Continuous monitoring of algal crops using minimal optical information

A continuous monitoring system using minimal optical information and machine learning addresses the challenge of late stress detection in microalgae farming, enhancing early stress detection and resource optimization.

JP7587513B2Active Publication Date: 2024-11-20

Patent Information

Application Number
JP2021550066
Authority / Receiving Office
JP · JP
Patent Type
Patents
Priority Date
2019-03-14
Filing Date
2020-03-12
Publication Date
2024-11-20
Estimated Expiration
2040-03-12

AI Technical Summary

Technical Problem

Current microalgae farming methods lack real-time, high-resolution biological monitoring, leading to late detection of stress and contamination, which can result in culture collapse and inefficient resource use.

Method used

Implementing a system for continuous, in-situ monitoring of microalgae using a minimal set of calculated fluorescent light wavelengths, combined with machine learning to identify stress and physiological conditions, comprising a central unit for learning and a peripheral unit for real-time monitoring.

Benefits of technology

Enables early detection of stress, optimizing growth conditions and resource use, reducing costs, and improving yield stability in microalgae farming.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007587513000001
    Figure 0007587513000001
  • Figure 0007587513000002
    Figure 0007587513000002
  • Figure 0007587513000003
    Figure 0007587513000003
Patent Text Reader

Abstract

A method for monitoring algae species for stress includes growing a test set of algae of a given species, applying a predetermined type of stress to a portion of the algae, and irradiating the algae with a first predetermined set of wavelengths. The algae are then monitored with a second predetermined set of wavelengths to detect fluorescence and / or light absorption by the stressed algae at the first wavelength set. The detected fluorescence and / or light absorption is compared between the stressed algae and non-stressed algae for each irradiation wavelength to find a signature of the applied stress. Next, each combination of irradiation wavelength and detection wavelength is searched to find the minimum set of irradiation wavelength and detection wavelength that detects stress. The minimum size set is then used to irradiate a further set of algae of the tested species to detect the given stress.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] Related Applications This application claims the benefit of priority to Israel Patent Application No. 265,385, filed March 14, 2019, the entire contents of which are incorporated herein by reference.

[0002] The present invention, in some embodiments thereof, relates to methods and apparatus for continuous monitoring of algal crops, and more particularly, but not exclusively, to methods of monitoring that use minimal optical information. [Background technology]

[0003] Microalgae are used in human and animal nutritional supplements, cosmetics, pharmaceuticals, CO 2 Microalgae can be considered one of the most promising feedstocks for the sustainable production of goods, including products for enabling biocapture, bioenergy production, and nutrient removal from wastewater. Microalgae do not need to be grown on arable land, they can be grown in seawater containing residual nutrients, have high areal productivity, are rich in oil, protein, and carbohydrates, and algal biomass can be fractionated into both food and non-food products. Certain species of microalgae are even candidates for biofuel production due to their high productivity and high oil content. Currently, the majority of microalgae biomass production is carried out in systems that are variations of the basic common systems, namely open raceway ponds (ORPs), horizontal tubular photobioreactors (PBRs), stacked horizontal tubular PBRs, and flat panel (FP-PBRs).

[0004] Successful cultivation of microalgae presents several challenges: one must produce the desired amount of biomass at a controlled level of quality while facing hazards and various potential stresses such as bacterial or viral contamination, potential grazers, nutrient depletion, and other suboptimal environmental conditions such as light regimes and temperature changes. Microalgae farmers often work with stressed or contaminated cultures and encounter some of these hazards, but by then it is too late to respond and recover the batch, which ultimately leads to the collapse of the culture.

[0005] Despite all the mentioned risks, optimizing growth and maintaining high quality and productivity are key factors necessary for successful and profitable cultivation of microalgae. Today, microalgae growers monitor their algae cultures offline, which requires samples to be taken to a laboratory for analysis, typically once a day using basic laboratory tools and costly labor.

[0006] Today, there are no known methods to perform online biological monitoring with high temporal resolution, which would allow the collection of data several times a day, or even on an hourly basis, but such monitoring could bring substantial benefits to growers in multiple aspects, by early detection of stress in time to rescue the crop, optimizing the growth cycle and use of resources (nutrients, electricity, labor, etc.), as well as precise management of the production process. Summary of the Invention

[0007] This embodiment may provide in situ monitoring of algae using a set of calculated minimal wavelengths of light for each species. The monitoring may indicate the health of the microalgae and identify either early stress formation or general physiological state. A set of wavelengths is defined separately for each microalgae species. This set may, for example, consist of one or more light sources (at different wavelengths) for excitation and also one or more detectors for different wavelengths for fluorescence emission. Furthermore, the absorption may be detected using a detector placed on the opposite side of the sample from the excitation source.

[0008] The embodiments further extend to methods of identifying any novel algal species and a minimal set of stresses.

[0009] According to an aspect of some embodiments of the present invention there is provided a method of monitoring an algal species for stress, comprising the steps of: growing a test set of a first species of algae; subjecting a portion of the algae to a predetermined type of stress; illuminating the algae with a first set of predetermined wavelengths; monitoring the algae at a second set of predetermined wavelengths to detect fluorescence and / or light absorption of the stressed algae; comparing the detected fluorescence and / or light absorption for each illumination wavelength between the stressed and non-stressed algae to find signatures indicative of the applied stress; searching through combinations of respective illumination wavelengths and detection wavelengths to find a minimal set of illumination wavelengths and detection wavelengths that detects stress; and Using the smallest size set to irradiate a further set of algae of the first species to detect a predetermined type of stress. A method is provided, comprising:

[0010] In an embodiment, the first wavelength set and the second wavelength set are incorporated into an optical matrix G(t) of detection at each wavelength combination, with each matrix containing results at a different time.

[0011] An embodiment can include constructing variance (V) images collected for each stress at different times and generating a normalized G (NG) matrix, a relative variance (εG) matrix, a global ranking matrix (RG) and a line ranking matrix (LG).

