site is not regarded as a motif region. If the plots overlap, the plot with the higher ratio is preferentially plotted. 13 and FIG.
The name shown on the left of the alignment data is
The entry name of the gene sequence registered in the gene sequence database DDBJ is shown. Also, Dihydrofo
late reductase signature
[LIF] -GX (4)-[LIVMF] -PW
Is data registered in the motif database PROSITE.

"122" shown on the left side in FIG.
And the like indicate the starting position of each motif region on the alignment data. "137" shown on the right side
And the like indicate the end position of each motif region on the alignment data. 16 to 18 show HUs in the same manner.
The other experimental result obtained using MTRX1-1 as a probe is shown. 16 to 18 show that the probe name is HUM.
The experimental results when TRX1-1, homologue, the number is 15, the initial area width is 11, the alignment data length is 110, the area width for obtaining the random level is 51, and the setting value at the time of extracting the motif part is 0.90 are shown. Show. FIG.
6 and FIG. 17 show plots for the same alignment data divided, and FIG. 18 shows a motif region obtained by the motif region identification process.

FIGS. 16 and 17 show the proportion of the motif site in the set region width in association with the alignment data. “O” is a plot of the initial value of the motif region width, that is, the set region width. The motif site appearance rate by width is shown, and "..." indicates a random level plot. That is, when the appearance rate of the motif site is lower than the random level, the motif site is not regarded as a motif region. If the plots overlap, the plot with the higher ratio is preferentially plotted. Furthermore,
In FIGS. 16 and 17, the name shown on the left side of the alignment data indicates the entry name of the gene sequence registered in the gene sequence database DDBJ. Also, T
hioredoxin family active
site [STA] -X- [WG] -C- [AGV]-
[PH] -C is a motif database PROSITE
This is the data registered in.

In FIG. 18, numerals such as "69" shown on the left side indicate the starting positions of the respective motif regions on the alignment data. The number such as “105” shown on the right side indicates the end position of each motif region on the alignment data. Thus, according to the present invention, a motif region can be mechanically (automatically) extracted and identified from gene sequence information, so that motif region extraction and identification can be performed at high speed. Therefore, it is easy to discover a new motif from a large amount of gene sequence data and create a motif database. Based on the motif information obtained in this way, it is possible to efficiently predict the function and structure of a gene sequence having an unknown function, and it is extremely useful to use the present invention to discover gene functions and identify functional regions. It is convenient.

The present invention has been described with reference to the embodiments.
The present invention is not limited to the above embodiments, and it goes without saying that various modifications and improvements can be made within the scope of the present invention.

[0050]

As described above, according to the present invention,
Since motif regions can be extracted and identified automatically (automatically) from gene sequence information, motif regions can be extracted and identified at high speed. New motifs can be discovered from a large amount of gene sequence data. Motif database can be easily created, and the function and structure of an unknown function gene sequence can be predicted efficiently based on the motif information thus obtained. It is very convenient to use the invention to discover gene functions and identify functional regions, and greatly contribute to the development of genetic engineering.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration example of the present invention.

FIG. 2 is a processing flowchart of an embodiment of the present invention.

FIG. 3 is a diagram showing an example of input alignment data.

FIG. 4 is a diagram showing a phylogenetic tree and sequence weight results obtained from the alignment data shown in FIG. 3;

FIG. 5 is a diagram showing calculation results of amino acid weights at each site.

FIG. 6 is a flowchart illustrating an example of a process performed by an array weight calculator.

FIG. 7 is a flowchart illustrating an example of a process performed by a score calculation unit.

FIG. 8 is a diagram showing motif sites extracted according to an example of the present invention.

FIG. 9 is a flowchart illustrating an example of a process performed by a feature information extracting unit.

FIG. 10 is a view for explaining an experimental result when a motif region identification process is performed (part 1).

FIG. 11 is a diagram for explaining an experimental result when a motif region identification process is performed (part 2).

FIG. 12 is a diagram illustrating an experimental result when a motif region identification process is performed (part 3).

FIG. 13 is a view for explaining an experimental result when a motif region identification process is performed (part 1).

FIG. 14 is a diagram illustrating an experimental result when a motif region identification process is performed (part 2).

FIG. 15 is a diagram illustrating an experimental result when a motif region identification process is performed (part 3).

FIG. 16 is a diagram illustrating an experimental result when a motif region identification process is performed (part 1).

FIG. 17 is a diagram for explaining an experimental result when a motif region identification process is performed (part 2).

