Hybrid manual-machine learning PCR curve analysis and classification
AssigneeQuantGene Inc.
InventorJohannes Bhakdi
Life of this patent
Estimating the term…
Abstract
Methods, systems and apparatus for the analysis of biological samples. Biological sample held in sample plates may be processed by a PCR machine. PCR curves are generated for each biological sample and analyzed by one or more machine learning models. The PCR curves are assigned a confidence level and classified based on the analysis and the confidence level. PCR curve with a confidence level below a predetermined threshold may be flagged for analysis by a lab director. PCR curves flagged for manual analysis may be displayed on an analysis interface. The lab director may view, analyze and classify the flagged PCR curves through interaction with the analysis interface.
Claims20
A system for analyzing biological samples, the system comprising:
- an analysis terminal comprising:
- at least one processor operatively connected to a memory;
- a network module;
- one or more input devices;
- and
- a display;
- one or more analysis servers comprising:
- at least one processor operatively connected to a memory;
- a network module;
- and
- one or more machine learning modules, trained to classify PCR curves of biological samples;
- a database, operatively connected to the analysis terminal and the one or more analysis servers;
- one or more polymerase chain reaction (PCR) systems, the one or more PCR systems are configured to:
- receive a sample plate, wherein the sample plate comprises a plurality of sample wells each holding a biological sample, wherein the sample plate comprises a positive control sample and a negative control sample, and wherein the sample plate comprises multiple labeled rows and multiple labeled columns;
- perform a PCR testing process on each biological sample included in the sample plate;
- generate a PCR curve for each biological sample;
- and
- transfer the generated PCR curves to the one or more analysis servers;
- the one or more analysis servers are configured to:
- receive the generated PCR curves from the one or more PCR systems;
- analyze, by the one or more machine learning models, the received PCR curves for each biological sample;
- classify each PCR curve based on the analysis;
- transfer the classified PCR curves to the analysis terminal;
- and
- providing for display, via the analysis terminal, a user interface depicting a representation of the sample plate with a graphical representation of the generated PCR curves being depicted at multiple row and column locations, wherein a graphical representation of a first PCR curve for the positive control sample is depicted and a graphical representation of a second PCR curve for the negative control sample is depicted.
A computer implemented method for analyzing biological samples, method comprising:
- receiving, at a polymerase chain reaction (PCR) machine, a sample plate, wherein the sample plate comprises a plurality of sample wells each holding a biological sample, wherein the sample plate comprises a positive control sample and a negative control sample, and wherein the sample plate comprises multiple labeled rows and multiple labeled columns;
- performing a PCR testing process on each biological sample in the sample plate;
- generating a PCR curve for each biological sample;
- and
- transferring the generated PCR curves to one or more analysis servers, wherein the one or more analysis servers are configured to:
- receive the generated PCR curves from the PCR machine;
- analyze, by one or more machine learning models, the received PCR curves for each biological sample;
- classify each PCR curve based on the analysis;
- transfer the classified PCR curves to an analysis terminal;
- and
- providing for display, via the analysis terminal, a user interface depicting a representation of the sample plate with a graphical representation of the generated PCR curves being depicted at multiple row and column locations, wherein a graphical representation of a first PCR curve for the positive control sample is depicted and a graphical representation of a second PCR curve for the negative control sample is depicted.
A Non-transitory computer-readable medium that stores executable program instructions that, when executed by one or more computing devices, configure the one or more computing devices to perform operations comprising:
- receiving, at a polymerase chain reaction (PCR) machine, a sample plate, wherein the sample plate comprises a plurality of sample wells each holding a biological sample, wherein the sample plate comprises a positive control sample and a negative control sample, and wherein the sample plate comprises multiple labeled rows and multiple labeled columns;
- performing a PCR testing process on each biological sample in the sample plate;
- generating a PCR curve for each biological sample;
- and transferring the generated PCR curves to one or more analysis servers, wherein the one or more analysis servers are configured to:
- receive the generated PCR curves from the PCR machine;
- analyze, by one or more machine learning models, the received PCR curves for each biological sample;
- classify each PCR curve based on the analysis;
- transfer the classified PCR curves to an analysis terminal;
- and
- providing for display, via the analysis terminal, a user interface depicting a representation of the sample plate with a graphical representation of the generated PCR curves being depicted at multiple row and column locations, wherein a graphical representation of a first PCR curve for the positive control sample is depicted and a graphical representation of a second PCR curve for the negative control sample is depicted.
Dependencies are inferred from the claim text.