A culture detection method and system

Abstract

The present invention discloses a culture detection method and system, belonging to the field of biological detection. The method comprises the following steps: dividing a sample into multiple microsamples in a microfluidic chip and culturing the samples to form multiple cultures; continuously collecting images of each culture to form an image set that changes over time; extracting the size, shape, density, and position features of the culture in each image, and performing time series analysis on multiple images in the image set; obtaining dynamic features of each culture based on the extracted features and the time series analysis results, the dynamic features including growth rate, morphological change, activity, movement speed, and movement trajectory; statistically analyzing the dynamic features of all cultures, calculating the average value, confidence interval, and standard deviation of each dynamic feature, and obtaining at least one detection result of the sample type, drug resistance, and quantity based on the average values, confidence intervals, and standard deviations of the multiple dynamic features. The detection result is fast, thereby solving the problem of slow bacterial detection speed in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of biological detection, and in particular to a culture detection method and system. Background Art

[0002] Bacterial infections are caused by pathogenic bacteria invading the human body. They can occur anywhere in the body, including the skin, respiratory tract, digestive tract, genitourinary tract, bloodstream, and meninges. With the continued emergence and spread of antibiotic-resistant bacteria, rapid and reliable infection identification is crucial for proper treatment. Once bacteria are detected, their species and antibiotic resistance profile must be determined to ensure the use of effective antibiotics and reduce reliance on broad-spectrum drugs. This is crucial for timely diagnosis and treatment.

[0003] Currently, the gold standard method for detecting bacterial resistance in clinical practice is the culture method. This method involves placing a patient's sample in a culture medium containing nutrients, using the nutrients in the culture medium to promote bacterial growth, and ultimately observing whether bacterial colonies appear in the culture dish to determine the type and number of bacteria in the infection. For positive samples confirmed to be bacterial infections, further sensitivity testing for different antibiotics is required to determine the optimal treatment. The culture method is highly accurate and reliable, and can clearly identify the type and number of bacteria, helping doctors develop targeted treatment plans.

[0004] However, due to the time-consuming testing process of culture methods, sample culture usually takes 24 to 48 hours or even longer to obtain results, which may delay treatment and is not suitable for emergency situations or situations requiring frequent testing. Therefore, although culture methods are the gold standard method with high accuracy and reliability, their time cost must be weighed against the timeliness of treatment guidance in actual application. Summary of the Invention

[0005] In order to overcome the deficiencies of the prior art, one of the objectives of the present invention is to provide a culture detection method with a fast detection speed.

[0006] In order to overcome the deficiencies of the prior art, a second object of the present invention is to provide a culture detection system with a fast detection speed.

[0007] One of the purposes of the present invention is achieved by the following technical solution:

[0008] A culture detection method comprises the following steps:

[0009] loading a sample into a microfluidic chip, where the sample is divided into a plurality of microsamples;

[0010] Each of the microsamples is cultured in the microfluidic chip to form a culture;

[0011] continuously acquiring images of each of the cultures to form a collection of images of each of the cultures over time;

[0012] identifying size, shape, density, and position features of the culture in each of the images in the image set, extracting the size, shape, density, and position features from a plurality of images in the image set, and performing a time series analysis on the plurality of images in the image set;

[0013] Obtaining dynamic characteristics of each of the cultures based on the extracted features and time series analysis results, wherein the dynamic characteristics include growth rate, morphological change, activity, movement speed, and movement trajectory;

[0014] The dynamic characteristics of all cultures are counted, and the average value, confidence interval, and standard deviation of each dynamic characteristic are calculated. At least one detection result of the sample type, drug resistance, and quantity is obtained based on the average values, confidence intervals, and standard deviations of multiple dynamic characteristics.

[0015] Furthermore, when the test result is drug resistance, the sample includes a sample containing antibiotics and a sample not containing antibiotics. When the upper limit of the confidence interval of the growth rate of the sample containing antibiotics is higher than the lower limit of the confidence interval of the growth rate of the sample not containing antibiotics, the drug resistance result is that the sample is sensitive to the antibiotic.

