An ai-based method development and data acquisition assistant for spectrometric chemical analysis

Abstract

According to a first aspect of the present disclosure, a computer-implemented method for determining a specific analytical protocol for a sample, the sample being of a type of a plurality of types of samples and each type of sample being associated with a corresponding specific analytical protocol is described Performing the method comprises: obtaining a baseline spectrum of a sample using a baseline analytical protocol, the baseline analytical protocol being the same for the plurality of types of samples; providing a machine learning model, such as a convolutional neural network, trained to output, in response to a spectrum of a sample, output data indicating a specific analytical protocol to use for the sample; applying the obtained spectrum of the sample, as an input to the machine learning model and obtaining an output of the machine learning model; and determining, based on the output, a specific analytical protocol for the sample. This minimises the burden of determining the correct analytical protocol for the sample.

Description

An Al-based method development and data acquisition assistant for spectrometric chemical analysisTECHNICAL FIELD

[0001] The present disclosure relates to useability of scientific instruments, such as spectrometers, and in particular facilitating selection of an analytical protocol for a sample.BACKGROUND

[0002] There is a general and ongoing need for systems and methods for reducing the level of analytical chemistry literacy required for effectively operating scientific instruments and analysing samples. Typically, analysing samples and operating scientific instruments for a sample requires at least knowledge of the components making up the sample and knowledge of the appropriate analytic techniques to use for such components. This means that there are two barriers to overcome in order to analyse samples effectively, both of which require expertise in the field of analytical chemistry.

[0003] There are cases when an analytical chemist may have a rudimental knowledge of the techniques for analysing a sample in its pure form but no knowledge of its common impurities or binding compounds. In some cases, an analytical chemist may know the exact components of a sample but not know which instruments to use and with which settings, or, for example, which measurement ranges to examine using these instruments in order to find out additional information about the compound, such as the purity of the compound. Both of these barriers may incur a time or financial cost due to the expense and availability of analytical chemists. Therefore, there is a need to be able to minimise the burden of determining properties of a sample and in particular the burden of setting up an instrument and specific analytical protocols.,SUMMARY

[0004] According to a first aspect of the present disclosure, a computer-implemented method for determining a specific analytical protocol for a sample, the sample being of a type of a plurality of types of samples and each type of sample being associated with a corresponding specific analytical protocol is described.

[0005] Performing the computer-implemented method comprises: obtaining a baseline spectrum of a sample using a baseline analytical protocol, the baseline analytical protocol being the same for the plurality of types of samples; providing a machine learning model, such as a convolutional neural network, trained to output, in response to a spectrum of a sample, output data indicating a specific analytical protocol to use for the sample; applying the obtained spectrum of the sample, as an input to the machine learning model and obtaining an output of the machine learning model; and determining, based on the output, a specific analytical protocol for the sample. Beneficially, this minimises the burden of determining the correct analytical protocol for the sample such that even if the sample is an unknownsample, the correct protocol can be followed in order to investigate it further. The method may ensure safe handling of potentially dangerous substances, for example, heavy metals, where they are present in the sample. It is understood that “spectrum” may refer to a spectrum obtained by any appropriate spectroscopic technique for example, using: optical emission spectrometry, such as inductively coupled plasma optical emission spectrometry; mass spectrometry, such as inductively coupled plasma mass spectrometry; gas chromatography - mass spectrometry; liquid chromatography - mass spectrometry; nuclear magnetic resonance spectrometry; or infrared spectrometry. For example, the baseline spectrum may be an echelle spectrum obtained by optical emission spectrometry; a spectrum of intensity as a function of mass over charge (m / z) obtained by mass spectrometry; or a spectrum of intensity as a function of time obtained by gas chromatography or liquid chromatography. Beneficially, very little information, needs to be known about the sample in order to obtain the baseline spectrum since the baseline spectrum is obtained by a baseline protocol which is the same for every sample of the plurality of samples. Hence, the baseline spectrum can be obtained without sample-specific knowledge or information.

[0006] In some examples, the method further comprises recording a further spectrum of the sample using the determined specific analytical protocol.

