Data transformation methods for multiplexed assays
Methods to improve the performance of an array assay are described. A correlation between fluorescence intensity-related parameters and negative control values of the assay is determined. The parameters are then adjusted as a function of the correlation. As a result, sensitivity of the assay is improved without changes in its specificity.
1. A method of developing an assay for a target, comprising:
providing an assay, comprising:
a plurality of negative samples known to lack a target;
a negative control signature that does not bind the target;
a positive signature that binds the target;
wherein each of the negative samples is combined with both the negative control signature and the positive signature;
measuring one or more paired fluorescence intensities in each sample, wherein the fluorescence intensity pair consists of a fluorescence intensity of the negative control signature and a fluorescence intensity of the positive signature;
determining, by linear regression, a correlation between the fluorescence intensities of the negative control signatures and the fluorescence intensities of the positive signatures;
determining the distribution of orthogonal distances from the regression line to the points consisting of the paired fluorescence intensity measures; and
choosing a threshold for the distance based on the distribution of orthogonal distances, wherein a distance between a measured fluorescence intensity pair and the regression line above the threshold the indicates a positive result in the assay, and a distance between a measured fluorescence intensity pair and the regression line below the threshold indicates a negative result in the assay.
2. The method of claim 1 , wherein the threshold is chosen to give a desired false positive rate.
3. The method of claim 1 , wherein the assay is a multiplexed array assay.
4. The method of claim 1 , wherein the fluorescence intensity measurements are median fluorescence intensity (MFI) measurements.
5. A method of assaying a sample for a target, comprising:
providing an assay, comprising:
a plurality of negative samples known to lack a target;
one or more samples to be assayed;
a negative control signature that does not bind the target;
a positive signature that binds the target;
wherein each of the samples is combined with both the negative control signature and the positive signature;
measuring one or more fluorescence intensities of each signature in each sample;
determining, by linear regression, a correlation between the fluorescence intensities of the negative control signatures in the negative samples and the fluorescence intensities of the positive signatures in the negative samples;
determining the orthogonal distance from the regression line to a point consisting of a measured fluorescence intensity of the negative control signature in an assayed sample, and the measured fluorescence intensity of the positive signature in the assayed sample; and
determining whether or not the assayed sample contains the target based on the distance.
6. The method of claim 5 , further comprising choosing a threshold value for the distance.
7. The method of claim 5 , wherein the assay is a multiplexed array assay.
8. The method of claim 5 , wherein the fluorescence intensity measurements are median fluorescence intensity (MFI) measurements.