[0016] Furthermore, when the test result is drug resistance, an upper confidence limit curve of the growth rate of samples containing antibiotics and a lower confidence limit curve of the growth rate of samples not containing antibiotics are drawn, and the time when the two intersect is the shortest detection time.

[0017] Furthermore, when the detection result is drug resistance, the microfluidic chip is provided with at least two sample chambers, and a sample containing antibiotics and a sample not containing antibiotics are added to the two sample chambers respectively.

[0018] Furthermore, the growth rate is calculated using any one of the culture quantity, culture density, and culture pixel ratio.

[0019] Furthermore, when the growth rate is calculated using the culture pixel ratio, the number of effective image pixels occupied by the culture corresponding to the i-th culture at the j-th time node is p ij , the number of effective pixels of the image is p i0 , the growth rate of the i-th culture at the j-th time point is

[0020] Furthermore, the sample is divided into multiple microsamples in the microfluidic chip specifically by: the sample forms any one of multiple independent droplets in the microfluidic chip, multiple micropores of the sample dispersion microfluidic chip, multiple microcavities of the sample dispersion microfluidic chip, and multiple microchannels of the sample dispersion microfluidic chip.

[0021] Furthermore, the step of continuously collecting images of each of the cultures specifically includes: performing time-lapse in situ imaging monitoring of the area where each culture is located using visible light imaging or fluorescence imaging.

[0022] Furthermore, in the step of identifying the size, shape, density, and position features of the culture in each image in the image collection, a deep neural network model is used for identification. The deep neural network model includes image feature extraction and time series analysis. The image feature extraction uses a convolutional neural network architecture to encode high-dimensional images into low-dimensional features while retaining the spatial attributes of the image features. The time series analysis uses a recurrent neural network.

[0023] The second object of the present invention is achieved by adopting the following technical solution:

[0024] A culture detection system for implementing any one of the above-mentioned culture detection methods, the culture detection system comprising

[0025] A microfluidic chip, wherein the microfluidic chip is used to disperse a sample into a plurality of microsamples, and the plurality of microsamples are cultured in the microfluidic chip to form a culture;

[0026] a culture module, which controls the temperature in the microfluidic chip to provide an environmental temperature suitable for culture;

[0027] a monitoring module, wherein the monitoring module collects images of each area where the culture is located that change over time;

[0028] A data processing module is used to analyze the images collected by the monitoring module to obtain at least one detection result of the sample type, drug resistance and quantity.

[0029] Compared with the existing technology, the culture detection method of the present invention is as follows: a sample is divided into multiple microsamples in a microfluidic chip; each of the microsamples is cultured in the microfluidic chip to form a culture; images of each culture are continuously collected to form an image set of each culture changing over time; the size, shape, density, and position characteristics of the culture in each image in the image set are identified, the size, shape, density, and position characteristics of multiple images in the image set are extracted, and time series analysis is performed on multiple images in the image set; dynamic characteristics of each culture are obtained based on the extracted characteristics and the time series analysis results, and the dynamic characteristics include growth rate, morphological change, activity, movement speed, and movement trajectory; the dynamic characteristics of all cultures are counted, and the average value, confidence interval, and standard deviation of each dynamic characteristic are calculated, and at least one detection result of sample type, drug resistance, and quantity is obtained based on the average value, confidence interval, and standard deviation of multiple dynamic characteristics. The detection result is fast, thereby solving the problem of slow speed of bacterial detection methods in the existing technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a flow chart of the culture detection method of the present invention;

[0031] Figure 2 is a schematic diagram of the culture detection system of the present invention;

[0032] Figure 3 Schematic diagram of the microfluidic chip of the present invention;

[0033] Figure 4 This is a schematic diagram of deep neural network model recognition in the present invention. DETAILED DESCRIPTION

[0034] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0035] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0036] like Figure 1 As shown, the present invention provides a culture detection method, comprising the following steps:

