Digital sampling apparatus and methods for sorting particles
AssigneeInguran, LLC
InventorsGary Durack, Jeffrey D. Wallace, Gary P. Vandre, Lon A. Westfall, Jeremy T. Hatcher, Niraj V. Nayak
Life of this patent
Estimating the term…
Abstract
A system and method for sorting a mixture of stained particles including a digital signal processor for analyzing and classifying the digital information generated from the particles and providing a sorting signal to a sorting system as a function of the analyzed and classified digital information.
Claims55
1-80. (canceled)
A system for sorting a mixture of stained particles, including stained particles having a characteristic A and stained particles having a characteristic B, the system comprising:
- a. a fluid delivery system for delivering a fluid containing the stained particles in a flow path;
- b. an electromagnetic radiation source for exciting fluorescence emissions from the stained particles having characteristic A and the stained particles having characteristic B in the flow path;
- c. a photodetector for detecting the fluorescence emissions from the stained particles;
- d. a processor in communication with the photodetector for classifying the stained particles according to their fluorescence emissions as either particles having characteristic A or particles having characteristic B;
- e. a sorting system for sorting the stained particles according to the classification to provide at least one population containing desired particles;
- f. an analog to digital converter for sampling a time-varying analog output from the photodetector and providing an output including digital information corresponding to the time-varying analog output wherein the time-varying analog output and the corresponding digital information are indicative of characteristic A or characteristic B;
- and
- g. a digital signal processor for analyzing and classifying the digital information and providing a sorting signal to the sorting system as a function of the analyzed and classified digital information.
A system as claimed in claim 1, wherein the time-varying analog output includes a series of waveform pulses, each of which is representative of characteristic of a particle, wherein the digital signal processor detects portions of the digital information corresponding to the waveform pulses and classifies the detected portions, and wherein the digital signal processor provides the sorting signal as a function of said detected and classified portions.
A system as claimed in claim 1, wherein the particles are cells and characteristics A and B relate to physical characteristics of the cells.
A system as claimed in claim 1, wherein the flow path is formed to contain particles following generally one after another in a series which includes sequential sets of particles, including first particle sets each including one or more particles having characteristic A, second particle sets each including one or more particles having characteristic B, and third particle sets each including two or more closely spaced particles at least one of which has characteristic A and at least one of which has characteristic B.
A system as claimed in claim 1, wherein the sorting system is a droplet sorting system or a photo-damage sorting system or a fluid-switching sorting system.
A system as claimed in claim 1, wherein the digital signal processor includes instructions for detecting pulses corresponding to the digital information, instructions for extracting features in the detected pulses, and instructions for discriminating the detected pulses as a function of their extracted features.
A system as claimed in claim 1, wherein that the digital signal processor includes instructions for defining a decision boundary discriminating between the extracted features representing characteristics A and the extracted features representing characteristic B.
A system as claimed in claim 1, wherein the digital signal processor includes a data management processor for assembling the digital information into a continuous stream.
A system as claimed in claim 1, wherein the digital signal processor includes a pulse detection processor for detecting waveform pulses represented by the digital information, and in that the digital signal processor classifies the digital information as a function of the detected waveform pulses.
A system as claimed in claim 1, wherein the digital signal processor includes a pulse detection processor for detecting waveform pulses represented by the digital information, and in that the digital signal processor classifies the digital information as a function of a coefficient of variation of a population of the particles having characteristic
- A or as a function of a coefficient of variation of a population of the particles having characteristic B.
A system as claimed in claim 1, wherein the system further includes a filter for filtering the analog output at a frequency equal to or less than one half the continuous sampling rate of the analog to digital converter, and in that the analog to digital converter converts the filtered analog output into corresponding digital information, and in that the digital signal processor classifies the digital information as a function of a discrimination boundary.
A system as claimed in claim 1, wherein the system further includes a filter for filtering the analog output at a frequency equal to or less than one half the continuous sampling rate of the converter, and in that the converter converts the filtered analog output into corresponding digital information.
A system as claimed in claim 1, wherein the continuous sampling rate is about 105 MHz or higher.
A system as claimed in claim 1, wherein the digital signal processor includes a feature extraction and discrimination processor for defining a decision boundary discriminating between extracted features representing characteristics A and extracted features representing characteristic B, and in that the feature extraction and discrimination processor extracts features represented by the digital information and classifies the features as a function of the decision boundary.