FIG. 18 is a diagram illustrating an experimental result when a motif region identification process is performed (part 3).

[Explanation of symbols]

 Reference Signs List 10 processing device 11 sequence information input means 12 phylogenetic tree creating means 13 phylogenetic tree 14 array weight calculating means 15 score calculating means 16 score table based on similarity of elements 17 feature information extracting means 18 output device

Continued on the front page (72) Inventor Kazuho Ikeo 1111 Yata, Mishima City, Shizuoka Prefecture Inside the National Institute of Genetics (72) Inventor Yuichi Kawanishi 1015 Uedanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Inside Fujitsu Limited (72) Inventor Kawai Masato 1015 Kamikodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Inside Fujitsu Limited (56) References Mol. Evol. , February 1994, 38, p. 188-203 Mol. Biol, June 24, 1994, 239, p. 698-712 bits, September 1993, 25 [9], pp. 87-94 (58) Fields investigated (Int. Cl. ⁷ , DB name) C12N 15/00-15/90 G06F 17/30 PubMed JICST File (JOIS) INSPEC (DIALOG)

Claims (9)

(57) [Claims] 1. A processing device for extracting a characteristic regularity specifying a gene function from gene sequence information, wherein the sequence calculates a weight of each sequence from the length of a branch of an evolutionary phylogenetic tree regarding a plurality of gene sequences. a weight calculation hand stage, for each site of the array, and score calculation means to calculate a score indicating the degree of similarity of array elements appearing at that site using the weight, based on the calculated score and a feature information extraction means to extract a portion having a characteristic regularities in gene sequence motif, gene motif extraction apparatus. 2. A motif extracting apparatus of a gene of claim 1, further phylogenetic tree creation means to create a phylogenetic tree based on the degree of difference between the gene sequences on the basis of the alignment data of a plurality of gene sequences Equipped with a gene motif extraction processing device. 3. A motif extracting apparatus of a gene according to claim 1 or 2, wherein said scoring hands stage, the weight of each sequence are calculated Te <br/> by the weight calculation hand stage of the sequence, and means for calculating a score based on a similarity Seijo paper between array elements are determined according to the type of pre-array elements, gene motif extraction apparatus. 4. A motif extracting apparatus of a gene of any one of claims 1 to 3, wherein the feature information extraction hand stage, the value of the score calculation hand stages thus calculated score is given A gene motif extraction processing device, comprising: means for extracting and outputting a portion as a motif portion when a threshold value or a set threshold value is exceeded. 5. A motif extracting apparatus of a gene according to claim 4, wherein the feature information extraction hand stage, the appearance rate of the motif sites were calculated by a predetermined continuous area width or the set continuous region width, that value When a predetermined random level or a set random level is exceeded, a continuous area is extracted as a motif area, and when both adjacent areas are motif areas, means for outputting those areas as one motif area is provided. Further, a gene motif extraction processing device. 6. The motif extracting apparatus of genes according to claim 5, the characteristic information of the feature information extraction hand stage or found by displaying the graph by plotting the values of the incidence and random level of at least the motif part A gene motif extraction processing device further comprising an output device for outputting. 7. A processing method for extracting a characteristic regularity specifying a gene function from gene sequence information by a computer, wherein the sequence information input means inputs alignment data of a plurality of gene sequences to be extracted. A processing step , wherein the phylogenetic tree creating means creates an evolutionary phylogenetic tree based on the alignment data, and an array weight calculating means calculates the weight of each array from the length of the branch in the phylogenetic tree A process in which a score calculating unit calculates a score of each part based on the calculated weight of each array and similarity information between array elements which is previously determined according to the type of the array element. and process, the feature information extraction unit, when the value of the calculated score exceeds a predetermined threshold or set threshold, and a process of extracting the site motif sites heritage Motif extraction processing method of the child. 8. The method for extracting a motif of a gene according to claim 7, wherein in the extracting step, the feature information extracting means determines the appearance rate of the motif portion at a predetermined continuous region width or a set continuous region width. When the calculated value exceeds a predetermined random level or a set random level, the continuous region is extracted as a motif region, and when both adjacent regions are motif regions, those regions are regarded as one motif region. A gene motif extraction processing method, further comprising a processing step of outputting as a region. 9. The motif extraction method of gene according to claim 8, process the output device, as the motif the area for displaying the graph by plotting the values of the incidence and random level of at least the motif part, said extraction A method for extracting a motif of a gene, further comprising the step of outputting feature information from the motif.