[0007] In some examples, the specific analytical protocol comprises one or more of: a method of acquisition of the sample, one or more analytic techniques for analysis of the sample, one or more preferred settings for analytic instruments used for analysis of the sample, one or more criteria for storing a sample, one or more analytic wavelengths ranges of functional groups of the sample, a concentration range for external standard preparation, one or more calibration factors, one or more sample introduction settings, the one or more sample introduction settings comprising a peristaltic pump speed, a tubing diameter and material, a nebulizer type, nebulizer gas flow, auxiliary gas flow, cooling gas flow and plasma power, one or more optical system settings, the one or more optical system settings comprising plasma view direction (axial or radial) , viewing height (if radial view is appropriate), exposure duration, subarray size, dilution settings for calibration curve, flags / print limits, a recommended number of repeats per sample. In examples where the method of acquisition comprises one or more preferred settings for analytic instruments used for analysis of the sample, the method may further comprise adjusting one or more settings on one or more analytical instruments based on the one or more preferred settings. Beneficially, this not only reduces the amount of analytical chemical knowledge required to determine an analytical protocol of a sample but reduces the amount of time to perform steps of the analytical protocol. Error is also reduced, for example, errors in calibration settings or errors as a result of a user not examining all the present analytically observable species of importance in the sample. The reduction of such errors avoids unnecessary extension of analysis time in addition to improving accuracy as analysis and reducing waste.

[0008] In some examples, a method of training a machine learning model to output, in response to an input of a spectrum of a sample, output data indicating a specific analytical protocol for the sample from a plurality of specific analytical protocols, each specific analytical protocol being adapted for a corresponding sample type is provided. The method comprises: obtaining a training data set comprising training data pairs, each training data pair comprising a spectrum obtained from a sample using abaseline analytical protocol and an indication of one of the specific analytical protocols to use for the sample, the training data comprising corresponding spectra for each one of the sample types obtained from samples of each one of the sample types; and adjusting parameters of the machine learning model to reduce a discrepancy between the indications of the training data pairs and the indications output by the machine learning model in response to the respective spectra of the training data pairs. The baseline analytical protocol is the same for all types of samples

[0009] In some examples, a method of obtaining training data for training a machine learning model is provided. The method comprises: for each of a plurality of specific analytical protocols, obtaining a baseline spectrum for each of a plurality of samples of a type for which the specific analytical protocol is applicable, the baseline spectra being obtained using a baseline analytical protocol the same for all of a plurality of types of samples; and storing pairs of input and output data for the machine learning model, the input data of each pair comprising an obtained baseline spectrum and the output data of each pair comprising an indication of the analytical protocol for the respective sample. The samples for the training data are selected to span the space of samples likely to be encountered when using the model.

[0010] According to another aspect of the present disclosure, there is provided a system comprising one or more processors and one or more memories having stored thereon computer-readable instructions configured to cause the one or more processors to perform any of the methods disclosed herein. In some examples, the system further comprises a spectrometer for recording a spectrum.

[0011] According to another aspect of the present disclosure, there is provided a computer program comprising instructions which, when the program is executed by a computer, cause the computer to perform any of the methods disclosed herein.BRIEF DESORPTION OF THE DRAWINGS

[0012] Disclosed implementations will now be described by way of example to illustrate aspects of the disclosure and with reference to the accompanying drawings, in which:

[0013] FIG. 1 shows an example of an optical spectrometer 100 for obtaining spectra of a sample;

[0014] FIG. 2 is a flow diagram of an example method 200 of obtaining training data for training a machine learning model in accordance with various examples;

[0015] FIG. 3 is a flow diagram of an example method 300 of training a machine learning model to output, in response to an input of a spectrum of a sample, output data indicating a specific analytical protocol for the sample, in accordance with various examples;

[0016] FIG. 4 is a flow diagram of an example method 400 of determining an analytical protocol for a sample, in accordance with various examples;

[0017] FIG. 5A shows an example of a spectrum for a sample containing acidified aluminium FIG. 5B shows an example of a spectrum for a sample containing acidified silicon.

[0018] FIG. 6 is an example of a graphical user interface that may be used in the performance of some or all of the support methods disclosed herein, in accordance with various embodiments;

[0019] FIG. 7 is a block diagram of an example computing device that may perform some or all of the scientific instrument support methods disclosed herein, in accordance with various embodiments;

[0020] FIG. 8 is a block diagram of an example scientific instrument support system in which some or all of the scientific instrument support methods disclosed herein may be performed, in accordance with various embodiments.DETAILED DESCRIPTION OF THE DRAWINGS

