JP3878503B2 - Nucleotide sequencing method - Google Patents

Description

ここで、上記数列の最初の値「９」は、１番目の塩基「C」と２番目の塩基「A」のピーク間隔を示す点数で、２番目の値「７」は、２番目の塩基「A」と３番目の塩基「A」のピーク間隔を示す点数、以下同様にして、各値が各ピークの間隔を示している。以下に、上記６種類の候補配列に対して各同定塩基のピーク間隔を算出したものを示す。

【００２９】
上記各候補配列のギャップ「：」を含む部分のピーク間隔の値を下に示す。なお、ギャップを含む部分が複数ある場合にはその平均値をとる。
候補配列１＝19
候補配列２＝18
候補配列３＝20
候補配列４＝12.5
候補配列５＝13.5
候補配列６＝13
【００３０】
図３のステップ４０において、上記ギャップを含む部分のピーク間隔の値が最も小さい候補配列を選択すると、候補配列４が選ばれる。候補配列４は、上記の最適な配列（候補配列２及び候補配列４〜６）の一つである。また、上記のピーク間隔が小さい順に候補配列を並べた場合、上位４つの配列（▲１▼候補配列４、▲２▼候補配列６、▲３▼候補配列５、▲４▼候補配列２）が上記の最適な候補配列となっており、「ギャップを含む部分のピーク間隔の値が最も小さい」という選択基準が、最適な配列を選択するための判定根拠として適していることが分かる。
【００３１】
図４では、このようにして最適な候補塩基（ここでは候補塩基４）との並置を決定したのち、候補塩基４のギャップを削除した候補配列４’（CAAGGA）を作成し、DB配列として表示している。なお、このDB配列の表示を行う際には、「２つのピークが重畳している状態」として誤って判定されていたピーク位置（「GG」のピーク位置）を補正するため、再度、１つのピークであることを考慮してピーク位置検索を行い、各ピークの最大信号強度の位置上に塩基種を示す文字が配置されるようにしてある。
【００３２】
なお、塩基配列の最終的な確定は、表示されているDB配列を参照して、人間がマニュアルで確定を行っても良いし、各ピーク位置での各塩基種の信号強度を比較して、自動的に確定を行っても良い。図４の例では、候補配列４’と同じ配列「CAAGGA」を決定配列として表示している。
【００３３】
もう一つの例として、図５に示したヘテロを含む蛍光強度波形（一部）の塩基配列を決定する場合について述べる。この図５の蛍光強度波形は、図４の場合と同様に、塩基長の長い部分で得られた波形データであるため、ピークどうしの分離が困難になりつつある部分の例である。また、一つのピークが変異を起こし、ヘテロが生じている場合の例でもある（５番目の塩基Gが変異を起こしてAとGのヘテロ（R）になっている）。このような波形に対してピーク検出を行うと、図４の場合と同様に、半値幅が広がった１つのピークが、「２つのピークが重畳している状態」として判定される。この場合には、「CAAGGAC」（＝DB配列）として判定されるべき配列が、４番目の塩基Gが２つのピークとして認識され、配列は「CAAGGRAC」として仮決定されている。この仮決定された配列「CAAGGRAC」と既知の配列「CAAGGAC」を「従来の技術」で述べた文字配列の情報のみで比較を行うスミス・ウォーターマンのホモロジー検索法で並置させた場合、下記の配列が最高スコアの候補として挙げられる。
（仮配列 ＝ＣＡＡＧＧＲＡＣ ）
候補配列１＝ＣＡＡＧＧ：ＡＣ
【００３４】
なお、上記の候補配列１は、「完全に一致した1文字のスコアは4点」（即ち、「R（=A+G）」と「A」や、「R」と「G」の組み合わせを一致と見なさない）とするスコア方法を用いた場合の結果であり、スコア方法を「一部でも一致した1文字のスコアは4点」（即ち、「R（=A+G）」と「A」や、「R」と「G」の組み合わせを一致と見なす）とした場合には、下記の候補配列２〜４もスミス・ウォーターマン法での候補配列となる（図３のステップ３７を経て、図３のステップ３８の判定でYESの場合）。
（仮配列 ＝ＣＡＡＧＧＲＡＣ）
候補配列２＝ＣＡＡ：ＧＧＡＣ
候補配列３＝ＣＡＡＧ：ＧＡＣ
候補配列４＝ＣＡＡＧＧＡ：Ｃ
ここで候補配列１は、５番目の塩基「G」が一つ目のGのピークに由来するものであるため最適な並置とは言えない。また候補配列４は、６番目の塩基「A」が「R」のピークに由来するものであるため最適な並置とは言えない。この４種類の候補の中では候補配列２〜３が最適な並置と言える。
【００３５】
しかしながら、従来のホモロジー検索では、文字配列の情報のみで判定を行うため、上記４種類の候補配列の中から、最適な配列（候補配列２〜３のどちらか）を選択するための判定根拠を見いだすことができない。
これに対して本発明では、検出した蛍光強度波形データから各塩基のピーク間隔を算出し、既知の塩基配列と並置させる際に、算出した各塩基種のピーク間隔を評価基準として用いることにより、最適な並置を行うことが可能となる。以下、上記の例に対して、本発明の方法を適用した場合について述べる。
【００３６】
まず初めに、仮配列のピーク間隔を以下のように算出しておく。

ここで、上記数列の最初の値「９」は、１番目の塩基「C」と２番目の塩基「A」のピーク間隔を示す点数で、２番目の値「７」は、２番目の塩基「A」と３番目の塩基「A」のピーク間隔を示す点数、以下同様にして、各値が各ピークの間隔を示している。
【００３７】
以下に、上記４種類の候補配列に対して各同定塩基のピーク間隔を算出したものを示す。

図３のステップ３９において算出した上記各候補配列のギャップ「：」を含む部分のピーク間隔の値を下に示す。
候補配列１＝20.0
候補配列２＝13.0
候補配列３＝15.0
候補塩基４＝19.0
【００３８】
図３のステップ４０において上記ギャップを含む部分のピーク間隔の値が最も小さい候補配列を選択した場合、候補配列２が選ばれる。候補配列２は、上記の最適な配列（候補配列２〜３）の一つである。また、上記のピーク間隔が小さい順に候補配列を並べた場合、上位２つの配列（▲１▼候補配列２、▲２▼候補配列３）が上記の最適な配列となっており、「ギャップを含む部分のピーク間隔の値が最も小さい」という選択基準が最適な配列を選択するための判定根拠として適していることが分かる。
【００３９】
図５では、このようにして最適な候補塩基（ここでは候補塩基２）との並置を決定したのち、候補塩基２のギャップを削除した候補配列２’（CAAGGAC）を作成し、DB配列として表示している。なお、このDB配列の表示を行う際には、「２つのピークが重畳している状態」として誤って判定されていたピーク位置（「GG」のピーク位置）を補正するため、再度、１つのピークであることを考慮してピーク位置検索を行い、各ピークの最大信号強度の位置上（ピーク位置の真上）に塩基種を示す文字が配置されるようにしてある。
【００４０】
なお、塩基配列の最終的な確定は、表示されているDB配列を参照して、オペレータがマニュアルで確定を行っても良いし、各ピーク位置での各塩基種の信号強度を比較して、自動的に確定を行っても良い。図５の例では、５番目の塩基において、既知配列である「A」の信号強度と、既知配列ではない「G」の信号強度が同等であることを判定の根拠として、「A」と「G」のヘテロ（R）であると確定し、候補配列２’とは１塩基異なる配列「CAAGRAC」を決定配列として表示している。
【００４１】
上記のようにして決定された塩基配列情報（ピーク番号、ピーク位置、塩基種等）は、上記図１のデータ格納部１５に記録される。記録する際の形式（フォーマット）として、既に様々なものが提案されているが、一例としてSCFフォーマットと呼ばれる形式について、以下、簡単に説明する。
SCFフォーマット（version 3.00）では、以下の項目に対応する値が、ファイルに順次、記録されている。
【００４２】
項目 内容
magic_number ＝ フォーマット識別数（文字列".SCF"を数値化したもの）
samples ＝ 波形点数
samples_offset ＝ 波形強度が記録されている最初の番地（バイトオフセット）
bases ＝ 塩基数
bases_left_clip ＝ 不使用（No. bases in left clip）
bases_right_clip＝ 不使用（No. bases in right clip）
bases_offset ＝ 塩基配列が記録されている最初の番地（バイトオフセット）
comments_size ＝ コメントの大きさ
comments_offset ＝ コメントが記録されている最初の番地（バイトオフセット）
version ＝ バージョン
sample_size ＝ 波形強度値のビットサイズ（１＝８ビット、２＝16ビット）
code_set ＝ 使用されているコードセット
private_size ＝ プライベートデータの大きさ
private_offset ＝ プライベート値が記録されている最初の番地（バイトオフセット）
