Detection of significance of somatic copy number variation based on two-dimensional statistical model

Abstract

The invention provides a somatic copy number alteration obviousness detection method based on a two-dimension statistic model. The somatic copy number alteration obviousness detection method based on the two-dimension statistic model comprises the steps that S1, SCNA data are acquired and pre-processed; S2, the coefficient of relationship between every two adjacent SCNA loca is worked out, and a chromosome is divided into a plurality of SCNA structural units which are independent relatively; S3, the statistic of each SCNA structural unit is worked out, and two-dimension random permutation is conducted on a whole genome; S4, according to different lengths L of the SCNA structural units, null distribution D[L] based on the L is established in a two-dimension space through calculation of the statistic of any SCNA structural unit with the length L in a permutation sample; the statistic of the corresponding SCNA is compared with the D[L], and the statistic of the corresponding SCNA and the D[L] are marked as a value p; if the value p is smaller than a set threshold value, the corresponding SCNA is obvious, and potential cancer exists.

Description

technical field [0001] The invention discloses a method for detecting the significance of somatic cell copy number variation based on a two-dimensional statistical model. Background technique [0002] Somatic copy number alteration (SCNA) is an important phenomenon in cancer genomes. It is mainly manifested in two states of copy number amplification and deletion, and is closely related to the occurrence and development of cancer cells. Therefore, the systematic analysis of SCNA provides an important way to study the pathogenic mechanism of cancer at the molecular level. The bottom and core problem is how to distinguish the SCNA pattern with cancer function from the randomly occurring SCNA. [0003] Numerous studies have shown that SCNA functional patterns are often hidden in the consistent variation regions of cancer genome samples, then a calculation method based on statistical theory is established to detect the recurrent (Recurrent) significance level of SCNA in multiple...

AI Technical Summary

Problems solved by technology

First, the difficulty of the problem itself: a) the number of sites is as high as 1.8 million and the number of samples is often small, forming a data pattern of high latitude and small samples; b) there is a strong correlation between SCNA sites, It is not independent, so that there is an interaction between the detection factors; c) the copy number amplification or deletion status includes two characteristics, namely the variation frequency and the variation magnitude, which requires a reasonable mechanism to balance these two characteristics; d) SCNA Structural patterns vary in length, which requires consideration of the different background distributions of SCNAs of different lengths

Second, the theories and methods for solving the problem are challenging: a) The scale of data is large, and it is a challenge to effectively control the complexity of computing time and space; b) How to fully consider the correlation between SCNA sites and reduce the significance of SCNA The conservatism of the estimation of the significance level is a difficult problem; c) How to establish a null hypothesis distribution that is consistent with the statistics and enhance the statistical significance of the estimation of the significance level is a key issue that has not yet been broken through

Images