JP3437521B2 - Apparatus and method for correcting frame shift error of cDNA sequence, and recording medium recording program for executing the method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus and method for correcting a frame shift error in a cDNA sequence, and a recording medium recording a program for executing the same.

[0002] Based on the related art many assumptions protein is estimated from the rough draft genome sequence, attempts to sequence a large scale began for mice full-length cDNA containing the complete CDS information. Calling error of base in CDS, especially CDS
Frame shift error in rough draft cDN
It is a significant problem in deducing amino acid sequences based on A or genomic sequences. The final determination of these sequences is a time consuming task. Therefore, information about the position of the putative frameshift error is very important for primer design in the final work. In addition, rough drafts of amino acid sequences are useful for gene classification and data exploration. Several methods have been developed to examine and prevent frame shift errors due to sequencing errors.

However, most computer software that corrects frame shifts in DNA sequences was developed for genomic DNA analysis and not for full-length cDNA sequences. In particular, attempts at full-length cDNAs such as the RIKEN project require high quality correction of rough drafts of cDNA sequences in order to minimize sequence information of amino acids from DNA sequences. Conventional software is designed to correct frameshifts based on the information of textual base analysis of DNA sequences. However, not only textual base information, but also information on the quality of the sequence from sequencing experiments is very important for unambiguous correction.

It has recently been stated that the quality of a sequence can be indicated by a numerical score, or fred. Such a method for cDNA was developed for genomic analysis (White,
O. et al., (1993) Nuc. Acids Res. 21, 3829-3838; Richte
rich, P. (1998) Genome Res.8,251-259; Medigue, C. et a
L., (1999) Genome Res. 9, 1116-1127). The information obtained by these methods can be used to suggest the location of frame shift errors (Ewing, B. et al., (199
8) Genome Res. 8,175-185; Ewing, B., And Green, P. (199
8) Genome Res.8, 186-194), which was still insufficient in terms of accuracy. For this reason, correction of frame shift errors must rely on the naked eye and experience of the human eye, and accurate prediction and verification of amino acid sequences required a great deal of labor.

[0005]

SUMMARY OF THE INVENTION The present invention has been made to solve such problems.

That is, an object of the present invention is to provide a means for identifying the position of a frame shift in a cDNA sequence, predicting an accurate amino acid sequence, and verifying the amino acid sequence.

The first aspect of the present invention is an apparatus for correcting a frame shift error in a cDNA sequence, which comprises:
An input unit for inputting the sequence and the image data of the electrophoresis used for the sequencing of this cDNA sequence, and the input cD
A modified sequence creating unit that creates a modified cDNA sequence by generating 1 to 4 artificial mutations selected from the group consisting of insertion and deletion in the NA sequence, and a Va score of the input cDNA sequence. And a calculation unit for calculating the Va score of the corrected cDNA sequence based on the formula (I), Va = logP1 + logP2 (I) (wherein, , Pcodon (i) represents the probability that a more preferred codon selection will occur at codon position (i), i represents the codon position number, W1 represents the start codon position, and W2 represents the stop codon position. , Perr (j) represents the probability of a base call error at the position j of the artificial mutation, and D represents a proportional constant of 0 to 120.) An output unit that selects and outputs the sequence having the smallest Va score. It is a device including.

A second aspect of the present invention is a computer-readable recording medium in which a program for correcting a frame shift error in a cDNA sequence is recorded.
Procedure for inputting the sequence and the image data of the electrophoresis used for the sequencing of this cDNA sequence, and the input cDNA
1 selected from the group consisting of insertions and deletions in the A sequence
Procedure for creating a modified cDNA sequence by making ~ 4 artificial mutations, and calculating Va score of input cDNA sequence and Va score of modified cDNA sequence based on formula (I) Procedure and Va = logP1 + logP2 (I) (wherein , Pcodon (i) represents the probability that a more preferred codon selection will occur at codon position (i), i represents the codon position number, W1 represents the start codon position, and W2 represents the stop codon position. , Perr (j) represents the probability of a base call error at the position j of the artificial mutation, and D represents a proportional constant of 0 to 120.) and a procedure of selecting and outputting the sequence having the smallest Va score. It is a computer-readable recording medium in which a program to be executed is recorded.