spare ＝ 予備
Samples for A trace ＝ アデニン(A)塩基の波形データ
Samples for C trace ＝ シトシン(C)塩基の波形データ
Samples for G trace ＝ グアニン(G)塩基の波形データ
Samples for T trace ＝ チミン(T)塩基の波形データ
Offset into peak index for each base ＝ 各塩基のピーク位置
Accuracy estimate bases being 'A' ＝ A塩基の同定信頼性
Accuracy estimate bases being 'C' ＝ C塩基の同定信頼性
Accuracy estimate bases being 'G' ＝ G塩基の同定信頼性
Accuracy estimate bases being 'T' ＝ T塩基の同定信頼性
The called bases ＝ 同定された塩基種（決定塩基配列）
Reserved for future use ＝ 予備
Comments ＝ コメント
Private data ＝ プライベートデータ
上記SCFフォーマット（version 3.00）で記録された情報（データファイル）を用いることにより、上記図５と同等の解析結果（新規に計測した蛍光強度波形と各ピーク位置に対応する塩基種文字）を再現することが可能となる。なお図５では、既知塩基配列と解析途中の仮決定配列が表示されているが、既知塩基配列については、上記のSCFフォーマットで別途記録されたデータ（波形データやピーク位置等は省かれているもの）を用いても良いし、既知塩基配列だけが単なる文字列（テキストファイル）として記録されたものを用いても良い。また、解析途中の仮決定配列に関しては、特に記録しておく必要は無い。
【００４３】
図６は本発明による核酸塩基配列検査システムの表示画面の例（ピーク番号表示）を示す図、図７はピークを拡大表示した表示例を示す図である。図６の表示例では１画面に870ピーク分の波形が表示されているのに対し、図７の表示例は１画面に19ピーク分の波形が表示されている（約46倍の拡大率）。拡大後の画面において１画面当たり１〜50個のピークが表示されるような拡大倍率で拡大を行えば、同様の効果を得ることが出来る。
【００４４】
なお、計測した蛍光強度波形データに、上記第２の例のようなヘテロを示す部位が多数（１つ以上）存在していた場合、図６の表示欄６１に示すように、ヘテロと同定された部位のピーク番号を纏めて表示しておくことにより、ヘテロの有無を容易にチェックすることが可能となる。更に、表示されているピーク番号を選択した場合に、図７に示すように、そのピーク番号に対応する蛍光強度波形の該当部分７１を拡大して表示することによって、ヘテロと判定された部分の波形のチェックが容易になる。なお、表示画面上でのピーク番号の選択方法としては、画面上の表示部分をマウスカーソル６３等で選択してクリックする方法や、ピーク番号入力ボックス６４にピーク番号を入力する方法等を用いれば良い。
【００４５】
また、計測した蛍光強度波形データに、上記第２の例のようなDB配列とは異なる配列を示す部位が多数（１つ以上）存在していた場合、図６の表示欄６２に示すように、DB配列と異なる塩基種に同定された部位のピーク番号を纏めて表示しておくことにより、DB配列との差異の有無を容易にチェックすることが可能となる。更に、上記ヘテロの場合と同様に、表示されているピーク番号を選択した場合に、そのピーク番号に対応する蛍光強度波形の該当部分を拡大して表示することによって、容易にDB配列と異なる塩基種に同定された部分の波形をチェックすることが可能となる。なお、上記ピーク番号の選択方法としては、上記ヘテロの場合と同様に、画面上の表示部分をマウス等でクリックする方法やピーク番号を入力する方法等を用いれば良い。
【００４６】
なお、本発明が適用される図１の核酸塩基配列決定装置の構成例では、蛍光標識した核酸断片群を電気泳動し塩基長の違いにより分離する核酸断片分離部１１、分離した核酸断片にレーザ光を照射する光学機器及び発生する蛍光を検出する検出器等からなる蛍光信号計測部１２を含む装置構成例が示されているが、これらの構成部分は必ずしも必要ではなく、別の蛍光強度波形計測装置等で測定された蛍光強度波形データを読み込む機能を、蛍光信号処理部１３に持たせた場合にも、同様の効果を得ることができる。なお、上記データの読み込み方法には、フロッピーディスクや光ディスク等の記録媒体を用いる情報伝達方法や、通信回線を用いる方法等を利用できる。
【００４７】
【発明の効果】
本発明によれば、核酸断片を測定して得られた蛍光強度波形データを解釈して、A、C、G、T（U）等の塩基配列を決定する際に、既知の塩基配列を正しく並置して参照することが可能となり、その結果として塩基配列の決定精度を向上させることができる。
【図面の簡単な説明】
【図１】本発明が適用される塩基配列決定装置の構成例を示す図。
【図２】蛍光強度波形データと塩基配列の例を示す図。
【図３】蛍光強度波形データに対する処理手順の例を示す図。
【図４】本発明による塩基配列決定の例を示す図。
【図５】本発明による塩基配列決定の他の例（ヘテロを含む場合）を示す図。
【図６】本発明による核酸塩基配列検査システムの表示例（ピーク番号表示）の図。
【図７】本発明による核酸塩基配列検査システムの表示例（ピーク拡大図）の図。
【図８】スミス・ウォーターマンの方法の説明図。
【符号の説明】
１１…核酸断片分離部、１２…蛍光信号計測部、１３…蛍光信号処理部、１４…データ表示部、１５…データ格納部、１６…装置制御部、２１…蛍光強度波形、２２…塩基配列、６１…ヘテロと同定されたピークの番号表示欄、６２…ＤＢと異なる塩基種として同定されたピークの番号表示欄、６３…マウスカーソル、６４…拡大表示するピーク番号の入力部[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a nucleic acid base sequence determination method for determining a base sequence by interpreting fluorescence intensity waveform data obtained by electrophoresis of a nucleic acid sample.
[0002]
[Prior art]
Recently, it was announced that the human genome project to decode human genetic information consisting of 3 billion bases was almost completed, and in parallel, various human diseases may be caused by mutations in nucleic acid (DNA) base sequences It is being elucidated. Among individuals, the nucleobase sequence is different at many sites as well as the physical characteristics are different, and this difference is called polymorphism. A polymorphism is defined as one base change occurring at a frequency of 1% or more in the population, where one base is replaced by another base (Single Nucleotide Polymorphisms: SNPs), 1 There are those in which tens of bases are deleted or inserted, and the number of repetitions of a site where a genetic sequence of 2 to several tens of bases is repeated differs among individuals. In the 3 billion bases of the human genome, it is estimated that mutations exist at a ratio of 500 to 1000 bases, and it is thought that there are more than 3 million single base mutation pairs (SNPs). . Such a genetic diagnosis (DNA marker) method using SNPs or the like as an index is expected to be used for searching for disease genes, determining disease susceptibility, and developing pharmaceuticals (tailor-made medicine). In particular, recently, upon completion of decoding of the human genome (99% or more), there is an increasing demand for elucidating differences (polymorphisms) in individual measurement data using this huge amount of decoded data.