[0021] With reference to FIG. 1 , an example optical spectrometer 100 for obtaining spectra of a sample comprises a sample introduction system 102, an optical system 104, and a detector 106. Although an optical spectrometer is used here as an illustrative example, it is understood that other types of spectrometers may be used to obtain other types of spectrum to measure a characteristic of a sample over a given range. The present disclosure relates to methods that can be performed using data obtained from one or more different types of spectrometer (i.e. they are not spectrometer-specific, per se, as will be discussed in more detail with reference to FIG. 2-4). For example, the optical spectrometer 100 may be an inductively coupled plasma optical emission spectrometer (ICP-OES) which operates by exciting atoms and ions to emit electromagnetic (EM) radiation at wavelengths characteristic of a particular element. A spectrum of frequencies of EM radiation is emitted due to electrons making a transition from a high energy state to a lower energy state. In such cases, the sample introduction system 102 may comprise a plasma chamber, connected to a radio frequency (RF) voltage source and a source of gas, for example argon gas. The voltage source is used to apply a voltage to a coil which induces an RF electromagnetic field The argon gas may be ionized inside the plasma chamber by the electromagnetic field to develop and maintain plasma inside the plasma chamber. The optical system 104 may comprise an echelle diffraction grating, a prism and multiple focusing mirrors to selectively focus light from the plasma chamber and diffract the light into multiple diffraction orders, creating a high- resolution 2D spectrum known as an echellogram, also referred to as full-frame when detected by detector 106. The optical spectrometer 100 may be part of a scientific instrument support system, such as the scientific instrument support system 800 described with reference to FIG. 8 and hence, may be coupled to computing devices 820, 830, 840 comprising processing devices 802, storage devices 804 and interface devices 806. The optical spectrometer 100 requires numerous settings to be adjusted for each sample. For example, in the case of the ICP-OES, settings include the temperature of the optical system, the rate of spectrum acquisition and the RF power supplied to the plasma chamber. In addition to the settings of the spectrometer, analytical protocols comprise details of ancillary steps, such as sample preparation steps. As will be discussed in more detail with reference to FIG. 2-4, many different types of spectrometer may be used other than an ICP-OES. For example, inductively coupled plasma mass spectrometry (ICP-MS) or total ion chromatography for gas chromatography-mass spectrometry (TIC for GC-MS).

[0022] With reference to FIG 2-4, methods of determining an analytical protocol for a sample and training a machine model for providing such a determination are described. These methods reduce the amount of analytical chemistry literacy required for effectively operating a chromatography, mass and / oroptical spectrometry instrument such as the optical spectrometer 100 described above with reference to FIG. 1 . In order to train a machine learning model for determining a specific analytical protocol for a sample, training data for training the machine learning model is obtained.

[0023] FIG. 2 is a flow diagram of an example method 200 of obtaining training data for training a machine learning model in accordance with various examples. The training data set comprises training data pairs wherein each training data pair comprises a baseline spectrum obtained from a sample and a label indicating of one of a plurality of specific analytical protocols corresponding to the sample. For example, a label of ‘protocol 1 ’ may correspond to a first specific analytical protocol, a label of ‘protocol 2’ may correspond to a second specific analytical protocol and so on.

[0024] The specific analytical protocols are known protocols corresponding to respective sample types (unlike the baseline protocol which is not sample-specific) and may refer to sample-specific analytical techniques, storage settings, one or more scientific instruments, such as optical spectrometer 100 for analysing the sample and corresponding instrument calibrations. More specifically, the analytical protocol may specify one or more of: a method of acquisition of the sample, one or more analytic techniques for analysis of the sample, one or more preferred settings for analytic instruments used for analysis of the sample, one or more criteria for storing a sample, one or more analytic wavelengths ranges of functional groups of the sample, a concentration range for external standard preparation, one or more calibration factors. For example, if the analytical protocol specifies sample analysis on a mass spectrometer, the protocol may further comprise one or more sample introduction settings, the one or more sample introduction settings comprising a peristaltic pump speed, a tubing diameter and material, a nebulizer type, nebulizer gas flow, auxiliary gas flow, cooling gas flow and plasma power, one or more optical system settings, the one or more optical system settings comprising plasma view direction (axial or radial) , viewing height (if radial view is appropriate), exposure duration, subarray size. For example, if the analytical protocol specifies UV analysis, the analytical protocol may further comprise dilution settings for calibration curve, flags / print limits or a recommended number of repeats per sample. The concentration range for external standard preparation may refer the range of concentrations used to establish a relationship between intensity of emission lines observed in the ICP-OES spectra and the concentration of an analyte in the sample. The flag / print limits may indicate regulatory requirements, guidelines and detection limits. Limits may be Quality Control (QC) limits or recovery limits. QC limits require the that measured concentration of QC standards should fall within an acceptable range around the known value (e.g., ±5%). Recovery limits require that recovery studies (spiking known amounts of analyte into samples) should demonstrate recovery within an acceptable range (e.g., 90-110%). Matrix spike recovery requires that matrix spike recovery studies should demonstrate that the method can accurately measure analytes in complex sample matrices. LOD and LOQ respectively referto the lowest analyte concentration that can be distinguished from the assay background and the lowest concentration at which the analytics can be quantitated at defined levels for imprecision and accuracy. Stability limits indicate instrument stability over time to ensure that machine drift and fluctuations are within acceptable limits. Duplicate precision indicates that duplicate measurements of the same sample should show good precision (low relative standard deviation, often less than 5%).