[0012] In an embodiment, the combinatorial search is performed by applying machine learning to the matrices.

[0013] In an embodiment, the machine learning includes using a combination of decision random forests and support vector machines.

[0014] In an embodiment, the machine learning includes using a convolutional neural network.

[0015] In an embodiment, the predetermined type of stress is nitrogen stress.

[0016] In an embodiment, the second set of wavelengths includes the range including 480 and 840 nm.

[0017] In an embodiment, the first set of wavelengths includes a range including 440 nm and 800 nm.

[0018] In an embodiment, the range is covered with jumps of 20 nm.

[0019] In an embodiment, the irradiation is performed using pulses.

[0020] In an embodiment, the first species is from the genus Chlorella.

[0021] In an embodiment, the minimum size set includes less than 5 illumination wavelengths and less than 10 detection wavelengths.

[0022] In an embodiment, the minimum size set includes fewer than three illumination wavelengths and fewer than six detection wavelengths.

[0023] In an embodiment, the minimum size set includes one illumination wavelength and two detection wavelengths.

[0024] In an embodiment, detecting stress using the minimum set includes detecting a scatter in amplitude at a wavelength of the first detection wavelength compared to a wavelength of the second detection wavelength.

[0025] An embodiment of the present invention may relate to the use of the minimal wavelength set obtained as described above for monitoring growing algae.

[0026] In aspects, the invention may relate to algae grown using monitoring as described above.

[0027] According to a further aspect of an embodiment of the present invention there is provided a method of monitoring algal growth in response to stress, comprising the steps of: illuminating the algae with at least one illumination wavelength for successive time intervals; measuring fluorescence and / or light absorption from the algae at at least two detection wavelengths, the at least two detection wavelengths being different from the at least one illumination wavelength; Comparing the amplitudes of at least two detection wavelengths to determine the presence of stress. A method is provided, comprising:

[0028] The method comprises: selecting a first stress factor for the test; and selecting a corresponding set of at least one illumination wavelength and at least two detection wavelengths; may include.

[0029] In an embodiment, determining the presence of stress includes detecting a scatter in amplitude at a first one of the detection wavelengths compared to a second one of the detection wavelengths.

[0030] In a further aspect of an embodiment of the present invention there is provided an apparatus for monitoring algal growth in response to stress, comprising: an illumination source positionable to illuminate the growing algae using at least one illumination wavelength; a detector positionable to detect fluorescence and / or light absorption from the growing algae, the detector being configured to detect amplitudes at a plurality of detection wavelengths after illumination, the plurality of detection wavelengths being different from the at least one illumination wavelength; a comparator configured to compare the amplitudes of the plurality of wavelengths and determine from the comparison whether the growing algae is experiencing a given stress; An apparatus is provided comprising:

[0031] In an embodiment, the at least one illumination wavelength and the multiple detection wavelengths are selected according to the species of algae being grown and the multiple stress types acting on the species.

[0032] In a further aspect of an embodiment of the present invention there is provided an apparatus for monitoring algal growth in response to stress, comprising: an illumination source positionable to illuminate the growing algae using at least one illumination wavelength; a detector positionable to detect fluorescence and / or light absorption from the growing algae, the detector being configured to detect amplitudes at a plurality of detection wavelengths after illumination, the plurality of detection wavelengths being different from the at least one illumination wavelength; and a communication link configured to transmit the detection result to a comparator, the comparator configured to compare the amplitudes of the multiple wavelengths and determine from the comparison whether the growing algae is under a given stress; An apparatus is provided comprising:

[0033] In yet a further aspect of an embodiment of the present invention, there is provided an apparatus for remote monitoring of algae growth in response to stress, comprising: an input for acquiring a detection result of a detector arranged to detect fluorescence and / or light absorption from the growing algae, the detector being configured to detect amplitude, after illumination with an illumination wavelength, the detection being at a plurality of detection wavelengths, the plurality of detection wavelengths being different from at least one illumination wavelength; a comparator connected to the input and configured to compare the amplitudes of the plurality of wavelengths and determine from the comparison whether the growing algae is experiencing a given stress; An apparatus is provided comprising:

[0034] Unless otherwise defined, all technical and / or scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention pertains.Although methods and materials similar or equivalent to those described herein can be used to carry out or test embodiments of the present invention, exemplary methods and / or materials are described below.In case of conflict, the patent specification, including definitions, will control.In addition, the materials, methods, and examples are merely illustrative and are not necessarily intended to be limiting.

[0035] Implementation of the method and / or system of the present invention may include performing or completing selected tasks manually, automatically, or a combination thereof. Further, depending on the actual instrumentation and equipment of the method and / or system of the present invention, some selected tasks may be performed by hardware, software or firmware, or a combination thereof using an operating system.

[0036] For example, hardware for performing selected tasks according to embodiments of the present invention may be implemented as a chip or circuit. As software, selected tasks according to embodiments of the present invention may be implemented as a number of software instructions executed by a computer using any suitable operating system. In an exemplary embodiment of the present invention, one or more tasks according to exemplary embodiments of the methods and / or systems described herein are performed by a data processor, such as a computing platform for executing a number of instructions. Optionally, the data processor includes a volatile memory for storing instructions and / or data, and / or a non-volatile storage device, such as a magnetic hard disk and / or a removable media, for storing instructions and / or data. Optionally, a network connection is also provided.