[0003]
Currently, many methods for detecting such polymorphisms at low cost and easily have been developed. However, since each method is a method for indirectly knowing the mutation by comparing the sizes of the nucleic acid fragments, the final method is used. In many cases, the base sequence is determined with high reliability so that mutation sites can be directly detected. Conventionally, in order to determine this base sequence, a DNA sequencing method combining a technique for fluorescently labeling a nucleic acid fragment, a high-resolution gel electrophoresis technique, and a high-sensitivity fluorescence detection technique has been widely used.
[0004]
In the conventional nucleic acid base sequence determination method, a fluorescence intensity waveform that often makes it difficult to determine the base sequence may be obtained. The reason is that when the amount of the nucleic acid fragment is small and the signal intensity is weak, or when the nucleic acid fragment itself has a secondary structure and an extra signal component is generated, the purity of the nucleic acid sample whose base sequence should be determined is low. Therefore, when a nucleic acid fragment that becomes an extra signal component is generated, there may be a case where a signal is distorted depending on conditions during a sequence reaction or electrophoresis. In addition, there is a limit to the base length that can be determined by a single measurement, and this limit is determined by the separation base length of the DNA fragment in gel electrophoresis. That is, in gel electrophoresis, it becomes difficult to separate peaks of DNA fragments that differ by one base length as the base length increases. This is because the increase in peak half-value width (peak half-value width in waveform data after sampling) with the increase in base length is larger than the increase in peak interval (peak interval in waveform data after sampling). This is because it becomes difficult to separate adjacent peaks.
[0005]
In general, for these problems, base sequences of nucleic acids having a base sequence complementary to the nucleic acid sample whose base sequence is to be determined (the sequence order (front and back) is also reversed) are determined, and 2 complementary to each other. In many cases, a sequence is determined by comparing two base sequences, or a skilled worker performs a sequence determination by visual discrimination based on experience. However, even if two samples are prepared and the base sequence is determined twice, or when visual discrimination is performed by an expert, a new problem arises that a lot of time and cost are required. Depending on the case, two base sequences complementary to each other may not be obtained. The above problems often become a problem when trying to decode a completely unknown base sequence. However, when determining the base sequence of an actual nucleic acid sample, there are many cases where at least a part of the base sequence of a nucleic acid sample to be determined is known, such as when examining a mutation at a specific site, and the human genome project Now that is almost complete, it can be said that there is an interest in elucidating the difference (individual difference = polymorphism) from the known human genome information. When such a known base sequence that can be referred to exists, the nucleic acid fragment detection data is interpreted by referring to the known base sequence by some method.