[0025] Obtaining the training data may comprise recording baseline spectra on the scientific instrument, such as the optical spectrometer 100 described with reference to FIG.1 , or retrieving baseline spectra from a database or other storage system or device, such as those described below with reference to, for example FIG. 8. Example baseline spectra are shown in FIG. 5A and 5B which show example inductively coupled plasma optical emission spectra for samples of acidified aluminium and acidified silicon respectively.

[0026] Unlike the specific analytical protocols, the baseline protocol, with which the baseline spectra are obtained is the same for every sample of the plurality of sample types and hence is defined in such a way to produce spectra suitable for the purpose of identifying an analytical protocol for a plurality of different sample types. The baseline spectra are recorded in a consistent manner and on the same type of scientific instrument, such as an ICP-OES, according to the same pre-defined calibrations across the entire training data set. These baseline spectra, recorded on the same type of scientific instrument, are then associated with respective analytical protocols and the pairs of baseline spectra and analytical protocols used to train a machine learning model. The trained model may then be used to predict a specific analytical protocol to use with a respective new sample.

[0027] At step 202, the baseline spectrum is obtained for each sample of a plurality of samples of a type for which a given specific analytical protocol is applicable. For example, where there are 100 different specific analytical protocols labelled ‘protocol T to ‘protocol 100’, a plurality of samples are obtained for each of the 100 different specific analytical protocols and a baseline spectrum obtained for each sample. In some examples, the plurality may be in the region of 100s - 1000s but it is understood that this number is machine learning model specific and in some cases the model may permit fewer or require more samples in order to converge or meet a stopping condition, as described in more detail with reference to FIG.3.

[0028] In some examples, obtaining a baseline spectrum may require the sample to be dissolved in a solvent such as water or an organic compound. In such cases, an additional spectrum of the solvent may also be obtained (from a computing device or by recording a spectrum). Where a spectrum of the solvent is obtained, in some examples, the peaks corresponding to the solvent are removed from the spectrum of the sample in the solvent in order to obtain the baseline spectrum. This may be done by any appropriate method, for example, by determining equivalent peaks between the spectrum of the solvent and the spectrum of the sample in the solvent by aligning the spectrum of the solvent with the spectrum of the sample in the solvent. The peaks may then be removed by subtracting, from the spectrum of the unknown sample, the equivalent peaks. Of course, if a baseline spectrum is obtained in this manner for a type of sample (by subtracting the solvent peaks from the spectrum) the same process is used to acquire the baseline spectrum for this type of sample in inference. In some examples, solvent peak subtraction is applied to the baseline spectra of all sample types, irrespective of the sample type.

[0029] At step 204, pairs of input and output data are stored, wherein the input data comprises the obtained baseline spectrum and wherein the output data of each pair comprises a label of the analytical protocol for the respective sample (such as labels ‘protocol T, ‘protocol 2’ ‘protocol 3’ corresponding to analytical protocols one, two and three). The stored pairs of input and output data provide the trainingdata for the machine learning model. For example, the data set may comprise N data pairs corresponding to each of M different analytical protocols (i.e. a total of NxM data pairs).

[0030] FIG. 3 is a flow diagram of an example method 300 of training a machine learning model to output, in response to an input of a baseline spectrum of a sample, output data indicating a specific analytical protocol for the sample. The output data indicates a specific analytical protocol from a plurality of specific analytical protocols by means of a corresponding label, each specific analytical protocol being adapted for a corresponding sample type.