[0037] Some embodiments of the invention are herein described, by way of example only, with reference to the accompanying drawings. Referring now specifically to the drawings in detail, it is stressed that the particulars shown are by way of example and for purposes of illustrative description of embodiments of the invention. In this regard, the description taken together with the drawings will make apparent to those skilled in the art how embodiments of the invention may be practiced. [Brief description of the drawings]

[0038] [Figure 1] 1 is a simplified flow chart showing a process according to the present invention for finding a minimum set of illumination and monitoring wavelengths, or optical parameters, for monitoring algae growth. [Diagram 2] 2 is a simplified flow chart showing a portion of the process of FIG. 1 in more detail. [Diagram 3] FIG. 2 is a simplified flow chart showing how to monitor growing microalgae using only the minimal set of parameters found in FIG. 1. [Figure 4]1 is a simplified flow chart illustrating a process for monitoring different stresses according to an embodiment of the present invention. [Diagram 5] FIG. 1 is a simplified schematic block diagram of an apparatus for monitoring algae growth in accordance with an embodiment of the present invention. [Figure 6] FIG. 2 is a simplified block diagram of all microalgae monitoring components using the minimum set of optical parameters found in FIG. 1. [Figure 7] FIG. 1 is an example of raw data measured in a learning component for three biological replicates of Chlorella desiccata at specific time points under either nitrogen starvation stress or optimal (control) conditions according to the present embodiment. [Figure 8] FIG. 1 is an example of raw data (including standard deviations) measured and visualized in a study component for three biological replicates of Chlorella decicata at multiple time points under either nitrogen starvation stress or optimal (control) conditions according to the present embodiment. [Figure 9] FIG. 9 is an example of early detection of stress in Chlorella desicata due to N starvation stress for the set shown in FIG. 8, according to the present embodiment, where each matrix shows the normalized difference between the control and stressed set as a function of time after stress onset (4 h, 24 h, 48 h, 72 h). [Figure 10A] Figures 8 and 9 show the results for two different fluorescence signals (excitation at 440 nm and emission at 480 nm or 680 nm) after 4 and 48 hours, respectively, of growing Chlorella desiccata samples subjected to stress. [Figure 10B] Figures 8 and 9 show the results for two different fluorescence signals (excitation at 440 nm and emission at 480 nm or 680 nm) after 4 and 48 hours, respectively, of growing Chlorella desiccata samples subjected to stress. [Figure 11]FIG. 1 illustrates how light sources and sensors can be installed within a bioreactor, according to an embodiment of the present invention. [Figure 12A] Results of heat, nitrogen and metal stress tests on P. tricornutum. [Figure 12B] Results of heat, nitrogen and metal stress tests on P. tricornutum. [Figure 12C] Results of heat, nitrogen and metal stress tests on P. tricornutum. [Figure 12D] Results of heat, nitrogen and metal stress tests on P. tricornutum. [Figure 12E] These are the results of heat, nitrogen and metal stress tests on P. tricornutum. [Figure 12F] Results of heat, nitrogen and metal stress tests on P. tricornutum. [Figure 13A] 13-13C show hypersaline, nitrogen starvation and metal starvation tests on N. Salina, respectively. [Figure 13B] 13A-13C show hypersaline, nitrogen starvation and metal starvation tests on N. salina, respectively. [Figure 13C] 13A-13C show hypersaline, nitrogen starvation and metal starvation tests on N. salina, respectively. [Figure 13D] 1 shows a nitrogen starvation test for C. desicatta according to the present embodiments. [Figure 14A] Binary sorting of P. tricornutum algae into stressed and non-stressed populations is shown, 4 and 24 hours after the onset of stress, respectively. [Figure 14B] Binary sorting of P. tricornutum algae into stressed and non-stressed populations is shown, 4 and 24 hours after the onset of stress, respectively. [Figure 14C] Binary sorting of P. tricornutum algae into stressed and non-stressed populations is shown, 4 and 24 hours after the onset of stress, respectively. [Figure 15A] FIG. 1 illustrates optical absorption and total data classification of P. tricornutum according to an embodiment of the present invention. [Figure 15B] FIG. 1 shows light absorption and total data classification of P. tricornutum according to an embodiment of the present invention. [Figure 15C] FIG. 1 shows light absorption and total data classification of P. tricornutum according to an embodiment of the present invention. [Figure 15D] FIG. 1 shows light absorption and total data classification of P. tricornutum according to an embodiment of the present invention. [Figure 15E] FIG. 1 shows light absorption and total data classification of P. tricornutum according to an embodiment of the present invention. [Figure 15F] FIG. 1 shows light absorption and total data classification of P. tricornutum according to an embodiment of the present invention. [Figure 16A] 1 shows specific stress classification of P. tricornutum based on light absorption of only 10 wavelengths according to the present invention. [Figure 16B] 1 shows specific stress classification of P. tricornutum based on light absorption of only 10 wavelengths according to the present invention. [Figure 16C] 1 shows specific stress classification of P. tricornutum based on light absorption of only 10 wavelengths according to the present invention. [Figure 17A] 17A and 17B show specific stress classification of P. tricornutum based on absorbance of only five wavelengths, 6 and 8 hours after stress, respectively, according to an embodiment of the present invention. [Figure 17B] 17A and 17B show specific stress classification of P. tricornutum based on absorbance of only five wavelengths, 6 and 8 hours after stress, respectively, according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0039] As described, the present invention, in some embodiments thereof, relates to methods and apparatus for continuous monitoring of algal crops that can be performed online and in situ, and more particularly, but not exclusively, to methods that utilize minimal optical information.

[0040] As discussed in the background art, biological monitoring of microalgae in current technology is performed offline. In the present embodiment, online in-situ monitoring can be provided, and the environmental and physiological dynamics of microalgae can be tracked, allowing growers to better and more accurately manage the production process. For example, online and / or in-situ monitoring can enable optimal use of resources such as cultivation facilities, nutrients, electricity, optimization of culture conditions, and provide the possibility for growers to still respond and remedy the culture while obtaining early detection of stress.

[0041] The monitoring approach of the present embodiment may rely on the specific passive and active optical properties of the cultivated microalgae. 1) Different algae strains contain different pigment compositions and exhibit different optical properties. 2) Under physiological perturbations of microalgae, whether fluorescence or light absorption, there are changes in the optical properties of the algal cultures, some of which are expressed at early stages. Indeed, patterns and clusters in the absorption-excitation-fluorescence matrix (AEF) and its derivatives are specific to I) the algae type, II) the type and amplitude of stress, and III) the time since stress onset.