[0006]
For example, first, a base sequence roughly determined from the signal intensity (which may contain an error) is determined for the fluorescence intensity waveform of a newly obtained nucleic acid fragment. Next, a known base sequence obtained when a similar nucleic acid fragment is measured is prepared. Then, a homology search (homology search) is performed on the tentatively determined base sequence and a known base sequence, and each site of the base sequence is associated. At this time, if the tentatively determined base sequence (Sequence 1 = AACGTTCG) and the known base sequence (Sequence 2 = AACGTTCG) completely match, display and compare them in two horizontal rows as shown below. (Juxtaposition) is possible.
Array 1 = AACGTTCG
Array 2 = AACGTTCG
[0007]
On the other hand, if there is an error in the tentatively determined base sequence (sequence 1 '= ACGTTCGG) (when noise is judged as a peak or when a small peak is overlooked), or some sequences are actually mutated In such a case, the combination (optimum juxtaposition) having the highest homology is searched in consideration of the gap (portion where the base is missing) and the like as described below.
Sequence 1 ′ = A: CGTTCGG
Sequence 2 = AACGTTCG:
Here, “:” in the sequence letters is a symbol representing a gap (deficiency).
[0008]
As a conventional method for comparing DNA sequences, Smith Waterman's method based on the dynamic programming (DP) method is known as the most rigorous method (Journal of Molecular Biology, Vol.147, pp.195-197, 1981). Smith Waterman's method, when comparing two character sequences, gives a positive score for character matches and a negative score for mismatches, deletions, and insertions, and then aligns the two character sequences. This is a method for obtaining a juxtaposition that maximizes the total score among all the juxtapositions.
[0009]
As an example, juxtaposition of array 1 ″ (AAGGTATC) and array 2 (AACGTTCG) by the DP method will be described with reference to FIG. 8. In the DP method, 2 along the X-axis and Y-axis directions of the two-dimensional mesh, respectively. An array of books is placed, and each point of the mesh is taken as a node, and an optimum route from the upper left to the lower right is obtained between any two nodes when considering three vertical, horizontal, and diagonal routes between the nodes. Since vertical and horizontal arcs (lines connecting lattice points) correspond to insertion / deletion, a penalty score is applied, and a diagonal arc with which array elements are paired is also given a score according to the type of pairing. By summing up these scores along the path, the path with the highest total score is solved by the DP method to find the optimal alignment.Score generally used for alignment of DNA sequences is the insertion / deletion score. Score n characters insertion / deletion -4N-8 points for, matched one character score 4 points, the score of 1 character is different -3 points. For example, the score of the path shown in FIG. 8 is nine points.
[0010]
Other than this Smith Waterman method, FASTA method (Academic Press, Doolittle, Method in Enzymology, 183, 63-98, 1990, 1990) BLAST method (Journal of Molecular Biology, 215, 403-410, 1990) is known as a representative method.
[0011]
[Problems to be solved by the invention]
In any of the above methods, the comparison is made only with the information of the character arrangement, and when the peak position is not correctly recognized (when noise is judged as a peak, or when a small peak is overlooked, etc.), the optimum juxtaposition As a result, the sequencing accuracy may be reduced.
The present invention has been made in view of the above-described problems of the prior art, and an object thereof is to provide a method capable of determining a nucleic acid base sequence with high accuracy.
[0012]
[Means for Solving the Problems]
A nucleobase sequencing apparatus for carrying out the method of the present invention comprises means for reading fluorescence intensity waveform data obtained by electrophoresis of a phosphor-labeled nucleic acid fragment, means for calculating the read fluorescence intensity waveform data, Means for displaying fluorescence intensity waveform data and information related to the base sequence, means for calculating the fluorescence intensity waveform data, a function for storing known base sequence information, and processing the detected fluorescence intensity waveform data Function to calculate the peak interval of each base, and when referring to the information of the known base sequence, the function to determine the method of juxtaposition with the known base sequence based on the calculated peak interval of each base type Have
[0013]
That is, the nucleic acid base sequencing method according to the present invention is based on the peak information of fluorescence intensity waveform data of four types of bases obtained by electrophoresis of fluorescently labeled nucleic acid fragments of various lengths obtained from nucleic acid samples. A step of tentatively determining a base sequence of a nucleic acid sample, a step of performing a homology search on the tentatively determined base sequence and a known base sequence, and selecting a known base sequence having high homology to the tentatively determined base sequence as a candidate sequence; When there are a plurality of candidate sequences, the step of calculating the peak interval of the fluorescence intensity waveform data of the four types of bases, and the candidate sequence having the minimum interval between two peaks sandwiching the site determined as the base-deficient portion are provisionally determined And a step of juxtaposing the base sequence.
[0014]
(This) If there is a site different from the known base sequence in the determined nucleic acid sample base sequence, it is preferable to display the peak number of that site. Similarly, if there is a part identified as containing multiple bases at the same peak position in the base sequence of the nucleic acid sample determined (this), the peak number of that part is displayed. preferable. Further, when the displayed peak number is selected, it is preferable to enlarge the portion corresponding to the selected peak number of the fluorescence intensity waveform data.
[0015]
The present invention also provides the nucleotide sequence of a nucleic acid sample based on the peak information of the fluorescence intensity waveform data of four types of bases obtained by electrophoresis of fluorescently labeled nucleic acid fragments of various lengths obtained from the nucleic acid sample. There are a plurality of candidate sequences, a step of tentatively determining, performing a homology search on the tentatively determined base sequence and a known base sequence, selecting a known base sequence having high homology to the tentatively determined base sequence as a candidate sequence, and In this case, the step of calculating the peak interval of the fluorescence intensity waveform data of the four types of bases and the candidate sequence having the minimum interval between two peaks sandwiching the site determined as the base-deficient portion are juxtaposed with the tentatively determined base sequence A program for causing a computer to execute the steps is provided.