[0037] The sample is loaded into a microfluidic chip, where the sample is divided into multiple microsamples;

[0038] Each microsample is cultured in a microfluidic chip to form a culture;

[0039] Continuously acquiring images of each culture to form a collection of images of each culture over time;

[0040] identifying size, shape, density, and position features of the culture in each image in the image set, extracting size, shape, density, and position features of multiple images in the image set, and performing time series analysis on multiple images in the image set;

[0041] The dynamic characteristics of each culture were obtained based on the extracted features and time series analysis results. The dynamic characteristics included growth rate, morphological changes, activity, movement speed, and movement trajectory.

[0042] The dynamic characteristics of all cultures are counted, the average value, confidence interval, and standard deviation of each dynamic characteristic are calculated, and at least one test result of the sample type, drug resistance, and quantity is obtained based on the average values, confidence intervals, and standard deviations of multiple dynamic characteristics.

[0043] Before loading the sample onto the microfluidic chip, it needs to be pre-treated to filter out unwanted particles. Alternatively, a filter area can be set up on the microfluidic chip to remove interfering substances from the sample and filter out unwanted particles.

[0044] The sample can be divided into multiple microsamples in the microfluidic chip by any of the following methods:

[0045] 1. The sample forms multiple independent droplets in the microfluidic chip;

[0046] 2. Sample dispersion into multiple microwells of a microfluidic chip;

[0047] 3. Multiple microcavities of sample dispersion microfluidic chip;

[0048] 4. The sample is dispersed into multiple microchannels of the microfluidic chip.

[0049] Among them, droplets, micropores, microcavities and microchannels can all be used as microchambers. The microchamber can accommodate single cells. The microchamber contains at least one transparent optical surface for imaging observation, which facilitates image acquisition.

[0050] When multiple microsamples are formed by forming multiple independent droplets within a microfluidic chip, the following steps are performed: A flow focusing method is used to prepare droplets as culture microchambers: an aqueous sample containing the culture serves as the discrete phase, and an immiscible mineral oil serves as the continuous phase. Based on liquid-liquid two-phase flow technology, the sample is sheared into independent reaction units, i.e., droplets, with volumes ranging from picoliters to nanoliters. The generated droplets enter a droplet collection structure through a transport channel. The height of the droplet collection structure's chamber matches the droplet diameter, confining the droplets to form a single layer for detection. At least one side of the chip's monitoring area is optically transparent to facilitate observation. To facilitate comparative studies, at least two sets of similar structures are typically present on the same chip.

[0051] When drug sensitivity testing is performed, that is, when the test result is drug resistance, the samples include samples containing antibiotics and samples without antibiotics. The structure of the microfluidic chip is as follows: Figure 3 As shown, the microfluidic chip is provided with two sample chambers, and samples containing antibiotics and samples without antibiotics are added to the two sample chambers respectively.

[0052] In the step of culturing each micro sample in the microfluidic chip to form a culture, the culture module provides an external environment suitable for the culture through temperature control and other methods.

[0053] Continuously acquiring images of each culture to form a collection of images of each culture changing over time is specifically performed by using visible light imaging or fluorescence imaging to perform time-lapse in situ imaging monitoring of the area where each culture is located.

[0054] The size, shape, density, and position characteristics of the culture in each image in the image set are specifically:

[0055] In order to accurately identify the microchambers and cultures in the microscopic images, traditional machine vision or deep learning methods are used for processing. In this embodiment, deep learning methods are used. Please continue to refer to Figure 4 The deep neural network model architecture is divided into two stages: image feature extraction and time series analysis. For the image feature extraction stage, a convolutional neural network architecture such as ResNet-50 can be used to encode high-dimensional images into low-dimensional features while preserving their spatial properties. For the time series analysis stage, a recurrent neural network architecture such as the LSTM architecture can be used.