A third aspect of the invention is a method of correcting frameshift errors in a cDNA sequence, the method comprising:
The sequence and the electrophoretic image data used for the sequencing of this cDNA sequence are prepared, and 1 to 4 artificial mutations selected from the group consisting of insertion and deletion are generated in the prepared cDNA sequence. Modified by
A sequence was prepared, and the Va score of the prepared cDNA sequence and the Va score of the corrected cDNA sequence were calculated based on the formula (I), and Va = logP1 + logP2 (I) (wherein, , Pcodon (i) represents the probability that a more preferred codon selection will occur at codon position (i), i represents the codon position number, W1 represents the start codon position, and W2 represents the stop codon position. , Perr (j) represents the probability of a base call error at position j of the artificial mutation, and D represents the proportional constant between 0 and 120). Selecting the sequence with the smallest Va score. Is.

According to the first aspect, the second aspect, and the third aspect of the present invention, the search work of the frame shift of the cDNA sequence is automated, and the accuracy of the search result is extremely high in comparison with the conventional method. Was excellent. Therefore, the labor of the frame shift search work, which is conventionally performed manually, is significantly reduced.

According to the first aspect, the second aspect, and the third aspect of the present invention, the search result of the frame shift is also quantified. Therefore, not only the accurate amino acid sequence can be predicted, but also the reliability of the prediction result can be known. Therefore, when rereading a nucleotide sequence for the purpose of sequence accuracy and conducting a re-experiment, which portion should be re-read intensively is indicated, which is advantageous in that the labor of the experiment is simplified.

[0012]

DETAILED DESCRIPTION OF THE INVENTION First Embodiment The origin of the cDNA sequence input in the input section is not particularly limited, and both cDNA derived from prokaryotic cells and those derived from eukaryotic cells are included in the input cDNA sequences. Cd input
The NA sequence is preferably a full length cDNA sequence. c
When the DNA sequence is derived from a eukaryotic cell, it preferably contains a region from the 5'terminal noncoding region to the 3'terminal noncoding region.

In the input section, the image data of electrophoresis used for determining the sequence of the input cDNA sequence is input. This data can be an image file of the electrophoresis result such as an SCF file or a trace file, and is used for calculating the probability of base call error at the position j of the artificial mutation, as described later. In the modified sequence generator of the device according to the present invention, 1 to 4 (preferably 1 to 2) artificial mutations selected from the group consisting of insertion and deletion are generated in the input cDNA sequence.

More preferably, a modification c containing one insertion
A DNA sequence and a modified cDNA sequence containing one deletion,
A modified cDNA sequence containing one insertion and one deletion is created. In this case, 2 × N sequences containing one artificial insertion or deletion are created, and 2 × N × (N-1) sequences containing a combination of artificial insertion and artificial deletion are created. (However, the number of base pairs of the input cDNA sequence is N). Therefore, the number of sequences created is 2N + 2
It is N (N-1).

The modified sequence preparation unit may be provided with a storage unit for temporarily storing the cDNA sequence in which the artificial mutation has occurred as a cDNA candidate group. It is also possible that all the candidate sequences are created in the modified sequence creation unit, the created sequence candidates are stored in the storage unit, and then Va of each candidate is calculated in the calculation unit.

In the calculation unit, the Va score of the input cDNA sequence and the Va score of the corrected cDNA sequence are calculated. Va can be calculated according to the above formula (I).
The Va score can preferably be calculated based on formula (II). Va = logP1 + logP2 + logPatg (II) (In the above formula, P1 and P2 have the same meanings as defined above, and Patg represents the probability of occurrence of a more preferable start codon position.) The Va score is preferably the formula (III). ) Can be calculated based on. Va = logP1 + logP2 + logPconsensus (III) (In the above formula, P1 and P2 have the same meanings as defined above, and Pconsensus represents the probability of a more preferred consensus sequence.) The Va score is more preferably the formula ( IV) can be calculated based on. Va = logP1 + logP2 + logPatg + logPconsensus (IV) (In the above formula, P1 and P2 have the same meanings as defined above, Patg represents the probability of occurrence of a more preferred start codon position, and Pkozak yields a more preferred consensus sequence. Represents the probability.) In formulas (I), (II), (III), and (IV),
Pcodon (i) is from Gribskov, M. et al., (1984) Nucl. Acids.
Res.12,539-549, and also Transterm (Tra
nsTerm) website (http: //biochem.otago.
ac.nz:800/Transterm/homepage.html) (Brown, CM et
al., (1993) Nuc. Acids Res. 21, 3119-3123; Dalphin, ME
et al., (1998) Nuc. Acids Res. 26, 335-337), and the probability data of codon preference can be selected according to the probability of codon preference, and preferably the probability data of codon selection can be selected according to the origin of the input cDNA sequence.