[0016]
According to the present invention, the nucleobase sequence can be determined with high accuracy. Then, by performing genetic diagnosis (DNA marker) using single nucleotide mutation pairs (SNPs) or the like as an index based on the nucleic acid base sequence determined by the method of the present invention, it becomes possible to easily detect mutations, Gene search, disease susceptibility determination, and drug development (tailor-made medicine) can be performed with high accuracy and speed.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 shows a configuration example of a nucleobase sequencing apparatus to which the present invention is applied. The apparatus includes a nucleic acid fragment migration unit 11, a fluorescence signal measurement unit 12, a fluorescence signal calculation unit 13, a data display unit 14, a data storage unit 15, and a device control unit 16 that controls each unit. The nucleic acid fragment migration unit 11 performs electrophoresis on a group of fluorescently labeled nucleic acid fragments and separates them based on the difference in base length. The fluorescence signal measuring unit 12 includes an optical device that irradiates a laser beam to a separated nucleic acid fragment, a detector that detects generated fluorescence, and the like. The fluorescence signal calculation unit 13 performs signal processing on the measured fluorescence intensity waveform data to determine a base sequence and the like. The data display unit 14 displays fluorescence intensity waveform data and information related to the determined base sequence. The data storage unit 15 records the fluorescence intensity waveform data and the determined base sequence. The apparatus control unit 16 controls the power supply of the nucleic acid fragment migration unit 11, controls the light source of the fluorescence signal measurement unit 12 and the sampling conditions of the detector, and controls between the fluorescence signal calculation unit 13 and the data display unit 14 and the data storage unit 15. Control of data transfer, control of data processing contents in the fluorescence signal calculation unit 13, and the like are performed.
[0018]
In order to determine (provisionally determine) a base sequence using the base sequence determination apparatus shown in FIG. 1, the nucleic acid fragment separation unit 11 uses various nucleic acid samples whose base sequences should be determined using the Sanger method or the like. A nucleic acid fragment group having a sufficient length is prepared. In the reaction, a primer labeled with a fluorescent dye or ddNTP labeled with a fluorescent dye is used to label the nucleic acid fragment group with the fluorescent dye.
[0019]
First, prepare the DNA (template DNA) whose base sequence you want to know. Usually, DNA with an unknown sequence is incorporated into a plasmid (DNA in the cytoplasm other than the nucleus inside cells such as bacteria, mainly having replication initiation information), or polymerase chain amplification (PCR) A nucleic acid fragment obtained by directly amplifying the base sequence by the method is used. Next, a template DNA and a primer (which has a base sequence complementary to the sequence of a specific part of the template DNA and corresponds to the one used in the reaction when using the PCR method) in a solution in a test tube And the temperature is controlled so that the primer and template form complementary double strands (annealing). Furthermore, the process proceeds to the process of replicating DNA starting from this primer, and replication is performed using an enzyme called DNA polymerase as a catalyst. In this reaction solution, dNTPs (various bases: adenine (A), cytosine (C), guanine (G), thymine (T) (or uracil (U)) monomer) necessary for DNA synthesis, and Four types of ddNTP (A, C, G, T (U) terminators) are mixed at a predetermined ratio and put at a predetermined concentration. Then, when DNA is synthesized, if ddNTP is incorporated, DNA synthesis will not proceed further (extension reaction). Here, fluorescent dyes having different colors are labeled on ddNTP according to the respective bases. As a result, nucleic acid fragments whose synthesis is stopped at various lengths (base lengths) having a ddNTP at the terminal are generated, and each fragment is labeled with a fluorescent color corresponding to the terminal base.
[0020]
Next, electrophoresis is performed on the labeled nucleic acid fragment group, and the fluorescence signal processing unit 12 detects the fluorescence signal to create fluorescence intensity waveform data. Specifically, the solution containing the nucleic acid fragment produced as described above is concentrated and purified, then denatured into single strands, and the nucleic acid fragments are separated for each base length using a gel electrophoresis apparatus. Hereinafter, a case where a capillary electrophoresis apparatus is used as an example of a gel electrophoresis apparatus will be described. First, a viscous polymer is filled in a capillary (glass capillary), and a voltage is applied to both ends thereof to introduce and migrate a nucleic acid fragment having a negative charge from one side of the capillary. At this time, since the nucleic acid fragment is a chain polymer polymer, it moves through the polymer at a speed that is inversely proportional to the molecular weight, and the shorter (smaller molecular weight) nucleic acid fragment is faster and the longer (larger molecular weight) nucleic acid fragment moves more slowly. Therefore, nucleic acid fragments can be separated for each base length. Then, near the end of the capillary (position where each nucleic acid fragment can be separated by the difference in length of one base), the nucleic acid fragment is irradiated with laser light, and the fluorescence generated from each fragment end base is measured by a detector. As described above, since fluorescence is generated in order from short nucleic acid fragments, a fluorescence intensity curve for each of the four base types is obtained. By comparing the four types of fluorescence intensities at each peak position, the base type ( A, C, G, T (U)) can be determined.
[0021]
FIG. 2 shows an example 21 of fluorescence intensity waveform data and an example 22 of a base sequence determined by interpreting it. Actually, data for several hundred bases can be obtained by one measurement, but some of them are shown here for explanation. The vertical axis represents fluorescence intensity, and the horizontal axis represents migration time. The height of the peak appearing in the fluorescence intensity waveform data 21 reflects the amount of nucleic acid fragments having a certain length. Usually, a longer nucleic acid fragment shows a peak at a later migration time, and the peak interval tends to increase as the nucleic acid fragment becomes longer. Therefore, it is also effective to perform correction using parameters determined by electrophoresis conditions such as electrophoresis voltage so that the display time axis is proportional to the base length.
[0022]
FIG. 3 is a diagram illustrating a process performed on fluorescence intensity waveform data in order to determine the base sequence of an unknown nucleic acid fragment. This process is performed by the fluorescence signal calculation unit 13.
The fluorescence signal calculation unit 13 performs smoothing processing (S31) and background correction (S32) on the fluorescence intensity waveform data of the unknown nucleic acid fragment. Thereafter, peak detection (S33) and peak interval determination (S34) are performed. In addition, since the peak interval is not always constant due to electrophoretic unevenness (smileing) during electrophoresis, peak position correction (smileing correction) is performed as necessary from the size of the obtained peak interval (S35). . Next, the signal intensity (or the area of each peak) of each base type at each peak position is compared, and the base type is sequentially determined according to a predetermined identification criterion (temporary determination of the base sequence) (S36).