[0056] The dynamic characteristics of each culture are obtained based on the extracted features and time series analysis results. The dynamic characteristics include growth rate, morphological changes, activity, movement speed, and movement trajectory. Specifically, after continuously capturing and extracting the above features, machine learning (such as SVM) or deep learning methods can be used to track the movement and division of the culture, thereby confirming the dynamic characteristics such as the movement speed and trajectory of the culture in the microchamber. Then, the type, density, activity, morphological changes and growth rate curve of the culture in each microchamber are counted to characterize the culture growth status.

[0057] The growth rate is calculated using any one of the culture number, culture density, and culture pixel ratio. When the growth rate is calculated using the culture pixel ratio, the number of effective image pixels p occupied by the culture corresponding to the i-th culture at the j-th time node is ij , the number of effective pixels of the image is p i0 , the growth rate of the i-th culture at the j-th time point is

[0058] The dynamic characteristics of all cultures are counted, and the mean, confidence interval, and standard deviation of each dynamic characteristic are calculated. Based on the mean, confidence interval, and standard deviation of multiple dynamic characteristics, at least one test result of sample type, drug resistance, and quantity is obtained as follows:

[0059] Statistical features of all microsamples, such as mean values ​​and confidence interval limits, are collected to provide an overall assessment of the culture's growth status. The deep neural network model architecture outputs predictions through a fully connected layer module. This model output is used to estimate growth rates over time and label culture categories. In the model output, samples are labeled based on the ratio of starting and ending concentrations and culture category manually identified by domain experts. Network weights can be fine-tuned using a relatively small custom dataset using model weights previously trained on a common large dataset. Based on these network outputs, a curve of the culture growth rate over time in a single microchamber can be plotted. By summarizing the growth rates of all cultures in the chip's microchambers over time, the mean, standard deviation, and 99% confidence interval of the growth rate curve can be obtained.

[0060] When the upper limit of the confidence interval for the growth rate of the sample containing the antibiotic is higher than the lower limit of the confidence interval for the growth rate of the sample without the antibiotic, the sample is considered sensitive to the antibiotic. If the test result is drug resistance, a curve for the upper confidence limit of the growth rate of the sample containing the antibiotic and a curve for the lower confidence limit of the growth rate of the sample without the antibiotic are plotted. The time at which these curves intersect is the minimum detection time.

[0061] When the above-mentioned culture detection method is used for Escherichia coli drug sensitivity testing, the specific steps are as follows:

[0062] (1) Take a sample and dilute the original bacterial solution to 1×105 CFU / mL using sterile liquid culture medium.

[0063] (2) Take the microfluidic chip for detection, add 10 μL of Escherichia coli liquid to the sample input port of the in-situ microfluidic chip, and add culture medium containing antibiotic ciprofloxacin and culture medium without antibiotics to the culture medium input port to generate droplets.

[0064] After droplet generation, the microfluidic chip is placed on a microscope stage to image the detection chamber. Image parameters vary depending on the growth rate; imaging system magnification is 40X or 20X, with an interval of 5 seconds to 15 minutes, for a total imaging time of 30 minutes to 4 hours. Depending on performance requirements, other sample volumes (e.g., 1-1000 μL) can be used, and pre-processed with enrichment and filtration before droplet generation.

[0065] Based on the continuously captured colony images, the number of effective pixels p occupied by live bacteria in the i-th chamber at the j-th time node is counted. ij and the number of effective pixels p of the image i0 Calculate the bacterial growth rate of chamber i at this moment Bacterial growth rate can also be characterized by other methods, such as the number of viable bacteria, density, etc.

[0066] Count the average growth rate of all effective chambers containing live bacteria at the jth time point Standard deviation S(L j ) and 99% confidence interval, Among them, P is the confidence interval, a j 、b j are the growth rates L at time j j The upper and lower limits of the 99% confidence interval.

[0067] The upper confidence limit curve of the growth rate of samples containing antibiotics and the lower confidence limit curve of the growth rate of samples without antibiotics were drawn respectively, and the time when the two intersected was the shortest detection time.