The cDNA sequence to be input is mammalian (for example,
Human, mouse, rat, cow, pig)
Pcodon can preferably be determined according to the occurrence probabilities listed in the table below.

[0018]

[Table 7] If the codon contains an indeterminate base ("N"), Pcodon can be set to 1/2.

Since Pkozak is not independent of Pcodon, the probabilities of equations (I), (II), (III), and (IV) are double calculated. Kozak consensus is one of the following codons
The second base, ie, the AnnATG belonging to the next codon
This is because the last G of G is included. Although the present inventor has applied the method according to the present invention to proteins of more than 30 amino acids, this double calculation has no significant effect on Va and the result of the double calculation is negligible.

Formulas (I), (II), (III), and (I
In V), Perr (j) represents the error probability of the jth base call and can be calculated based on the image data of the electrophoresis result used for the sequencing of the cDNA sequence.

Specifically, the formulas (I), (II) and (II
In I) and (IV), Perr (j) can be calculated according to the following formula. Per (j) = 10 ^−0.1 Q (j), Q (j) is a Fred score representing the reliability of the base call at the position j of the artificial mutation,
It can be calculated from the distance and height of the signal peak of each base and the distance from the nearest unassigned base (N), based on the electrophoresis image data used for sequencing the cDNA sequence. (B. Ewing, et al., G
enome Research 8: 175-185 (1998); B. Ewing and P. Gree
n Genome Research 8: 186-194 (1998)). Fred score calculation software is http: //bozeman.genome.wash
Available at ington.edu/ (University of Washington, USA).

Q (j) is also commercially available software corresponding to Fred (eg, BaseImagIR (Licor, ww
w.licor.com), Trace Tuner (Parcel, www.paracel.
com)) can also be used for the calculation.

The calculation unit of the device according to the present invention may include the above-mentioned software for calculating the Fred score.

Formulas (I), (II), (III), and (I
In V), D is 0 to 120, preferably 60 to 1
00 represents a proportional constant. Formulas (I), (II), (III),
In and (IV), Patg can be determined according to the occurrence probability of the start codon.

When the input cDNA sequence is derived from a mammal (eg, human, mouse, rat, cow, pig), Patg can be determined according to the probability of occurrence described in Table 2.

When the input cDNA sequence is derived from a species other than mammal, the occurrence probability data can be constructed and used according to the origin of the input cDNA sequence. Of course, even if the input cDNA sequence is derived from a source other than mammalian, Table 2
Appearance probability data of can be used.

[0027]

[Table 8] Table 2 shows which position of ATG is selected as the start codon, and the probability (P _atg ) of a random sequence indicating a more preferable position of ATG.
"N" is the start codon n from the 5th end of all frames
It means that it is the first item. 60. of the start codon
3% is the first ATG from the 5'end, and 18.1% of the start codon is the second ATG from the 5'end.
The 5'upstream ATG is preferred over the downstream. On the other hand, the 5'UTR of the test data set provided in the example below
Has an average length of 151.6 bp. For convenience, the number of ATG is applied, and the denominator of P _atg is an arbitrary number, here 7
Set to.

Formulas (I), (II), (III), and (I
In V), Pconsensus can be determined according to the appearance probability of the more preferred consensus sequence, and preferably the appearance probability data of the consensus sequence can be selected according to the origin of the input cDNA sequence.

Examples of consensus sequences include Kozak sequences and Shine-Dalgarno sequences.

If the input cDNA sequence is of mammalian origin (eg human, mouse, rat, bovine, porcine), Pconsensus can be the probability of a more preferred Kozak consensus sequence: Pkozak. Pkoz
ak is Kozak, M., Nucleic Acids Research, 15: 8125-8148
It can be determined according to the probability of occurrence of the preferred Kozak consensus sequence disclosed in (1987).

Pkozak can preferably be determined according to the table below.

[0032]

[Table 9] Table 3 shows the percentage of consensus around the start codon and the probability that a random sequence with a more favorable consensus (Pkozak) will occur. Consensus is Xnn
It is represented by atgY (X, Y = A, G, C, T), where X and Y are the bases in question. AnnatgG
Is the most preferred consensus of 16 alternatives that gives a random consensus.
A more favorable consensus occurs with a probability of 1/16.

