Method and system for calibration of immunoassay systems through application of bayesian analysis
First Claim
1. A method for measuring a concentration of an analyte in a sample having an unknown concentration of the analyte using a binding assay in which a labelled binding partner for the analyte is allowed to react with the analyte to form a labeled complex in an assayed sample, wherein the labeled binding partner includes a tag which emits an experimental indicator and wherein the assayed sample emits in a predetermined period of time an amount of the experimental indicator, which amount is a response emitted by the assayed sample, the method comprising the steps of:
- assaying a plurality of known samples using the binding assay, wherein each known sample has a known concentration of the analyte and is allowed to react with the labeled binding partner for the analyte to form a first labeled complex and the first labeled complex is detected to obtain a response emitted by each assayed known sample;
assaying the unknown sample using the binding assay, wherein each unknown sample is allowed to react with the labeled binding partner for the analyte to form a second labeled complex and the second labeled complex is detected to obtain a response emitted by the assayed unknown sample;
supplying a prior probability density for the unknown concentration and a probability model which specifics a standard curve relating an expected response to a concentration of the analyte;
generating a posterior density for the unknown concentration based on the supplied prior density, the supplied probability model, and the responses for the assayed known samples and the assayed unknown sample, by applying Bayes'"'"' rule; and
providing a characteristic of the posterior density as a measure of the concentration of the analyte in the unknown sample.
0 Assignments
0 Petitions
Accused Products
Abstract
A method for improved measurement of a concentration of an analyte in a sample having an unknown concentration of the analyte includes using an affinity assay having an experimental indicator. Standard and control samples, each having a known concentration of the analyte are assayed and a response, an amount of experimental indicator emitted for each sample is detected. The unknown sample is also assayed and its response is detected. A prior probability density is supplied for the unknown concentration, along with a model for a standard curve relating an expected response to a concentration of the analyte. A posterior density, having a median, for the unknown concentration is generated based on the supplied prior density, the supplied model, and the responses for the standard samples, the control samples and the unknown samples, by applying Bayes'"'"' rule. The median of the posterior density may be provided as the measure of concentration of analyte in the unknown sample, with an accuracy defined by the standard deviation of the posterior density. The standard curve preferably accounts for relative error due to experimental variation by using a random variable to represent variations on the analyte concentration scale due to variability in the preparation of samples in the laboratory. The minimal detectable dose for a single assay run may also be determined.
-
Citations
35 Claims
-
1. A method for measuring a concentration of an analyte in a sample having an unknown concentration of the analyte using a binding assay in which a labelled binding partner for the analyte is allowed to react with the analyte to form a labeled complex in an assayed sample, wherein the labeled binding partner includes a tag which emits an experimental indicator and wherein the assayed sample emits in a predetermined period of time an amount of the experimental indicator, which amount is a response emitted by the assayed sample, the method comprising the steps of:
- assaying a plurality of known samples using the binding assay, wherein each known sample has a known concentration of the analyte and is allowed to react with the labeled binding partner for the analyte to form a first labeled complex and the first labeled complex is detected to obtain a response emitted by each assayed known sample;
assaying the unknown sample using the binding assay, wherein each unknown sample is allowed to react with the labeled binding partner for the analyte to form a second labeled complex and the second labeled complex is detected to obtain a response emitted by the assayed unknown sample; supplying a prior probability density for the unknown concentration and a probability model which specifics a standard curve relating an expected response to a concentration of the analyte; generating a posterior density for the unknown concentration based on the supplied prior density, the supplied probability model, and the responses for the assayed known samples and the assayed unknown sample, by applying Bayes'"'"' rule; and providing a characteristic of the posterior density as a measure of the concentration of the analyte in the unknown sample. - View Dependent Claims (3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 34)
- assaying a plurality of known samples using the binding assay, wherein each known sample has a known concentration of the analyte and is allowed to react with the labeled binding partner for the analyte to form a first labeled complex and the first labeled complex is detected to obtain a response emitted by each assayed known sample;
-
2. A method for measuring a concentration of an analyte in a sample having an unknown concentration of the analyte using a competitive binding assay in which a the analyte and a labeled competitor of the analyte are allowed to react with a binding partner for the analyte, wherein the labeled competitor includes a tag which emits an experimental indicator, wherein the analyte and the labeled competitor compete with one another for binding to the binding partner to form a labeled complex in an assayed sample, wherein the assayed sample emits in a predetermined period of time an amount of the experimental indicator, which amount is a response emitted by the assayed sample, the method comprising the steps of:
-
assaying a plurality of known samples using the competitive binding assay, wherein each known sample has a known concentration of the analyte and is allowed to react with the binding partner and the labeled competitor to form a first labeled complex and the first labeled complex is detected to obtain a response emitted by each assayed known sample; assaying the unknown sample using the competitive binding assay, wherein each unknown sample is allowed to react with the binding partner and the labeled competitor to form a second labeled complex and the second labeled complex is detected to obtain a response emitted by the assayed unknown sample; supplying a prior probability density for the unknown concentration and a probability model which specifies a standard curve relating an expected response to a concentration of the analyte; generating a posterior density for the unknown concentration based on the supplied prior density, the supplied probability model, and the responses for the assayed known samples and the assayed unknown sample, by applying Bayes'"'"' rule; and providing a characteristic of the posterior density as a measure of the concentration of the analyte in the unknown sample. - View Dependent Claims (18, 19)
-
-
20. A system for measuring a concentration of an analyte in a sample having an unknown concentration of the analyte using a binding assay including a labeled binding partner having a tag which emits an experimental indicator and wherein an assayed sample emits, in a predetermined period of time, an amount of the experimental indicator, which amount is a response emitted by the assayed sample, comprising:
-
a detector arranged to detect the response of the assayed sample, and providing an output indicative of the response detected; and a computer system having an input connected to receive the output from the detector, and including a memory and a processing unit, the memory having stored therein a prior probability density for the unknown concentration and a probability model which specifies a standard curve relating an expected response to a concentration of the analyte, the processing unit including; means for controlling the binding assay and the detector to detect a response of a plurality of known samples and the unknown sample in the predetermined period of time; means for generating a posterior density for the unknown concentration based on the stored prior density, the stored probability model, and the responses for the known samples and the unknown samples, by applying Bayes'"'"' rule; and means for communicating a characteristic of the posterior density as a measure of the concentration of analyte in the unknown sample. - View Dependent Claims (21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 35)
-
Specification