[0031] The machine learning model may be any appropriate architecture for processing spectral data. For many of the scientific instruments described herein, for example, the ICP-OES described with reference to FIG.1 , the obtained spectra are in the form of image data, as shown, for example, in FIG.5. In such cases, the machine learning model may be one of any appropriate known architectures of convolutional neural network (CNN) with convolutional and pooling layers that have an output layer adapted to be appropriate for image classification. For example, Res-Net (Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun: “Deep Residual Learning for Image Recognition”. arXiv preprint arXiv:1512.03385, 2015) is a convolutional neural network that is, in some implementations, trained according to method 300. In some implementations, the spectra are not in the form of image data, but instead in the form of, for example, a list of spectral peak coordinates, peak positions on a plot of intensity against mass / charge (m / z) or peak positions on a plot of time against m / z and can in general be one or two-dimensional. In these implementations, the disclosed methods still apply but use a machine learning model adapted to handle the data in question such that the disclosed methods apply to any type of spectra data in the appropriate respective implementation.

[0032] At step 302, the training data set comprising training data pairs is obtained, either using method 200 on-line or retrieving the result from a previous application of method 200 from digital storage.

[0033] At step 304 the parameters of the machine learning model are adjusted to reduce a discrepancy between the labels of the training data pairs and the labels output by the machine learning model in response to the respective spectra of the training data pairs. For example, the discrepancy may be represented by a loss function which represents differences between labels and predictions. Any appropriate loss function may be used, for example, cross-entropy or focal loss. The parameters may be adjusted using any appropriate optimization algorithm, for example stochastic gradient descent, such as Adam.

[0034] Step 304 may be performed until a stopping condition is met. For example, step 304 may be iteratively performed for a pre-determined number of iterations (in batches or epochs) or until a threshold discrepancy (for example, represented by a threshold on the magnitude of the loss function) is met, or both. For completeness, although common terms in the art, batch size and epoch are briefly described herein. The batch size of the model refers in this context to the number data pairs to processed before the parameters of the model are adjusted. The number of epochs defines the number of complete passes through the data set. Of course, it is understood that the stopping condition for step 304 is machine learning model dependent and therefore that any appropriate stopping condition may be used corresponding the machine learning model in question.

[0035] The trained machine learning model may be used to determine a specific analytical protocol for the sample in accordance with method 400 described with reference to FIG. 4The method 400 provides, to a user, any information that may be required to analyse, store or react a sample, regardless of whether the user knows the components in the sample.

[0036] At step 402, a baseline spectrum of a sample is obtained using the baseline analytical protocol. As discussed above with reference to the method 200 of obtaining the training data, the baseline analytical protocol is a set of pre-determined steps and instrument calibrations to be performed to get a baseline spectrum using a predefined scientific instrument.

[0037] At step 404, a machine learning model is provided. The machine learning model has been trained to output, in response to a spectrum of a sample, output data indicating an analytical protocol specific to that sample. The machine learning model may have been trained, for example, by the method 300 described above with reference to FIG. 3.

[0038] At step 406, the obtained baseline spectrum of the sample is applied as an input to the machine learning model and an output of the machine learning model is obtained. The output provides an indication of an analytical protocol to use with the sample, based on the baseline spectrum of the sample. The indication may be a label indicating the analytical protocol to use. For example, labels ’protocol T, ‘protocol 2’, ‘protocol 3’ may correspond to different respective analytical protocols that have been associated with (via the baseline spectra) with sodium salts with different halide components (fluorine, chlorine, bromine etc). In another example, classifications ’protocol 4’, ‘protocol 5’, ‘protocol 6’ may correspond to different analytical protocols that are associated with different concentrations of the same compound that require different dilutions for further analysis or different procedures for storing. Of course, these are simply illustrative examples and the indications of respective analytical protocols may in practice correspond to any analytical protocols for different types samples, associated with the different types of samples by the corresponding baseline spectra that the machine learning model is trained on.

[0039] At step 408, a specific analytical protocol for the sample is obtained, based on the output. For example, the label identifying one of the plurality of analytical protocols is used to search for a data record corresponding to the analytical protocol, for example, through a dictionary or look-up table or through any other appropriate method, such as any form of database access or query. In this manner, the label identifying the analytical protocol may be used to retrieve the sample-specific analytical techniques, storage settings, one or more scientific instruments, such as optical spectrometer 100 for analysing the sample and corresponding instrument calibrations corresponding to the specific analytical protocol label. In some examples, the protocol may be an analytical protocol from a scientific body or journal such that the specific analytical protocol comprises a method of acquisition, analysis and storing of the sample according to a regulation or standard in the industry.