In Table 3, A, G in the CDS region,
It was assumed that the proportions of C and T were all equal to each other. CD
The ratio of 4 bases in S (code sequence) is
A: G: C: T = 25.9%: 26.4%: 25.9
%: 21.8%. The ratio of 5'UTR is
A: G: C: T = 20.8%: 29.2%: 29.8
%: 20.3%. The 5'untranslated region (UTR) is actually rich in GC, but the proportions of A, G, C and T are almost equivalent in CDS.

When the input cDNA sequence is derived from a prokaryotic cell (eg Escherichia coli, thermostable bacterium), Pconsensus
Can be the probability that a more preferred Shine-Dalgarno consensus sequence occurs: Ps-d.

At the output section, the sequence most likely as a cDNA sequence is selected from these sequences based on the Va score. Specifically, the cDNA sequence candidate showing the smallest Va score is "a sequence more like a cDNA sequence".
Is selected as.

Second and Third Aspects The input section, the correction sequence creation section in the above-mentioned first aspect,
The arithmetic unit, the output unit, and, in some cases, the storage unit can be realized as a program module that operates on the computer system 20 as shown in FIG. A frame shift error correction program for a cDNA sequence including such a program module is a recording medium such as a floppy (registered trademark) disk or a CD-R.
It is recorded in an OM (Compact Disk-Read Only Memory) or the like and read by the computer system 20 to correct the frame shift error as described above.

As shown in FIG. 1, the computer system 20 includes a computer main body 21 housed in a casing such as a mini tower, a display device 22 such as a CRT (Cathode Ray Tube), and a recording output device. Printer 23
And a keyboard 24a and a mouse 2 as input devices
4b, a floppy disk drive device 26 for reading information from a floppy disk 31 as a recording medium, and a CD-ROM drive device 27 for reading information from a CD-ROM 32 as a recording medium.

When these configurations are shown as a block diagram,
As shown in FIG. 2, an internal memory 25 such as a RAM (Random Access Memory) and an external storage device such as a hard disk unit 28 are further provided in the housing in which the computer main body 21 is housed. In addition, cDNA
A floppy disk (recording medium) 31 in which the array frame shift error correction program is recorded is inserted into a slot of the floppy disk drive device 26 and installed in the computer main body 21 by a predetermined procedure, as shown in FIG. The recording medium for recording the program according to the present invention is not limited to the floppy disk 31 and may be C
In addition to the D-ROM 32, the internal memory 25, the hard disk unit 28, etc., a MO (Magnet Optic
al) Disc, optical disc, DVD (Digital Versatil)
e Disk) or the like.

Example 1 Here, an example of calculating the Va score of a virtual cDNA sequence will be shown. The following two open reading frames cDN
A sequence 5'-TnnATGCTATGAnnGnnAT
Assume GGAGTGA-3 ', which includes Sequence 1 and Sequence 2. Sequence 1 = TnnATGCTA Sequence 2 = GnnATGGAG There are 16 alternatives around the start codon, XnnAT
GY (where X and Y are four base pairs (A,
G, C and h and T))). TnnATGC is the rarest consensus around the start codon of the 16 sequences, and all other sequences are more than TnnATGC.
It can be the start codon. Therefore, Pkozak is equal to 1. On the other hand, GnnATGG is the second most probable consensus around the start codon among the 16 sequences. P
kozak is equal to 2/16. Sequence 1 CTA is a rare codon with an occurrence frequency of the 57th out of all 64 codons. GAG has the second highest probability of 64 codons. CTA's Pcodon is 57/64,
It is 2/64 in GAG. Sequence 1 is found downstream of Sequence 2, Patg of Sequence 1 is 1/7 and Sequence 2 is 2 /
7 Pkozak, Patg, and Pcodo for each frame
In the array 1, n is Pkozak = 16/16, Patg = 1/7, P
codon = 57/64 In array 2, Pkozak = 2/16, Patg = 2/7, Pco
don = 2/64.

The Va value of array 1 and array 2 is Va (array 1) = log (16/16) + log (1 /
7) + log (57/64) = − 0.89 Va (sequence 2) = log (2/16) + log (2 /
7) + log (2/64) = − 2.95. Therefore, it can be said that the array 2 is closer to the CDS than the array 1.

