Method for detecting micro-deletion and micro-repetition of chromosome

A method for detecting the chromosomal copy number variation, comprising:

• a) randomly breaking genomic DNA molecules obtained from a test sample and a normal sample to obtain DNA fragments, and sequencing said DNA fragments to obtain reads from sequencing;
• b) aligning the DNA sequences determined in step a) to a genomic reference sequence of the species of said test and normal samples, locating the determined DNA sequences on the reference sequence, and only selecting and using reads with a unique position on the reference sequence to perform analysis;
• c) seeking breakpoints on the reference sequence, wherein the breakpoint is a site with a difference in the copy number variation ratio on the two sides of the site compared with the alignment result of the normal sample, comprising:
• i) for each site b on the reference sequence, forcing local windows on left and right sides thereof to contain w normal reads so that N(xL,b)=N(b,xR)=w, where N(xL,xR) is the alignment number falling within the window (xL,xR) for the normal sample, and w is an integer greater than 1;
• ii) among these positions, screening sites which meet
• b
• =
• min
• x
• 
• p
• 
• (
• 
• D
• x
• 
• (
• x
• L
• ,
• x
• R
• )
• 
• )
• ,
• and excluding sites which meet Di(xL,xR)=0 and b−w<i<b+w, where D(xL,xR)=log(R(xL,x))−log(R(x,xR)) and
• R
• 
• (
• x
• L
• ,
• x
• R
• )
• =
• T
• 
• (
• x
• L
• ,
• x
• R
• )
• /
• a
• T
• N
• 
• (
• x
• L
• ,
• x
• R
• )
• /
• a
• N
• ,
• where the numbers of reads of the normal sample and of reads of the test sample that are aligned with the reference sequence uniquely are aN and aT respectively, and the numbers of reads that fall within the window (xL,xR) and are aligned with the reference sequence uniquely are N(xL,xR) and T(xL,xR) respectively, and through the two-sided significance test for normal distribution on the test statistic D(xL,xR), obtaining p(|D(xL,xR)|) for each site;
• iii) setting Pbkp, and repeating the above steps until all sites meeting p(|D(xL,xR)|)>pbkp are obtained, so as to obtain a collection of candidate sites which is BcBc={b1, b2, . . . , bN}, wherein pbkp is selected by:
• taking the normal sample as a sample to be tested, executing the aforementioned steps a) to ii) in c), filtering all p(|D(xL,xR)|) through false discovery rate (FDR) control, and
• taking the last p(|D(xL,xR)|) breaking an FDR threshold in post-filtration sites as pbkp;
• wherein the steps for the false discovery rate control comprise:
• sorting datasets to be tested by significance (P value) in an ascending order to obtain their ranks (r);
• performing the test from top to bottom until a stop at the last site k which meets
• P
• k
• ≤
• r
• k
• N
• 
• α
• where Pk is the P value of the kth position, rk is the rank of the kth position, N is the total number of the sites, and α is the significance level, e.g. 0.01;
• and
• retaining k and all sites before k, and removing false-positive sites after k;
• d) for the collection of the candidate sites on the reference sequence obtained in step c which is Bc, Bc={b1, b2, . . . , bN}, the windows (bk−1,bk−1) and (bk,bk+1) existing on both sides of each site k, removing sites with a relatively small difference in the copy number variation ratio between the windows on the two sides, i.e., deleting the site k with the maximum p(|Dbk(bk−1,bk+1)|) each time, updating the p value of the merged interval (bk−1,bk+1), and through setting pmerge and repeating the step until all sites meet p(|Dbk(bk−1,bk+1)|)<pmerge, so as to obtain the sites where the chromosomal copy number variation occurs.