[0040] In examples where the specific analytical protocol comprises one or more preferred settings for analytic instruments used for analysis of the sample, in some implementations the method further comprises adjusting one or more settings on the one or more analytical instruments based on the one or more preferred settings. The respective calibration instructions are obtained from the data record of the analytical protocol retrieved using the label identifying one of the plurality of specific analyticalprotocols. The calibration instructions may be flagged with an automation flag indicating that upon retrieval, the respective calibration instructions may be provided to one or more scientific instruments 810 from computing devices 820, 830, 840 via one or more communication pathways 808 shown in FIG. 8. The scientific instrument 810, the user local computing device 820, the service local computing device 830, and the remote computing device 840 may be in communication with other elements of the scientific instrument support system 800 via communication pathways 808 (as described in more detail with reference to FIG. 8). A scientific instrument control region 606 coupled to the instrument (described in more detail with reference to FIG. 6) may then control the calibration settings on the scientific instrument.

[0041] In other examples, the user may be provided with the calibration settings on a GUI 600 (FIG. 6) coupled to the spectrometer 810. In such examples, the user may be able to approve the calibration settings before the automatic update of calibration settings on the instrument 810 via the scientific instrument control region 606.

[0042] In a detailed example, an experiment was conducted to determine specific analytic protocols for aluminium, silicon and boron. Aluminium, silicon and boron have different properties, for example, they show peaks at different wavelengths and have different boiling points and consequently, have specific protocols that must be followed to obtain useful information about the purity, functional groups, etc of the samples.

[0043] A training data set was obtained using samples of acidified element forms of one of: aluminium; silicon; and boron. In addition, samples of a blank solution containing none of these elements were also used in the training data set. The purpose of the blank solution in this experiment was primarily to investigate the accuracy of the model but the blank solution could also be used to subtract the solvent peaks as described above with reference to step 202.

[0044] A total of 353 samples, comprising 102 samples of acidified aluminium, 81 samples of acidified silicon, 106 samples of acidified boron and 64 blank samples were used in the experiment. Baseline spectra were obtained for each of the 353 samples according to the same baseline protocol for every sample, where the baseline protocol specified pre-defined calibrations for an ICP-OES. Examples of the baseline spectra can be seen in FIG. 5A and FIG. 5B which show baseline spectra for a sample containing aluminum and silicon respectively.

[0045] Each of the 353 baseline spectra were labelled with a label indicating of one of a plurality of specific analytical protocols corresponding to the sample to obtain training data as discussed above with reference to FIG. 2, associating the acidified aluminium, acidified silicon and acidified boron samples with a respective analytical protocol for each type of sample.

[0046] The training data was split into a training set comprising 213 samples (roughly 60% of the data), a validation set comprising 70 samples (roughly 20% of the data) and a test set comprising 70 samples (roughly 20% of the data).

[0047] The model was then tested on the remaining 70 test samples. For each of the test samples, after a label indicating an analytical protocol was obtained as the output of the model, the analytical protocol was retrieved from a database and provided to the user, for example displayed on a display. The analytical protocol comprised information that may be required to analyse, store or react a sampleor provided to the control region of scientific instrument triggering the automatic update of calibration settings on a scientific instrument.

[0048] FIG. 6 depicts an example GUI 600 that may be used in the performance of some or all of the support methods disclosed herein, in accordance with various embodiments. As noted above, the GUI 600 may be provided on a display device (e.g., the display device 710 discussed herein with reference to FIG. 7) of a computing device (e.g., the computing device 700 discussed herein with reference to FIG. 7) of a scientific instrument support system (e.g., the scientific instrument support system 800 discussed herein with reference to FIG. 8), and a user may interact with the GUI 600 using any suitable input device (e.g., any of the input devices included in the other I / O devices 712 discussed herein with reference to FIG. 7) and input technique (e.g., movement of a cursor, motion capture, facial recognition, gesture detection, voice recognition, actuation of buttons, etc.).

[0049] The GUI 600 may include a data display region 602, a data analysis region 604, the scientific instrument control region 606, and a settings region 608. The particular number and arrangement of regions depicted in FIG. 6 is simply illustrative, and any number and arrangement of regions, including any desired features, may be included in the GUI 600.