[0042] Therefore, monitoring the optical parameters of algae using new methods of representing the data and new methods for image classification can provide an innovative technique for monitoring the physiological status of algae in situ and can provide early stress-specific identification.

[0043] This embodiment may allow for continuous optical monitoring of microalgae and their physiological state using light sources and detectors for fluorescence at a small number of species-specific wavelength combinations. This embodiment minimizes the required optical data required to monitor the physiological state of the algae and can optimize growth, i.e., find a minimum number of excitation light sources for detectors that can be combined to allow early detection of stress situations. The system then monitors the algae using only that minimum set of parameters. The minimum set may be specific to the species of algae being monitored, as described above. The system is composed of two units: central and peripheral. The central unit is composed of two components: learning, which is essentially performed once per species per stress type, and monitoring, which is performed continuously using a minimum set of fluorescent signal sources and detectors optimized in the learning phase. The learning component is typically a laboratory-based unit that monitors the growth dynamics of a specific algae type a) during optimal growth conditions (control), and b) under abiotic and biotic stress conditions. At each time step, the system measures the fluorescence response using a separate array of well-defined excitation sources and measures the spectral fluorescence emission. The present embodiment can use the relatively weak fluorescence information to construct an image-like matrix of optical responses that serves as input for deep learning image analysis and classification methods. Early signs of different stresses have varying patterns and clustering characteristics in the light-fluorescence matrix. The present embodiment can use the matrix, its variance, and several additional matrices derived from nonlinear manipulation of the initial matrix, as described in more detail below, to detect early responses to specific stress types. The data collected in the matrix can then provide input to a deep learning system that aims to minimize the essential spectral information.

[0044] The second peripheral unit is a lightweight, small, weatherproof modular unit containing a minimal (essential) spectral source and detector. Such a unit may be installed at the algae growing site and may continuously transfer optical measurements to an interpretation component in the central unit. The real-time spectral interpretation may be based on the network optimization obtained during the training state of the first unit.

[0045] Before describing at least one embodiment of the invention in detail, it is to be understood that the invention is not necessarily limited in its application to the details of construction and the arrangement of components and / or methods set forth in the following description and / or illustrated in the drawings and / or examples, as the invention is capable of other embodiments or of being practiced or carried out in various ways.

[0046] The present embodiment may provide a system for monitoring the physiological state of microalgae using continuous, adaptive measurements and providing early warning of stress. Using artificial intelligence, the method according to the present embodiment can minimize the required optical data required to monitor algal physiology and growth and provide early indications of stress.

[0047] Reference is now made to Figure 1, which is a simplified flow chart illustrating a process according to the present embodiment for finding a minimum set of illumination and monitoring wavelengths (optical parameters) for monitoring the growth of microalgae. The method involves first growing a test set of microalgae of a particular species 10. A predetermined type of stress, such as a lack of essential nutrients, or some type of contamination, is applied 12 to a portion of the remaining algae that is kept as a control.

[0048] In box 14, growing algae is excited with a spectrum of light wavelengths, for example the set of wavelengths on the vertical axis of FIG.

[0049] In box 16, the microalgae are monitored for emission over a corresponding set of wavelengths, e.g., those along the horizontal axis of FIG. 7. At each wavelength, the amplitude of fluorescence after excitation at a given illumination wavelength is detected and a matrix is ​​filled. Figures are obtained for both the stressed algae and the control.

[0050] In box 18, the detected amplitudes for each illumination wavelength between the stressed and control microalgae are compared, and the combination of wavelengths that gives a difference between the stressed and control subjects can be used as an indicator of the applied stress.

[0051] In box 20, a process is performed to search through various combinations of each illumination wavelength and detected fluorescence wavelength to find the smallest size set of illumination and detection wavelengths that reliably detects the applied stress. The idea is to generate a large matrix of combinations and then from that matrix find a very small number of combinations that best detect the stress. For example, something like a 20x20 matrix could be a starting point and a 2x2 matrix could be an end point, depending on variables such as the particular stress being applied and the species of algae being grown.

[0052] In box 22, once a minimal size set of illuminating and monitoring wavelengths is found, that set is used for routine real-time in situ monitoring of algae growth.

[0053] Reference is now made to Figure 2, which illustrates one way in which machine learning can be performed. In box 24, the two sets of wavelengths and the corresponding detected amplitudes are incorporated into a detection optical matrix G(t). The matrix is ​​set up as a function of time by inputting the results of monitoring over successive times.

[0054] In box 26, a variance (V) matrix is ​​collected for a variety of different times. From the variance matrix, a normalized G (NG) matrix is ​​set in box 28. A relative variance (εG) matrix is ​​set in box 30. A global ranking matrix (RG) is set in box 32, and a line ranking matrix (LG) is set in box 34.

[0055] The various derived matrices are then applied as inputs to machine learning in box 36. A search through the combinations can then be performed by applying machine learning to the matrices. The machine learning can be supervised machine learning and can be informed which outcomes are stress and which are not. The process then finds out what is the minimum set of inputs from the matrices that reliably distinguishes between stressed and unstressed samples.

[0056] In other embodiments, combinatorial analysis or direct inspection of the matrices can be used.

[0057] In an embodiment, the machine learning in box 36 uses a combination of decision random forests and support vector machines, as described below. Alternatively, the machine learning may use convolutional neural networks.

[0058] One form of stress that can be applied is nitrogen deprivation, where a stressed sample is deprived of nitrogen from a certain point in its growth.

[0059] The monitoring device may initially monitor growing algae over a wavelength range including 480 and 840 nm, for example in regular jumps, for example 20 nm jumps.

[0060] The illumination wavelengths used to set up the matrix may include a range including 440 nm and 800 nm, again these may be in regular jumps, for example 20 nm.

[0061] Irradiation may be performed using pulses.

[0062] The algae may be, for example, a species from the genus Chlorella.