Example 2 The following is an example of repairing a cDNA sequence having a frame shift according to the present invention. A frameshifted cDNA sequence (test data set) was prepared by previously inserting / deleting in the complete cDNA sequence.

The test data set consisted of 612 known mouse complete CDS sequences downloaded from the NCBI nt database, and artificial insertions / deletions were created for each sequence as base call errors. A random number of generators creates a Fred score for each base,
The probabilities of insertions and deletions follow its Fred score. 1
A base call error of / 6 is a deletion and 1/6 is an insertion. Base call accuracy of test data is about 99%
Set to. Average length of cDNA in test dataset is 2
It is 370 bp. Table 4 shows the distribution of the number of insertions and deletions in CDS (rate of frameshift error). 14.4% of the sequences contain no frameshift error in the CDS and 85.6% of the sequences contain one or more frameshift errors in the CDS. This statistic does not include insertions / deletions in the UTR.

[0043]

[Table 10] The calculation of Va was performed according to the above formula (IV). Pcodon
(I) was determined according to Table 1.

Q (j) is a phred version 0.961028.m
(Obtained from http://bozeman.genome.washington.edu/)
Was calculated using.

Patg was determined according to Table 2. Pconsens
us was calculated as Pkozak. Pkozak was determined according to Table 3.

For each of the prepared cDNA sequences,
A modified cDNA sequence containing one artificial insertion, a modified cDNA sequence containing one artificial deletion, and a modified cDNA sequence containing one artificial insertion and one artificial deletion were created and input. The Va score of the cDNA sequence and the Va score of the cDNA candidate sequence were calculated.

The estimation accuracy was evaluated by the ratio of the deduced amino acid sequence showing homology higher than the threshold value to the correct amino acid sequence. Homology threshold is 100%, 98
%, 95%, 90% and 85%. D is
After trying various D's, it was set to 80 as the optimum value.

The results are shown in Table 5.

[0049]

[Table 11] Table 5 shows the estimation accuracy of several methods.

For the analysis of the estimation accuracy, Genscan ve
r1.0 (Burge, C. and Karlin, S. (1997) J. Mol. Biol. 2
68, 78-94; Burge, C. and Karlin, S. (1998) Curr. Opin. Str
uct.Biol.8,346-354) was used. DECORDER
^{In a} , the number of artificial frame shifts was limited to two. In DECORDER ^b the number of artificial frameshifts was limited to 4.

In Table 5, "longest frame" means that the longest frame is selected as the CDS for each test sequence without any data collection. Gensc
an was not developed for cDNA analysis,
It was developed for genome analysis. Even if this program is applied to cDNA, it will be a CDS rather than a simple selection of the longest frame.
Works even better for partial estimation. The inventor could not find any other suitable program for estimating the protein of a cDNA containing a frameshift error. Genscan works well with a putative promoter, but there is no promoter in the cDNA. In such applications, Genscan performance is very limited.

The program for carrying out the method according to the invention is called DECORDER. DECORDER
68-70% of the putative proteins according to (417 of 612 test samples) show greater than 85% identity with the correct protein. When the required estimation accuracy is greater than 85%, DECORDER shows the best estimation of the three methods. Although Genscan and DECORDER show almost equivalent results with an estimation accuracy of 85%, these estimation results show different percentages of well-estimated proteins. In the case of the estimation by DECOREDER, 35 to 36% of the estimated proteins are 9
Shows identity over 8%, 32 ~ 35% (= 68% ~
36% or 70% to 35%) of the estimated protein is 85
Shows ~ 98% identity. On the other hand, in the case of Genscan, 15% of the putative proteins show greater than 98% identity and 54% (= 69% -15%) of the putative proteins show 85-98% identity. Genscan and DEC
It can be combined with ORDER. Gensc
When an predicted protein based on a cDNA modified by DECORDER, 73% of the predicted proteins showed greater than 85% identity, the highest value at the 85% predicted level (data not shown).

[Brief description of drawings]

FIG. 1 is a perspective view showing a computer system using a computer-readable recording medium in which a frame shift error correction program for a cDNA sequence is recorded.

FIG. 2 is a block diagram showing a hardware configuration of the computer system of FIG.

