Method of selecting optimized SNP marker sets from multiple SNP markers associated with a complex disease
AssigneeSamsung Electronics Co., Ltd.
InventorsKyung-hee Park, Kyoung-a Kim, Kyu-sang Lee
Life of this patent
Estimating the term…
Abstract
Provided is a method of selecting an optimized SNP marker set from a plurality of SNP markers. The method includes selecting an SNP marker set having a high association from a plurality of SNP markers using set-association, selecting SNP markers having a high association by performing discrimination analysis on arbitrary combinations of the markers included in the SNP marker set, and selecting an optimized SNP marker combination satisfying predetermined selection criteria by using receiver operating characteristics (ROC) curve analysis of arbitrary combinations of the selected SNP markers.
Claims8
A method of selecting a SNP marker set to discriminate presence or absence of a disease in a subject, comprising:
- amplifying and determining sequences of a plurality of SNPs in DNA samples from a case population having a disease and from a control population and determining a plurality of SNP markers associated with the disease;
- selecting a maximum number of SNP markers from the plurality of SNP markers associated with the disease for a SNP marker set to discriminate presence or absence of the disease in a subject by;
- calculating for each SNP marker in the plurality of SNP markers a difference in frequency of occurrence between cases and controls to obtain an allelic association ti;
- ordering the allelic associations of the SNP markers so that the allelic association with the highest value has rank 1, such that S(n)=ti, wherein n represents a rank, and ti represents the allelic association of the ith SNP marker, which has rank n;
- calculating ΣS(n)=S(1)+S(2)+ . . . +S(n) for each SNP marker;
- calculating a degree of contribution for each SNP marker, wherein the degree of contribution=S(i)/ΣS(i);
- selecting a minimum value for the degree of contribution of a ranked SNP marker;
- and
- selecting as the maximum number (j) of SNP markers for a SNP marker set the rank i having the smallest degree of contribution greater than the minimum value;
- performing discrimination analysis on combinations of the SNP markers consisting of up to j SNP markers;
- and
- analyzing each combination by receiver operating characteristics (ROC) curve analysis to determine an area under the curve;
- selecting a SNP marker set having the highest association with the disease from the analyzed combinations, wherein the selected SNP marker set is the analyzed combination having the largest area under the curve, the highest sensitivity, the lowest specificity, or the smallest number of SNP markers;
- and,
- outputting the selected SNP marker set to a user.
Dependencies are inferred from the claim text.