[0023]
As an example of this identification criterion, the signal strength of a certain base type (for example, A) is highest at a certain peak position, and the signal strength of the largest base type (for example, C) among the remaining three base types is the largest. When it is less than T% of the signal value (A here) (T is a threshold value, for example, 50%), it is identified as the maximum base type (A here). In addition, the second base type (here, C) is T% (for example, 50%) or more, and the intensity of the third base type (for example, G) has a signal value T of the maximum base type (here, A). If it is less than 50% (for example, 50%), if the largest base type (here A) and the second base type (here C) are hetero (mixed base = multiple bases are included in the same peak position) (Identified site) (here, M (= A + C): IUB standard mixed base notation). Similarly, a mixed base display method (IUB standard mixed base display method) is determined for all combinations, but no clear value is shown as the criterion.
[0024]
As described above, in determining the base sequence of an actual nucleic acid sample, at least a part of the base sequence of the nucleic acid sample to be determined is often known, as in the case of examining a base mutation at a specific site. . When there is such a known base sequence that can be referred to, a homology search is performed on the tentatively determined base sequence and the known base sequence, and each part of the tentatively determined base sequence is compared with the known base sequence. By associating and referring to known base sequences having high homology in parallel, it is possible to improve the accuracy of base sequence determination. Hereinafter, specific processing contents of the homology search will be described with reference to the drawings.
[0025]
As an example, the case where the base sequence of the fluorescence intensity waveform (part) shown in FIG. 4 is determined will be described. The fluorescence intensity waveform in FIG. 4 is an example of a portion where separation of peaks is becoming difficult as the base length increases because it is waveform data obtained in a portion having a long base length (slow migration time). When peak detection is performed on such a waveform, one peak with a widened half-value width is often determined as “a state in which two peaks are superimposed”. In the case of FIG. 4, the sequence to be determined as “CAAGGA” (= database (DB) sequence) is identified as two peaks in which the fourth base G and the fifth base G are both “CAAGGGGA”. It has been provisionally determined. When the tentatively determined sequence “CAAGGGGA” and the known sequence “CAAGGA” are juxtaposed by the Smith Waterman homology search method in which only the character sequence information described in “Prior Art” is compared (FIG. 3). Step 37), the following three types of sequences are listed as candidates for the same score (in the case of YES in the determination of step 38 in FIG. 3).
(Tentative arrangement = CAAGGGGA)
Candidate sequence 1 = CAA :: GGA
Candidate sequence 2 = CAAG :: GA
Candidate sequence 3 = CAAGG :: A
[0026]
Here, candidate sequence 1 is not optimal juxtaposition because the sixth and seventh letters “GG” are both derived from the second G peak. Similarly, candidate sequence 3 Since the fourth and fifth letters “GG” are both derived from the first G peak, they cannot be said to be optimal juxtaposition. That is, it can be said that candidate sequence 2 is the optimum juxtaposition among these three types of candidates. The above candidate sequences 1 to 3 are the results when using a scoring method of “−4n−8 points for insertion / deletion of n characters”. When “−4n points for deletion” is used, the following candidate sequences 4 to 6 are also candidate sequences by the Smith-Waterman method, and these three types of candidate sequences can also be said to be one of the optimal juxtapositions. .
(Tentative arrangement = CAAGGGGA)
Candidate sequence 4 = CAA: G: GA
Candidate sequence 5 = CAA: GG: A
Candidate sequence 6 = CAAG: G: A
However, in the conventional homology search, the determination is made only with the information of the character sequence, and therefore, an optimal sequence (any of candidate sequence 2 and candidate sequences 4 to 6) is selected from the above six types of candidate sequences. Cannot find the basis of judgment.
[0027]
On the other hand, in the present invention, by calculating the peak interval of each base from the detected fluorescence intensity waveform data, and using the calculated peak interval of each base type as an evaluation criterion when juxtaposed with a known base sequence, Optimal juxtaposition can be performed. Hereinafter, a case where the method of the present invention is applied to the above example will be described.
[0028]
First, in step 39 of FIG. 3, the peak interval of the temporary array is calculated as follows.

Here, the first value “9” in the above sequence is a score indicating the peak interval between the first base “C” and the second base “A”, and the second value “7” is the second base. Points indicating the peak interval between “A” and the third base “A”, and similarly, each value indicates the interval between peaks. In the following, the peak intervals of each identified base are calculated for the above six types of candidate sequences.

[0029]
The value of the peak interval of the part including the gap “:” of each candidate sequence is shown below. If there are a plurality of portions including gaps, the average value is taken.
Candidate sequence 1 = 19
Candidate sequence 2 = 18
Candidate sequence 3 = 20
Candidate sequence 4 = 12.5
Candidate sequence 5 = 13.5
Candidate sequence 6 = 13
[0030]
In step 40 of FIG. 3, when a candidate sequence having the smallest peak interval value in the portion including the gap is selected, candidate sequence 4 is selected. Candidate sequence 4 is one of the above-described optimal sequences (candidate sequence 2 and candidate sequences 4 to 6). In addition, when the candidate sequences are arranged in the order of decreasing peak interval, the top four sequences ((1) candidate sequence 4, (2) candidate sequence 6, (3) candidate sequence 5, (4) candidate sequence 2) It can be seen that the above-mentioned optimal candidate sequence is suitable, and the selection criterion “the value of the peak interval of the portion including the gap is the smallest” is suitable as a determination basis for selecting the optimal sequence.
[0031]
In FIG. 4, after determining the juxtaposition with the optimal candidate base (here, candidate base 4) in this way, candidate sequence 4 ′ (CAAGGA) from which the gap of candidate base 4 has been deleted is created and displayed as a DB sequence is doing. When displaying this DB array, in order to correct the peak position (the peak position of “GG”) that was erroneously determined as “a state where two peaks are superimposed”, again one The peak position is searched in consideration of the peak, and characters indicating the base type are arranged on the position of the maximum signal intensity of each peak.