[Explanation of symbols]

20 computer system 21 Computer body 22 Display 23 Printer 24a input device 24b mouse 25 Recording medium (internal memory) 26 Floppy disk drive 27 CD-ROM drive device 28 Recording medium (hard disk unit) 31 recording medium (floppy disk) 32 recording medium (CD-ROM)

Claims (28)

(57) [Claims] 1. An apparatus for correcting a frame shift error in a cDNA sequence, comprising: an input unit for inputting the cDNA sequence and the image data of electrophoresis used for the sequencing of the cDNA sequence; and the input cDNA sequence. A modified sequence generator for generating a modified cDNA sequence by generating 1 to 4 artificial mutations selected from the group consisting of insertion and deletion, and Va score and modification of the input cDNA sequence. CD
An operation unit for calculating the Va score of the NA sequence based on the formula (I), and Va = logP1 + logP2 (I) (wherein: Where Pcodon (i) represents the probability that a more preferred codon choice will occur at codon position (i), i represents the codon position number, W1 represents the start codon position, and W2 represents the stop codon position. Perr (j) represents the probability of the base call error at the position j of the artificial mutation, and D represents the proportional constant of 0 to 120.) An output unit for selecting and outputting the sequence having the smallest Va score. A device comprising. 2. The apparatus according to claim 1, wherein the calculation unit calculates the Va score of the input cDNA sequence and the Va score of the corrected cDNA sequence based on formula (II). Va = logP1 + logP2 + logPatg (II) (In the above formula, P1 and P2 have the same meanings as defined in claim 1, and Patg represents the probability of occurrence of a more preferable start codon position.) 3. The apparatus according to claim 1, wherein the arithmetic unit calculates the Va score of the input cDNA sequence and the Va score of the corrected cDNA sequence based on the formula (III). Va = logP1 + logP2 + logPconsensus (III) (In the above formula, P1 and P2 have the same meanings as defined in claim 1, and Pconsensus represents the probability that a more preferable consensus sequence will occur.) 4. The apparatus according to claim 1, wherein the arithmetic unit calculates the Va score of the input cDNA sequence and the Va score of the corrected cDNA sequence based on formula (IV). Va = logP1 + logP2 + logPatg + logPconsensus (IV) (In the above formula, P1 and P2 are synonymous with the content defined in claim 1, Patg represents the probability of occurrence of a more preferable start codon position, and Pconsensus is a more preferable consensus sequence. Represents the probability that occurs.) 5. The apparatus according to claim 1, wherein Pcodon (i) is determined in the arithmetic unit according to the following table. [Table 1] 6. In the arithmetic unit, Perr (j) is Perr.
(J) = 10 ^{−0.1Q (j)} (wherein Q (j) represents the Fred score at the position j of the artificial mutation), and is calculated according to any one of claims 1 to 5. The device according to claim 1. 7. The apparatus according to claim 2, wherein Patg is determined according to the following table in the arithmetic unit. [Table 2] 8. A probability that Pconsensus produces a more preferable Kozak consensus sequence in the arithmetic unit: Pkozak
The device according to any one of claims 3 to 7, characterized in that 9. Device according to claim 8, characterized in that Pkozak is determined according to the following table. [Table 3] 10. The modified array creation unit inputs the cD.
10. The device according to any one of claims 1 to 9, characterized in that it produces 1-2 artificial mutations in the NA sequence. 11. A modified cDNA sequence containing one insertion and a modified cDNA containing one deletion in the modified sequence generator.
Modified cDNA containing A sequence and one insertion and one deletion
Device according to claim 10, characterized in that it creates an A array. 12. The apparatus according to claim 1, wherein the cDNA sequence is derived from mammals. 13. A computer-readable recording medium in which a program for correcting a frame shift error of a cDNA sequence is recorded, wherein the cDNA sequence and the electrophoretic image data used for the sequencing of the cDNA sequence are input. And a procedure for preparing a modified cDNA sequence by introducing 1 to 4 artificial mutations selected from the group consisting of insertion and deletion into the input cDNA sequence, and the input cDNA sequence Va score and modified cd