[0068] If at the end of the final detection time, the upper confidence limit of the growth rate of the sample containing antibiotics is still higher than the lower confidence limit of the growth rate of the sample without antibiotics, it proves that the two cannot be distinguished and the sample has a high resistance to antibiotics. Otherwise, it can be considered that the sample is sensitive to antibiotics.

[0069] Please continue reading Figure 2 The present application also discloses a culture detection system for implementing the above-mentioned culture detection method. The culture detection system includes

[0070] Microfluidic chip, the microfluidic chip is used to disperse the sample, disperse the sample into multiple micro samples, and the multiple micro samples are cultured in the microfluidic chip to form a culture;

[0071] The culture module controls the temperature inside the microfluidic chip to provide a suitable culture environment temperature for the culture;

[0072] A monitoring module, which collects images of each culture area changing over time;

[0073] The data processing module analyzes the images collected by the monitoring module to obtain at least one detection result of the sample type, drug resistance and quantity.

[0074] Specifically, microfluidic chips perform delicate sample manipulations, such as dispersing cells and other cultured organisms within a sample into individual droplets, microwells, microcavities, microchannels, and other microchambers capable of accommodating single cells. Each chamber contains at least one transparent optical surface for imaging. Microfluidic chips also include filtration areas to remove interfering substances from the sample.

[0075] The monitoring module uses visible light imaging, fluorescence imaging and other detection methods to perform time-delayed in-situ imaging monitoring of each microchamber.

[0076] The culture module provides an external environment suitable for culture through temperature control and other methods.

[0077] The data processing module analyzes the data obtained by the monitoring module and extracts useful information, such as cell number, growth rate, morphological changes, etc. to characterize the growth status of the culture.

[0078] The culture detection system also includes a fluid drive component, a fluid pipeline, an automatic control module and a human-computer interaction module.

[0079] The fluid drive module controls the flow of samples and reagents such as culture fluid, drugs, and cleaning fluids through various fluid control and drive devices such as peristaltic pumps, syringe pumps, one-way valves, and multi-way valves, and disperses the culture into the microchamber through the fluid pipeline.

[0080] The automatic control module is used to manage the functions of the culture detection system, such as managing the operation of the fluid drive module, temperature control, scheduling the monitoring module for timed imaging and any other instruments related to the micro-nanofluidic chip.

[0081] The human-computer interaction module is responsible for the communication between the user and the system. Through a friendly graphical user interface (GUI), the user can intuitively control and monitor the operating status of the entire system.

[0082] The subsystem consisting of the automatic control module, data processing module, and human-computer interaction module can be integrated within the instrument or controlled externally by a general-purpose computer, dedicated computer, personal computer, tablet device, smart mobile device, microprocessor, or other programmable data processing device. In addition to the functions described above, the subsystem can also provide processing capabilities, such as storing, interpreting, and / or executing software instructions. The controller can be configured and programmed to control the data and / or power aspects of the culture monitoring system. The data storage device can be built into the subsystem or provided separately from the subsystem. The instrument can be connected to a network. For example, the controller's communication interface can communicate with a networked computer. The networked computer can be any centralized server or cloud-based server.

[0083] Compared with the existing technology, the culture detection method of the present invention is as follows: a sample is divided into multiple microsamples in a microfluidic chip; each microsample is cultured in a microfluidic chip to form a culture; images of each culture are continuously collected to form an image set of each culture changing over time; the size, shape, density, and position characteristics of the culture in each image in the image set are identified, the size, shape, density, and position characteristics of multiple images in the image set are extracted, and time series analysis is performed on the multiple images in the image set; the dynamic characteristics of each culture are obtained based on the extracted characteristics and the time series analysis results, and the dynamic characteristics include growth rate, morphological change, activity, movement speed, and movement trajectory; the dynamic characteristics of all cultures are counted, and the average value, confidence interval, and standard deviation of each dynamic characteristic are calculated. At least one detection result of the sample type, drug resistance, and quantity is obtained based on the average value, confidence interval, and standard deviation of the multiple dynamic characteristics. The detection result is fast, thereby solving the problem of slow speed of the bacterial detection method in the existing technology.