[0063] Machine learning can determine the minimum number of wavelengths required for both source and detector for robust and statistically significant early stage stress detection. The minimum size set that reliably identifies stress may vary between different algal species and may encompass different applied stresses. In the examples herein, the minimum size set was one excitation wavelength and two emission wavelengths, but other examples may be of different sizes. For example, there may be five, three or one excitation wavelength with ten, six or two detection wavelengths. A particular detection wavelength may be accompanied by a particular excitation wavelength. Some excitation wavelengths may be associated with multiple detection wavelengths and some may be associated with only a single detection wavelength. Again, some excitation wavelengths may be associated with more detection wavelengths than others.

[0064] As shown in Figures 10A and 10B, and described in more detail below, detecting stress using the minimum size set can include detecting amplitude scattering at one of the detection wavelengths compared to a second wavelength that can be used as a reference or baseline. Thus, the minimum size wavelength set obtained as described above can be used to monitor growing algae, for example commercially. As shown in Figure 10A, the effects of stress are evident only 4 hours after the stress is applied, thus alerting growers to any problems much earlier than is currently possible. Furthermore, monitoring can be applied much more broadly to growing algae.

[0065] Reference is now made to FIG. 3, which is a simplified flow chart illustrating how to monitor microalgae growth using only the minimum set of parameters found in FIG. 1 according to the present embodiment.

[0066] The microalgae are grown (box 40) and monitoring is applied in box 42 by illuminating the algae at successive time intervals with the illumination wavelength(s) recommended for a given species and stress type. Fluorescence and / or light absorption are then measured in box 44 at the illumination wavelength and detection wavelengths that match the algae species and stress type. The number of monitoring wavelengths per illumination wavelength may be one or more, and at least one monitoring wavelength may be different from the illumination wavelength. Note that in the case of fluorescence, all detector wavelengths are longer than the light source. In the case of absorption measurements, the detector is generally at the same wavelength as the light source.

[0067] In box 46, the amplitudes of the detected wavelengths are compared to see if there are patterns indicative of stress. If there is an indication of stress (box 48), an alarm is generated so that the grower can take the necessary measures. In either case, monitoring continues.

[0068] Reference is now made to Figure 4, which is a simplified flow chart showing how the method of Figure 3 may be performed for multiple stress factors on growing algae at 60. A first stress factor is selected at box 62 and then a corresponding set of test wavelengths is selected at box 64. Illumination and monitoring is performed as before at 66 and a notification is issued 70 if a comparison indicates stress. The stress factor can then be changed and a different set of test wavelengths used.

[0069] Referring now to Figure 5, a simplified block diagram illustrates a monitoring apparatus for monitoring microalgae growth for stress according to the present embodiments. Figure 5 illustrates an algae growth 80, which may be in a pool, pond or photobioreactor or any other type of microalgae growth unit. The algae may be grown commercially. An illumination source 82 is positioned to illuminate the growing algae 80 with one or more selected illumination wavelengths. The illumination wavelengths may be selected, for example, using a minimization process as described herein.

[0070] The detector 84 is positioned to detect the fluorescence and / or light absorption from the growing algae 80 produced as a result of the illumination. Fluorescent signals are expected to be detected only at longer wavelengths compared to the light source. Light absorption is generally measured at the same wavelength. A comparator 86 compares the amplitudes of the various detection wavelengths to determine whether the growing algae is experiencing a given stress. As described with respect to FIG. 4, the device can test for multiple stresses and different wavelength combinations can be selected for each different stress.

[0071] Furthermore, the same device can be used for different species of algae, and different illumination and detection wavelengths can be selected depending on the current species of interest and the current stress.

[0072] The system according to this embodiment is composed of two main units: a central unit and a peripheral unit. The central unit has two functional components: a learning component and a signal interpretation component. The system continuously learns and improves the algae monitoring efficiency, increasing the level of detail, for example adding new algae and additional stress types. The optical signal interpretation component can collect spectral information from the peripheral units, process the information, and provide information on algae health and early warning for various stress types.

[0073] The peripheral units may consist of lightweight, small, weatherproof and relatively inexpensive Optical Measurement Devices (OMDs) containing a minimal (essential) spectral source and detector. The OMDs may be designed to be modular, allowing easy modification of optical specifications to provide the best spectral settings for each microalgae species and growth conditions. Several peripheral units may be installed at the algae growth site and may collect optical signals that may be transmitted continuously or intermittently to a central unit. Interpretation, and if necessary real-time early warning signals, may be provided by the central unit.

[0074] Reference is now made to FIG. 6, which is a simplified block diagram that illustrates in schematic form the components of a continuous monitoring system according to an embodiment of the present invention. The system 110 comprises a central unit 112 and a peripheral unit 114. The central unit 112 includes biological and optical instruments that allow the growth of cultures under optimal and stress conditions and measure their detailed optical response in time to stress onset, this part forming the learning element 122 described later in this specification. The peripheral unit is composed of a simple optical measurement device (OMD) that includes a minimum of essential spectral sources and detectors, a minimum set of which is obtained as described herein. The OMD optical signals 117 are continuously transmitted from the microalgae growth units, e.g., microalgae photobioreactors 118, to the processing element 120 of the central unit, where the signals are analyzed and then the central unit continuously surveys the physiological state of the algae. This provides the user with a real-time assessment of the health of the cultures and an early warning of stress.

[0075] The learning / training component 122 of the central unit contains biological and optical instruments, described in more detail herein below, that allow the cultures to be grown under optimal conditions to provide control, as well as additional samples grown under typical stresses and stress conditions of interest. In both cases, the optical responses are measured in real time. At each time step, the system measures the spectral absorption, scattering and fluorescence responses using a continuous array of monochromatic light sources as well as light sources with broad spectral ranges.

[0076] The central unit may output results indicating the identified physiological state of the algae. The output 124 may indicate that the sample is healthy 126, or that it is under various types of stress, such as pollution 128, low nitrogen 130, or low phosphate 132.

[0077] It will be appreciated that the learning / training component may be omitted by end users who simply wish to monitor algae growth based on results already available.