[0032]
The final determination of the base sequence may be performed manually by a human with reference to the displayed DB sequence, or by comparing the signal strength of each base type at each peak position, You may decide automatically. In the example of FIG. 4, the same sequence “CAAGGA” as the candidate sequence 4 ′ is displayed as the determined sequence.
[0033]
As another example, a case will be described in which the base sequence of the fluorescence intensity waveform (part) including the hetero shown in FIG. 5 is determined. The fluorescence intensity waveform in FIG. 5 is an example of a portion where separation of peaks is becoming difficult because it is waveform data obtained in a portion with a long base length, as in FIG. This is also an example in which one peak is mutated and heterogeneity occurs (the fifth base G is mutated to become heterozygous (R) between A and G). When peak detection is performed on such a waveform, as in the case of FIG. 4, one peak whose half width is widened is determined as “a state in which two peaks are superimposed”. In this case, in the sequence to be determined as “CAAGGAC” (= DB sequence), the fourth base G is recognized as two peaks, and the sequence is provisionally determined as “CAAGGRAC”. When this tentatively determined sequence “CAAGGRAC” and the known sequence “CAAGGAC” are juxtaposed by the Smith Waterman homology search method, which compares only with the information of the character sequence described in “Prior Art”, the following sequence: Is listed as the highest score candidate.
(Tentative sequence = CAAGGRAC)
Candidate sequence 1 = CAAGG: AC
[0034]
The above candidate sequence 1 has a score of “completely matching one character of 4 points” (that is, “R (= A + G)” and “A”, or a combination of “R” and “G”. This is a result of using a scoring method that does not regard as a match, and the scoring method is “4 points for a single character that matches even part” (ie, “R (= A + G)” and “A , Or the combination of “R” and “G” is regarded as a match), the following candidate sequences 2 to 4 are also candidate sequences in the Smith Waterman method (through step 37 in FIG. 3, (If YES in step 38 of FIG. 3).
(Tentative arrangement = CAAGGRAC)
Candidate sequence 2 = CAA: GGAC
Candidate sequence 3 = CAAG: GAC
Candidate sequence 4 = CAAGGA: C
Here, candidate sequence 1 is not optimally juxtaposed because the fifth base “G” is derived from the first G peak. Candidate sequence 4 is not optimally aligned because the sixth base “A” is derived from the peak “R”. Among these four types of candidates, candidate sequences 2 to 3 can be said to be optimal juxtaposition.
[0035]
However, in the conventional homology search, the determination is made only with the information of the character sequence, and therefore the determination basis for selecting the optimum sequence (one of the candidate sequences 2 to 3) from the above four types of candidate sequences. I can't find it.
On the other hand, in the present invention, by calculating the peak interval of each base from the detected fluorescence intensity waveform data, and using the calculated peak interval of each base type as an evaluation criterion when juxtaposed with a known base sequence, Optimal juxtaposition can be performed. Hereinafter, a case where the method of the present invention is applied to the above example will be described.
[0036]
First, the peak interval of the temporary array is calculated as follows.

Here, the first value “9” in the above sequence is a score indicating the peak interval between the first base “C” and the second base “A”, and the second value “7” is the second base. Points indicating the peak interval between “A” and the third base “A”, and similarly, each value indicates the interval between peaks.
[0037]
In the following, the peak intervals of each identified base are calculated for the above four types of candidate sequences.

The value of the peak interval of the portion including the gap “:” of each candidate sequence calculated in step 39 of FIG. 3 is shown below.
Candidate sequence 1 = 20.0
Candidate sequence 2 = 13.0
Candidate sequence 3 = 15.0
Candidate base 4 = 19.0
[0038]
When the candidate sequence having the smallest peak interval value in the portion including the gap is selected in step 40 of FIG. 3, candidate sequence 2 is selected. Candidate sequence 2 is one of the above-described optimal sequences (candidate sequences 2 to 3). In addition, when the candidate sequences are arranged in the order of decreasing peak interval, the top two sequences ((1) candidate sequence 2, (2) candidate sequence 3) are the above optimal sequences, and “include gaps” It can be seen that the selection criterion “the smallest peak interval value” is suitable as a judgment basis for selecting the optimum arrangement.
[0039]
In FIG. 5, after determining the juxtaposition with the optimal candidate base (here, candidate base 2) in this way, candidate sequence 2 ′ (CAAGGAC) in which the gap of candidate base 2 is deleted is created and displayed as a DB sequence. is doing. When displaying this DB array, in order to correct the peak position (the peak position of “GG”) that was erroneously determined as “a state where two peaks are superimposed”, again one A peak position search is performed in consideration of the peak, and characters indicating the base type are arranged on the position of the maximum signal intensity of each peak (directly above the peak position).
[0040]
The final determination of the base sequence may be performed manually by the operator with reference to the displayed DB sequence, or by comparing the signal intensity of each base type at each peak position, You may decide automatically. In the example of FIG. 5, on the fifth base, “A” and “A” and “A” that are known sequences are equal to the signal strength of “G” that is not known sequences. The sequence “CAAGRAC”, which is determined to be hetero (R) of “G” and differs from the candidate sequence 2 ′ by one base, is displayed as the determined sequence.
[0041]
The base sequence information (peak number, peak position, base type, etc.) determined as described above is recorded in the data storage unit 15 in FIG. Various formats have already been proposed as recording formats, but a format called an SCF format will be briefly described below as an example.
In the SCF format (version 3.00), values corresponding to the following items are sequentially recorded in the file.