A procedure for calculating the Va score of the NA sequence based on the formula (I), and Va = logP1 + logP2 (I) (wherein, Where Pcodon (i) represents the probability that a more preferred codon choice will occur at codon position (i), i represents the codon position number, W1 represents the start codon position, and W2 represents the stop codon position. , Perr (j) represents the probability of base call error at the position j of the artificial mutation, and D represents the proportional constant of 0 to 120.) A procedure of selecting and outputting the sequence having the smallest Va score. A computer-readable recording medium in which a program to be executed is recorded. 14. The recording medium according to claim 13, wherein the arithmetic unit calculates the Va score of the input cDNA sequence and the Va score of the corrected cDNA sequence based on formula (II). . Va = logP1 + logP2 + logPatg (II) (In the above formula, P1 and P2 have the same meanings as defined in claim 13, and Patg represents the probability that a more preferable start codon position will occur.) 15. The recording medium according to claim 13, wherein the calculation unit calculates the Va score of the input cDNA sequence and the Va score of the corrected cDNA sequence based on formula (III). . Va = logP1 + logP2 + logPconsensus (III) (In the above formula, P1 and P2 have the same meanings as defined in claim 13, and Pconsensus represents the probability that a more preferred consensus sequence will occur.) 16. The recording medium according to claim 13, wherein the calculation unit calculates the Va score of the input cDNA sequence and the Va score of the corrected cDNA sequence based on formula (IV). . Va = logP1 + logP2 + logPatg + logPconsensus (IV) (In the above formula, P1 and P2 are synonymous with the content defined in claim 13, Patg represents the probability of occurrence of a more preferable start codon position, and Pconsensus represents a more preferable consensus sequence. Represents the probability that occurs.) 17. The calculation unit determines Pcodon (i) according to the following table.
16. The recording medium according to any one of 16. [Table 4] 18. In the arithmetic unit, Perr (j) is Perr.
(J) = 10 ^{−0.1Q (j)} (wherein, Q (j) represents a Fred score at the position j of the artificial mutation). The recording medium according to 1 above. 19. The calculating unit determines Patg according to the following table.
The recording medium according to any one of to 18. [Table 5] 20. Probability that Pconsensus produces a more preferable Kozak consensus sequence in the arithmetic unit: Pkoza
The recording medium according to any one of claims 15 to 19, wherein k is k. 21. The recording medium according to claim 20, wherein Pkozak is determined according to the following table. [Table 6] 22. The modified array creation unit inputs the cD
The recording medium according to any one of claims 13 to 21, characterized in that one to two artificial mutations are generated in the NA sequence. 23. A modified cDNA sequence containing one insertion and a modified cDNA containing one deletion in the modified sequence generator.
Modified cDNA containing A sequence and one insertion and one deletion
The recording medium according to claim 22, wherein an A array is created. 24. The recording medium according to claim 13, wherein the cDNA sequence is derived from a mammal. 25. A method for correcting a frame shift error in a cDNA sequence, which comprises preparing a cDNA sequence and image data of electrophoresis used for determining the sequence of the cDNA sequence, and inserting the prepared cDNA sequence into the inserted cDNA sequence. A cDNA sequence modified by generating 1 to 4 artificial mutations selected from the group consisting of:
The Va score of the A sequence is calculated based on the formula (I), and Va = logP1 + logP2 (I) (wherein, Where Pcodon (i) represents the probability that a more preferred codon choice will occur at codon position (i), i represents the codon position number, W1 represents the start codon position, and W2 represents the stop codon position. Where Perr (j) represents the probability of base call error at position j of the artificial mutation and D represents the proportionality constant of 0 to 120) A method comprising the step of selecting the sequence with the smallest Va score. . 26. The method according to claim 25, characterized in that the Va score of the prepared cDNA sequence and the Va score of the modified cDNA sequence are calculated based on formula (II). Va = logP1 + logP2 + logPatg (II) (In the above formula, P1 and P2 have the same meanings as defined in claim 25, and Patg represents the probability of occurrence of a more preferable start codon position.) 27. The method according to claim 25, wherein the Va score of the prepared cDNA sequence and the Va score of the modified cDNA sequence are calculated based on formula (III). Va = logP1 + logP2 + logPconsensus (III) (In the above formula, P1 and P2 have the same meanings as defined in claim 25, and Pconsensus represents the probability that a more preferable consensus sequence will occur.) 28. The method according to claim 25, wherein the calculation unit calculates the Va score of the input cDNA sequence and the Va score of the corrected cDNA sequence based on formula (IV). Va = logP1 + logP2 + logPatg + logPconsensus (IV) (In the above formula, P1 and P2 are synonymous with the content defined in claim 25, Patg represents the probability of occurrence of a more preferable start codon position, and Pconsensus represents a more preferable consensus sequence. Represents the probability that occurs.)