[0078] In the learning phase, the system infers a robust signature of early stress stages using the weak fluorescence responses measured in several channels. For each algal species, each stress condition is applied and after stress onset, the system generates an image-like absorption-excitation-fluorescence matrix (AEF matrix, here denoted as G(t)) composed of the fluorescence and absorption responses to all excitation sources. At each stage, three repeats of the optical measurements are performed at preset intervals, allowing the calculation of the mean and variance for each time step.

[0079] This embodiment may show that different stress types induce different fluorescence and absorption signatures, producing different patterns and clusters in G(t). Images of the optical matrix G(t) and its variance (V) collected for each stress at different time steps are converted into several relevant matrices, such as a normalized G(NG) matrix, a relative variance (εG) matrix, a global ranking matrix (RG), and a line ranking matrix (LG). These image matrices serve as inputs to a deep learning system that classifies the data into stress types and stages and provides the minimum required spectral information for stress detection, as described below. Model System

[0080] To demonstrate the effect of stress on optical changes, we use Chlorella desicata (CDX) as a model microalgae from the genus Chlorella, which is widely used in industry. We choose a mild nitrogen limitation, i.e. f / 2 medium diluted 5x against fresh medium, where the nitrogen component is omitted to generate stressed samples. Nitrogen deficiency is chosen as a key stress due to its known effect on optical changes, and because nitrogen starvation commonly occurs in microalgal lipid production. Microalgae Cultivation and Growth

[0081] Microalgal cultures were grown in artificial seawater (ASW) supplemented with f / 2 medium, leaving stressed samples without nitrogen. Samples were kept at 24 °C with a 16:8 h light:dark cycle. Light was provided by cool white LED light at 150 μmol photons m-2 sec-1 between 07:00 and 23:00. All experiments were performed in triplicate. Measuring Instruments

[0082] A microplate reader (Tecan infinite pro 200) is used to measure both the light absorption and fluorescence scans of the microalgae cultures. Three biological replicas of both control and stressed samples are used, each 200 μL in volume, all set in a flat, clear 96-well plate. result

[0083] In extracting the absorption-excited fluorescence (AEF) matrix and its derivatives

[0084] The quality of optical measurements is highly dependent on the spectral bandwidth and accuracy of the excitation source. Therefore, we provide a tailor-made light source and spectral measurement probe, configured with a well-specified light source and narrow filtering methods to generate spectrally isolated narrow excitation sources on demand. For each discrete emission of the narrow light source, a continuous spectral measurement is made, and the readout can form an individual line of the G matrix.

[0085] The light source is capable of generating narrow bandwidth pulses in a wide spectral range, with each pulse intended to have a width of approximately 2 nm, ranging from 200 to 3000 nm.

[0086] For each excitation pulse (f) centered at a wavelength (φ), i.e., f_φ, a weak fluorescence signal (g) can be measured for all wavelengths (λ) greater than φ, i.e., g_λ, such that λ>φ.

[0087] The output luminescence of such measurements can be arranged in a triangular matrix (G) as shown in FIG. 7, for which the y-axis is φ and each line represents the luminescence yield at a longer wavelength g_λ. The first line of G is the spectral absorption measurement. More specifically, FIG. 7 shows an example of an averaged measured absorption excitation fluorescence (AEF) matrix. The Y-axis indicates the excitation wavelength and the X-axis columns show the fluorescence response at all wavelengths greater than the light source. Each absorption excitation fluorescence measurement n=3 (biological replicates) is provided for each sample at each time step.

[0088] Referring now to FIG. 8, an example of measured raw data (including standard deviation) is shown, providing a visualization of the learning components of three biological replicates of Chlorella desicata at multiple time points, either under nitrogen starvation stress or optimal (control) conditions according to the present embodiment. The first column on the left is the control data, the second column is for the stressed set, and the third and fourth columns are the standard deviations of the three replicates for the control (column 3) and stressed (column 4). Each line refers to the time of measurement from the start of the stress (4 hours, 24 hours, 48 ​​hours, 72 hours, and 144 hours). Each color indicates the logarithm (unitless) of the normalized reflectance.

[0089] Now, referring to FIG. 9, FIG. 9 shows an example of early detection of stress in CDX due to nitrogen starvation stress for the set shown in FIG. 8 according to the present embodiment. Each matrix shows the normalized difference between the control group and the stressed group as a function of time after the stress starts (4 hours, 24 hours, 48 ​​hours, 72 hours). A non-monotonic response of the central wavelength group is shown. During the period from 4 hours to 24 hours, there is a relative increase in the fluorescence response of the central group, followed by a sharp decrease. Such a non-trivial response as an initial increase in the signal can serve as an early warning of stress, providing a clear indication at about 12 hours compared to standard stress detection methods that detect such stress at more than about 72 hours (see signal in FIG. 8). Machine learning techniques

[0090] The present embodiment can use machine learning (ML) techniques to extract the minimum signal required to find a robust mathematical relationship between algal stress and optical readout as a function of time since stress application. We perform machine learning and data minimization for two culture growth modes as follows: 1) Growth kinetics from the initial stage of dilution culture to the exponential and saturation stages 2) In chemostat mode, where there is a balance between sink (dilution) and growth and the culture is in a steady state

[0091] ML is usually classified into supervised learning and unsupervised learning. In this embodiment, we mainly use concepts from the supervised branch of ML. Specifically, we use advanced regression techniques, random forest (RF) decision trees, and support vector machines (SVM) as the main methods. In parallel, we adopt a method based on deep learning (DL).

[0092] In the learning phase, two sets are measured simultaneously: control (undisturbed cultures) and stressed. For each set, five matrices are measured and calculated. 1) G, the average absorption excitation fluorescence of g_λ for all light sources f_φ of the measured emission (Gc for control and Gs for stressed). 2) Vc and Vs, the standard deviation matrix 3) εGc and εGd, the relative variance matrices 4) Rc and Rs, the global ranking matrices where each pixel is replaced by its ordered rank index. 5) Line (or column) ranking matrices of Lc and Ls, where each pixel is replaced by its ordered rank index per line (or column).