[0042]
Item Contents
magic_number = number of format identifications (character string ".SCF")
samples = number of waveform points
samples_offset = the first address where the waveform strength is recorded (byte offset)
bases = number of bases
bases_left_clip = Not used (No. bases in left clip)
bases_right_clip = Not used (No. bases in right clip)
bases_offset = first address where the base sequence is recorded (byte offset)
comments_size = size of comments
comments_offset = The first address where the comment is recorded (byte offset)
version = version
sample_size = bit size of waveform intensity value (1 = 8 bits, 2 = 16 bits)
code_set = code set being used
private_size = private data size
private_offset = the first address where the private value is recorded (byte offset)
spare = spare
Samples for A trace = Adenine (A) base waveform data
Samples for C trace = Waveform data of cytosine (C) base
Samples for G trace = Waveform data of guanine (G) base
Samples for T trace = Waveform data of thymine (T) base
Offset into peak index for each base = Peak position of each base
Accuracy estimate bases being 'A' = A base identification reliability
Accuracy estimate bases being 'C' = C base identification reliability
Accuracy estimate bases being 'G' = G base identification reliability
Accuracy estimate bases being 'T' = T base identification reliability
The called bases = identified base species (decided base sequence)
Reserved for future use = Reserved
Comments = Comments
Private data = Private data
By using the information (data file) recorded in the SCF format (version 3.00), the analysis results equivalent to those in Fig. 5 (the newly measured fluorescence intensity waveform and the base species character corresponding to each peak position) are reproduced. It becomes possible to do. In FIG. 5, a known base sequence and a tentatively determined sequence in the middle of analysis are displayed. However, for the known base sequence, data separately recorded in the above SCF format (waveform data, peak positions, etc. are omitted). May be used, or a known base sequence recorded as a simple character string (text file) may be used. Further, it is not necessary to record the temporary determination sequence in the middle of the analysis.
[0043]
FIG. 6 is a view showing an example (peak number display) of a display screen of the nucleic acid base sequence inspection system according to the present invention, and FIG. In the display example of FIG. 6, the waveform for 870 peaks is displayed on one screen, whereas in the display example of FIG. 7, the waveform for 19 peaks is displayed on one screen (approximately 46 times magnification). . The same effect can be obtained by performing the enlargement at such an enlargement magnification that 1 to 50 peaks are displayed per screen on the enlarged screen.
[0044]
If the measured fluorescence intensity waveform data includes a large number (one or more) of sites showing heterogeneity as in the second example, it is identified as heterogeneous as shown in the display column 61 of FIG. It is possible to easily check for the presence or absence of heterogeneity by collectively displaying the peak numbers of the sites. Further, when the displayed peak number is selected, as shown in FIG. 7, the corresponding portion 71 of the fluorescence intensity waveform corresponding to the peak number is enlarged and displayed, so that the portion determined to be hetero is displayed. Waveform check becomes easy. As a method for selecting a peak number on the display screen, a method of selecting and clicking a display portion on the screen with the mouse cursor 63 or the like, a method of inputting a peak number in the peak number input box 64, or the like may be used. good.
[0045]
In addition, when the measured fluorescence intensity waveform data includes a large number (one or more) of parts showing an array different from the DB array as in the second example, as shown in the display column 62 of FIG. By collectively displaying the peak numbers of the sites identified in the base species different from the DB sequence, it is possible to easily check whether there is a difference from the DB sequence. Furthermore, as in the case of the above heterogeneity, when a displayed peak number is selected, a base different from the DB sequence can be easily displayed by enlarging the corresponding portion of the fluorescence intensity waveform corresponding to the peak number. It becomes possible to check the waveform of the part identified by the species. As a method for selecting the peak number, a method of clicking a display portion on the screen with a mouse or the like, a method of inputting a peak number, or the like may be used as in the case of the hetero.
[0046]
In the configuration example of the nucleic acid base sequence determination apparatus of FIG. 1 to which the present invention is applied, a nucleic acid fragment separation unit 11 that electrophoreses and separates a fluorescence-labeled nucleic acid fragment group according to a difference in base length, and lasers the separated nucleic acid fragments. Although an apparatus configuration example including a fluorescence signal measuring unit 12 including an optical device for irradiating light and a detector for detecting generated fluorescence is shown, these components are not always necessary, and another fluorescence intensity waveform is shown. The same effect can be obtained when the fluorescence signal processing unit 13 has a function of reading fluorescence intensity waveform data measured by a measuring device or the like. As the data reading method, an information transmission method using a recording medium such as a floppy disk or an optical disk, a method using a communication line, or the like can be used.
[0047]
【The invention's effect】
According to the present invention, when interpreting the fluorescence intensity waveform data obtained by measuring a nucleic acid fragment and determining the base sequence of A, C, G, T (U), etc., the known base sequence is correctly It becomes possible to refer to them side by side, and as a result, the accuracy of determining the base sequence can be improved.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration example of a base sequence determination apparatus to which the present invention is applied.
FIG. 2 is a diagram showing examples of fluorescence intensity waveform data and base sequences.
FIG. 3 is a diagram showing an example of a processing procedure for fluorescence intensity waveform data.
FIG. 4 is a diagram showing an example of base sequence determination according to the present invention.
FIG. 5 is a diagram showing another example of the nucleotide sequence determination according to the present invention (when heterogeneous is included).
FIG. 6 is a diagram of a display example (peak number display) of the nucleic acid base sequence inspection system according to the present invention.
FIG. 7 is a diagram showing a display example (peak enlarged view) of the nucleic acid base sequence inspection system according to the present invention.
FIG. 8 is an explanatory diagram of Smith Waterman's method.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 11 ... Nucleic acid fragment separation part, 12 ... Fluorescence signal measurement part, 13 ... Fluorescence signal processing part, 14 ... Data display part, 15 ... Data storage part, 16 ... Apparatus control part, 21 ... Fluorescence intensity waveform, 22 ... Base sequence, 61 ... Number display column of a peak identified as hetero, 62 ... Number display column of a peak identified as a base type different from DB, 63 ... Mouse cursor, 64 ... Input section of peak number to be enlarged

Claims (1)

Temporarily determining the base sequence of the nucleic acid sample based on the peak information of the fluorescence intensity waveform data of four types of bases obtained by electrophoresis of fluorescently labeled nucleic acid fragments of various lengths obtained from the nucleic acid sample; ,
Performing a homology search on the tentatively determined base sequence and the known base sequence, and selecting a known base sequence having high homology to the tentatively determined base sequence as a candidate sequence;
When there are a plurality of candidate sequences, calculating a peak interval of the fluorescence intensity waveform data of the four types of bases;
And a step of juxtaposing a candidate sequence having a minimum interval between two peaks sandwiching a site determined as a base-deficient portion with the provisionally determined base sequence.

Images