A system as claimed in claim 1, wherein the digital signal processor includes a feature extraction and discrimination processor for defining a decision boundary discriminating between extracted features representing characteristics A and extracted features representing characteristic B, and in that the feature extraction and discrimination processor extracts features represented by the digital information and classifies the features as a function of a coefficient of variation of a population of the particles having characteristic A or as a function of a coefficient of variation of a population of the particles having characteristic B.
A system as claimed in claim 1, wherein the digital signal processor includes a sort processor responsive to the classifying for providing the sort signals to the sorting system.
A system as claimed in claim 1, wherein the processor enumerates the number of classified particles having characteristic A or having characteristic B.
A system as claimed in claim 1, wherein the digital signal processor classifies the digital information as a function of reducing a coefficient of variation of a population of the particles having characteristic A to be substantially equal to or less than a preset amount or as a function of minimizing a coefficient of variation of a population of the particles having characteristic B to be substantially equal to or less than a preset amount.
A system as claimed in claim 1, wherein the digital signal processor classifies the digital information such that a population of the particles having characteristic A and a population of the particles having characteristic B correspond to a computer model of three populations including a first model population of particles having characteristic A, a second model population of particles having characteristic B and a third model population of unaligned particles, said model estimating population statistics for each of the first, second and third model populations.
A system as claimed in claim 1, wherein the digital signal processor includes a pulse detection processor for detecting waveform pulses represented by the digital information, and in that the digital signal processor classifies the digital information as a function of a coefficient of variation of a population of the particles having characteristic A or as a function of a coefficient of variation of a population of the particles having characteristic B.
A system as claimed in claim 1, wherein the digital signal processor determines background characteristics of the time-varying analog output from the digital information;
- detects waveform pulses represented by the digital information as a function of the determined background characteristics;
- and provides a sorting signal to the sorting system as a function of the detected waveform pulses.
A system as claimed in claim 1, wherein the digital signal processor employs an iterative procedure for determining a pulse detection threshold for defining the background characteristics, the iterative procedure including:
- a. computing background statistic estimates from the digital information;
- using the computed estimates to apply a pulse detection logic to the digital information in order to identify pulses indicative of characteristic A or characteristic B;
- and
- b. recomputing the estimates without using the digital information corresponding to the identified pulses;
- and repeating the above procedure until the background statistic estimates converge or a fixed maximum number of iterations occurs.
A system as claimed in claim 1, wherein the digital signal processor generating initial discrimination parameters from the digital information, discriminating the digital information as a function of the initial discrimination parameters, and providing a sorting signal to the sorting system as a function of the discriminated digital information.
A system as claimed in claim 1, wherein the digital signal processor employs at least one of the following algorithms to generate the initial discrimination parameters:
- k-Means, Fuzzy k-Means and Agglomerative Hierarchical.
A system as claimed claim 1, wherein the digital signal processor includes instructions for detecting waveform pulses represented by the digital information, instructions for extracting features in the detected waveform pulses and instructions for discriminating the detected waveform pulses as a function of their extracted features.
A system as claimed in claim 1, wherein the digital processor employs a detection threshold for defining the waveform the pulses, and wherein said detection threshold is a function of a background mean estimate and a standard deviation of the digital information computed within a moving window of samples ending with the current sample.
A system as claimed claim 1, wherein the digital control employs a statistical anomaly detection analysis, and in that the digital information statistically anomalous from background characteristics of the digital information is considered part of a digital pulse.
A system as claimed claim 1, wherein the system further comprises a pulsed illumination device.
A system as claimed claim 1, wherein the digital signal processor extracts features from the digital information and proves the sorting signal to the sorting system as a function of the extracted features.
A method sorting a mixture of stained particles including stained particles having a characteristic A and stained particles having a characteristic B, the method including the steps of:
- a. delivering a fluid containing the stained particles;
- b. forming the fluid into a stream containing stained particles;
- c. generataing an analog output with a photodetector detecting stained particles in the stream;
- d. sampling the analog output from the photodetector with an analog to digital converter and generating corresponding digital information wherein the analog output is indicative of characteristic A or characteristic B;
- e. analyzing and classifying the digital information and sorting the stained particles as a function of the analyzed and classified digital information;
- a
- f. sorting the particles in the stream and according to provide at least one population containing desired particles;
Dependencies are inferred from the claim text.