[0093] The inputs for each time in the above 10 images (5 matrices for stress and 5 matrices for control) reflect the state of the culture at a given time from the beginning of the stress when compared to the reference. In order to extract the minimum spectral information required for robust detection of stress for each algae type and each stress type, the present embodiment can selectively reduce the information content. Once the system is trained to robustly detect stress at an early stage using all the data, it analyzes the patterns of the input matrices as the main indicator. The training is repeated, but with more limited information, i.e. fewer channels and fewer detectors. The goal is to extract a signal that is more repeatable and therefore more robust indicative of stress. To that end, an analysis method is used that gradually reduces the information content of the matrices.

[0094] As an example, in such an approach, we can detect CDX in nitrogen starvation stress after 4 hours by using one light source at 440 nm and two detectors, one at 480 nm and one at 680 nm. Stressed cultures enhance the fluorescent signal at 680 nm in the first few hours and only decrease it afterwards. This allows the indication of stress to be shown by comparing the two fluorescent signals, the unaffected (480 nm) and the affected (680 nm). We note that the typical time to detect such stress in the current state of the art is 3 days.

[0095] The results are shown in Figures 10A and 10B. The minimum signal approach involves detecting CDX at 4 hours of nitrogen starvation stress using one light source at 440 nm and two detectors at 480 nm and 680 nm. Blue dots indicate the difference in 680 (stressed control) and red dots indicate the difference in 480 (stressed control). At 4 hours (left panel), the 680 signal of the stressed culture is increased compared to the control, while at 48 hours (right panel), the 680 signal of the stressed culture is decreased. Purpose

[0096] Continuous monitoring of algae can optimize production costs and increase growers' overall yields by: 1) Preventing collapses due to algal mortality events by enabling producers to respond to stress in real time. 2) Optimizing the growing cycle by improving the use of time, labor, and equipment. 3) Scaling growth systems and reducing marginal costs.

[0097] Together, the present solution can provide greater precision in the rapidly growing microalgae industry, increasing reproducibility and stability in production. Additionally, the present embodiments can improve versatility in cultivating specific types of algae.

[0098] The present embodiment can provide improved early detection and identification of certain stresses.

[0099] So, in summary, the initial stage involves detecting the entire range from 480 to 840 nm in 20 nm jumps using radiation sources of 20 different wavelengths, again in 20 nm jumps, with the relevant combinations shown in Figure 7. The question then is, how much of this data can be ignored and still provide a reliable answer?

[0100] Therefore, the present embodiment can enable data reduction and reduced processing. The present embodiment also aims to reduce the number of radiation sources and detectors, providing a cost-effective and practically feasible method of continuous monitoring.

[0101] As shown in Figures 10A and 10B, in this example, stress is successfully detected using a single radiation source at 440 nm and two detectors, one at 480 nm and the other at 680 nm. The fluorescence data at 480 nm serves as a baseline. The data at 680 nm changes under stress compared to the baseline, providing a robust indication of stress within 4 hours.

[0102] 11, a bioreactor 200 with sensors inserted is shown. The sensors are arranged as two panels, a light source panel 202 and a detection panel 204. The two panels extend from a circuit box 206. In the light source panel 202, a central white light source 208 is surrounded by four LED light sources with narrow filters L1...L4. The exact wavelength of each filter L1...L4 is specific to each algae strain.

[0103] The detection panel 204 has four detectors D1...D4, each specific to one of the four LED light sources L1...L4. The light sources illuminate the algae and the detectors detect the resulting wavelengths after absorption, as described above.

[0104] Referring now to Figures 12A-F, the stress response of P. Tricornutum is shown. Figure 12a shows the effect of heat stress. Specifically, the growth curves of P. Tricornutum under heat stress (28°C) (right hand bar) are compared to the control (24°C) (left hand bar). Cells were counted using a flow cytometer to obtain at least 5000 cells per count. Error bars represent ±1 standard deviation.

[0105] FIG. 12B shows the effect of high salinity.

[0106] Figures 12C and 12D show nitrogen starvation. Figure 12C shows the growth curves of P. tricornutum under nitrogen starvation (only 20% nitrogen amount) (right hand bar) compared to the control (left hand bar). Cells were counted using a flow cytometer to obtain at least 5000 cells per count.

[0107] Figure 12D shows the chlorophyll fluorescence response of P. tricornutum to nitrogen starvation (only 20% nitrogen amount) compared to the control (right hand bar). Chlorophyll fluorescence was measured using a flow cytometer (e.g., 488 nm em:700 nm) with at least 5000 cells per count.

[0108] Figures 12E and 12F show metal starvation. Figure 12E shows the growth curve of P. tricornutum under metal starvation (only 20% metal amount) (right hand bar) compared to the control (left hand bar). Cells were counted using a flow cytometer to obtain at least 5000 cells per count. Figure 12F shows the forward scatter (reflecting cell size) response of P. tricornutum to metal starvation (only 20% metal amount) (right hand bar) compared to the control (left hand bar). Fluorescence was measured using a flow cytometer (e.g., 488 nm) to obtain at least 5000 cells per count.

[0109] Figures 13A-C show hypersaline, nitrogen starvation and metal starvation tests on N. salina, respectively. Figure 13D shows nitrogen starvation tests on C. desiccata.

[0110] 14A and 14B, there is shown a binary classification of algae into stressed and non-stressed populations 4 and 24 hours after stress initiation, respectively. Figure 14A shows a binary classification based on all light absorption data 4 hours after stress initiation (P. tricornutum), and Figure 14B shows a binary classification based on all light absorption data 24 hours after stress initiation (P. tricornutum). Figure 14C shows the specific stress identification.