[0084] The above embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patented invention. It should be noted that those skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention. These variations and improvements are equivalent modifications and improvements to the above embodiments based on the essential technology of the present invention and fall within the scope of protection of the present invention.

Claims

1. A culture detection method, characterized in that: The following steps are involved: loading a sample into a microfluidic chip, where the sample is divided into a plurality of microsamples; Each of the microsamples is cultured in the microfluidic chip to form a culture; continuously acquiring images of each of the cultures to form a collection of images of each of the cultures over time; Using a deep neural network model to identify the size, shape, density, and position features of the culture in each image in the image collection, the deep neural network model includes image feature extraction and time series analysis, the image feature extraction uses a convolutional neural network architecture to encode high-dimensional images into low-dimensional features while retaining the spatial properties of the image features, and the time series analysis uses a recurrent neural network to extract the size, shape, density, and position features of multiple images in the image collection, and perform time series analysis on the multiple images in the image collection; Obtaining dynamic characteristics of each of the cultures based on the extracted features and time series analysis results, wherein the dynamic characteristics include growth rate, morphological change, activity, movement speed, and movement trajectory; The dynamic characteristics of all cultures are counted, and the average value, confidence interval, and standard deviation of each dynamic characteristic are calculated. At least one test result of sample type, drug resistance, and quantity is obtained based on the average value, confidence interval, and standard deviation of multiple dynamic characteristics. When the test result is drug resistance, the sample includes samples containing antibiotics and samples not containing antibiotics. When the upper limit of the confidence interval of the growth rate of the sample containing antibiotics is higher than the lower limit of the confidence interval of the growth rate of the sample not containing antibiotics, the drug resistance result is that the sample is sensitive to the antibiotic. When the test result is drug resistance, an upper confidence limit curve of the growth rate of the sample containing antibiotics and a lower confidence limit curve of the growth rate of the sample not containing antibiotics are drawn, and the time when the two intersect is the shortest detection time.

2. The culture detection method according to claim 1, wherein: When the detection result is drug resistance, the microfluidic chip is provided with at least two sample chambers, and a sample containing antibiotics and a sample not containing antibiotics are added to the two sample chambers respectively.

3. The culture detection method according to claim 1, wherein: The growth rate is calculated using any one of the culture number, culture density, and culture pixel ratio.

4. The culture detection method according to claim 3, characterized in that: When the growth rate is calculated using the culture pixel ratio, the number of effective pixels of the image occupied by the culture corresponding to the i-th culture at the j-th time node is , the number of effective pixels of the image is , the growth rate of the i-th culture at the j-th time point is .

5. The culture detection method according to claim 1, wherein: The sample is divided into multiple microsamples in the microfluidic chip specifically by: the sample forms any one of multiple independent droplets in the microfluidic chip, multiple micropores of the sample dispersion microfluidic chip, multiple microcavities of the sample dispersion microfluidic chip, and multiple microchannels of the sample dispersion microfluidic chip.

6. The culture detection method according to claim 1, wherein: The step of continuously collecting images of each culture is specifically: using visible light imaging or fluorescence imaging to perform time-lapse in situ imaging monitoring of the area where each culture is located.

7. A culture detection system for implementing the culture detection method according to any one of claims 1 to 6, characterized in that: The culture detection system comprises A microfluidic chip, wherein the microfluidic chip is used to disperse a sample into a plurality of microsamples, and the plurality of microsamples are cultured in the microfluidic chip to form a culture; a culture module, which controls the temperature in the microfluidic chip to provide an environmental temperature suitable for culture; a monitoring module, wherein the monitoring module collects images of each area where the culture is located that change over time; A data processing module is used to analyze the images collected by the monitoring module to obtain at least one detection result of the sample type, drug resistance and quantity.

Images