[0111]

[0013] Referring now to Figures 15A-15F, light absorption and all data classifications are shown. Figures 15A and 15B relate to a particular stress classification 6 hours after stress initiation. Figure 15B shows the relative contribution of a particular wavelength in nm for a particular stress classification based on light absorption. Figures 15C and 15D relate to a particular stress classification 8 hours after stress initiation. Figure 15D shows the relative contribution of a particular wavelength in nm for a particular stress classification based on light absorption. Figures 15E and 15F relate to a particular stress classification 24 hours after stress initiation. Figure 15F shows the relative contribution of a particular wavelength in nm for a particular stress classification based on light absorption.

[0112] Figures 16A-C show a specific stress classification of P. tricornutum based on light absorption at only 10 wavelengths, specifically 345 nm, 855 nm, 660 nm, 375 nm, 960 nm, 870 nm, 300 nm, 315 nm, 435 nm, and 675 nm. Figure 16A shows 6 hours after stress initiation. Figure 16B shows 8 hours after initiation, and Figure 16C shows classification based on absorption at the 10 wavelengths 24 hours after initiation.

[0113] Figures 17A and 17B show specific stress classification of P. tricornutum based on absorbance at only five wavelengths at 6 and 8 hours post-stress, respectively.

[0114] As used herein, the terms "comprises," "comprising," "includes," "including," "having," and combinations thereof mean "including but not limited to."

[0115] The term "consisting of" means "including and limited to."

[0116] As used herein, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise.

[0117] It is understood that certain features of the invention that are described for clarity in the context of separate embodiments may also be provided in combination in a single embodiment, and the text should be construed as if such embodiments were specifically described. Conversely, various features of the invention that are described for brevity in the context of a single embodiment may also be provided separately, or in any suitable subcombination, or as appropriate in any other described embodiment of the invention, and the text should be construed as if such embodiments were specifically described. Certain features described in the context of various embodiments should not be considered essential features of those embodiments, unless the embodiment is inoperable without those elements.

[0118] While the present invention has been described in conjunction with specific embodiments thereof, it is evident that many alternatives, modifications, and variations will be apparent to those skilled in the art. Accordingly, it is intended to embrace all such alternatives, modifications, and variations that fall within the spirit and broad scope of the appended claims.

[0119] All publications, patents, and patent applications mentioned in this specification are incorporated herein by reference in their entirety to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated herein by reference. Furthermore, citation or identification of any reference in this application should not be construed as an admission that such reference is available as prior art to the present invention. To the extent section headings are used, they should not be construed as necessarily limiting.

[0120] Additionally, any priority documents to this application are incorporated herein by reference in their entirety.

Claims

1. 1. A method for monitoring an algal species for stress to grow the algae, the method comprising: growing a test set of a first species of algae; subjecting a portion of the test set of algae to a predetermined type of stress and leaving the remainder of the test set of algae unstressed; illuminating the algae with a first predetermined set of illumination wavelengths, the first set including a first plurality of different wavelengths; monitoring the algae at a second predetermined set of fluorescence / light absorption wavelengths to detect fluorescence of stressed algae and fluorescence and / or light absorption of unstressed algae, the second set including a second plurality of different wavelengths; comparing the detected fluorescence and / or light absorption for each illumination wavelength between the stressed and unstressed algae to find indicative signatures of the applied stress; generating a plurality of combinations of respective illumination wavelengths and fluorescence / light absorption wavelengths; recording how well each of said combinations distinguishes between said stressed and unstressed portions; searching through each said combination of illumination and detection wavelengths to find a minimum set of illumination and detection wavelengths that detects the applied stress to a required level of confidence based on how well each of said combinations distinguishes between the stressed and unstressed portions, said minimum set including one illumination wavelength and at least two fluorescence / light absorption wavelengths; and using said minimum set to illuminate a further set of said first species of algae to detect said predetermined type of stress, thereby managing said stress and supporting growth of said algae. A method comprising:

2. 2. The method of claim 1, wherein the first wavelength set and the second wavelength set are incorporated into an optical matrix G(t) of detection at each wavelength combination, each matrix including results at a different time.

3. 3. The method of claim 2, comprising constructing variance (V) images collected for each stress at different times, and generating a normalized G (NG) matrix, a relative variance (εG) matrix, a global ranking matrix (RG), and a line ranking matrix (LG).

4. The method of claim 3 , wherein the searching through combinations is performed by applying machine learning to the matrix.

5. The method of claim 4 , wherein the machine learning comprises using a combination of decision random forests and support vector machines.

6. 6. The method of claim 4, wherein the machine learning comprises using a convolutional neural network.

7. 7. The method according to any one of claims 1 to 6, wherein the predetermined type of stress is nitrogen stress.

8. 8. The method of claim 1, wherein the second set of wavelengths comprises a range including 480 nm and 840 nm.

9. 9. The method of claim 1, wherein the first set of wavelengths comprises a range including 440 nm and 800 nm.

10. 10. The method of claim 8 or claim 9, wherein the range is covered with jumps of 20 nm.

11. 11. The method according to any one of claims 1 to 10, wherein the irradiation is performed using pulses.

12. 12. The method of any one of claims 1 to 11, wherein the first species is from the genus Chlorella.

13. 13. The method of claim 1, wherein the minimal set comprises less than 5 illumination wavelengths and less than 10 detection wavelengths.

14. 14. The method of claim 1, wherein the minimal set comprises less than three illumination wavelengths and less than six detection wavelengths.

15. 15. The method of any one of claims 1 to 14, wherein detecting stress using the minimal set comprises comparing an amplitude at a second wavelength among the at least two fluorescence / light absorption wavelengths that can be used as a reference with an amplitude at a first wavelength among the at least two fluorescence / light absorption wavelengths that is different from the second wavelength.

16. 16. Use of the set of smallest wavelengths obtainable according to any one of claims 1 to 15 for monitoring growing algae.

Citation Information

Patent Citations

  • Growth diagnostic system and method for large sized algae

    JP2015049227A

  • Estimation of alga growth stage and lipid content growth rate

    US8244477B1

Cited By

  • Method for detecting contamination in microalgae and cyanobacteria cultures

    ES3078635A1