[0050] The data display region 602 may display data generated by a scientific instrument (e.g., the scientific instrument 810 discussed herein with reference to FIG. 8). For example, the data display region 602 may display calibration settings such as temperature or wavelength.

[0051] The data analysis region 604 may display the results of data analysis (e.g., the results of analyzing the data illustrated in the data display region 602 and / or other data). For example, the data analysis region 604 may display wavelength readings or timings. In some embodiments, the data display region 602 and the data analysis region 604 may be combined in the GUI 600 (e.g., to include data output from a scientific instrument, and some analysis of the data, in a common graph or region).

[0052] The scientific instrument control region 606 may include options that allow the user to control a scientific instrument (e.g., the scientific instrument 810 discussed herein with reference to FIG. 8). For example, the scientific instrument control region 606 may include calibration controls.

[0053] The settings region 608 may include options that allow the user to control the features and functions of the GUI 600 (and / or other GUIs) and / or perform common computing operations with respect to the data display region 602 and data analysis region 604 (e.g., saving data on a storage device, such as the storage device 704 discussed herein with reference to FIG. 7, sending data to another user, labelling data, etc.). For example, the settings region 608 may include the option to retrieve stored spectra, for example spectra of a known solvent.

[0054] FIG. 7 is a block diagram of a computing device 700 that may perform some or all of the methods disclosed herein, in accordance with various embodiments.

[0055] The computing device 700 of FIG. 7 is illustrated as having a number of components, but any one or more of these components may be omitted or duplicated, as suitable for the application and setting. In some embodiments, some or all of the components included in the computing device 700 may be attached to one or more motherboards and enclosed in a housing (e.g., including plastic, metal, and / or other materials). In some embodiments, some these components may be fabricated onto a single system-on-a-chip (SoC) (e.g., an SoC may include one or more processing devices 702 and one ormore storage devices 704). Additionally, in various embodiments, the computing device 700 may not include one or more of the components illustrated in FIG. 7, but may include interface circuitry (not shown) for coupling to the one or more components using any suitable interface (e.g., a Universal Serial Bus (USB) interface, a High-Definition Multimedia Interface (HDMI) interface, a Controller Area Network (CAN) interface, a Serial Peripheral Interface (SPI) interface, an Ethernet interface, a wireless interface, or any other appropriate interface) . For example, the computing device 700 may not include a display device 710, but may include display device interface circuitry (e.g., a connector and driver circuitry) to which a display device 710 may be coupled.

[0056] The computing device 700 may include a processing device 702 (e.g., one or more processing devices). As used herein, the term "processing device" may refer to any device or portion of a device that processes electronic data from registers and / or memory to transform that electronic data into other electronic data that may be stored in registers and / or memory. The processing device 702 may include one or more digital signal processors (DSPs), application-specific integrated circuits (ASICs), central processing units (CPUs), graphics processing units (GPUs), cryptoprocessors (specialized processors that execute cryptographic algorithms within hardware), server processors, or any other suitable processing devices.

[0057] The computing device 700 may include a storage device 704 (e.g., one or more storage devices). The storage device 704 may include one or more memory devices such as random-access memory (RAM) (e.g., static RAM (SRAM) devices, magnetic RAM (MRAM) devices, dynamic RAM (DRAM) devices, resistive RAM (RRAM) devices, or conductive-bridging RAM (CBRAM) devices), hard drive-based memory devices, solid-state memory devices, networked drives, cloud drives, or any combination of memory devices. In some embodiments, the storage device 704 may include memory that shares a die with a processing device 702. In such an embodiment, the memory may be used as cache memory and may include embedded dynamic random-access memory (eDRAM) or spin transfer torque magnetic random access memory (STT-MRAM), for example. In some embodiments, the storage device 704 may include non-transitory computer readable media having instructions thereon that, when executed by one or more processing devices (e.g., the processing device 702), cause the computing device 700 to perform any appropriate ones of or portions of the methods disclosed herein.

[0058] The computing device 700 may include an interface device 706 (e.g., one or more interface devices 706). The interface device 706 may include one or more communication chips, connectors, and / or other hardware and software to govern communications between the computing device 700 and other computing devices. For example, the interface device 706 may include circuitry for managing wireless communications for the transfer of data to and from the computing device 700. The term "wireless" and its derivatives may be used to describe circuits, devices, systems, methods, techniques, communications channels, etc., that may communicate data through the use of modulated electromagnetic radiation through a nonsolid medium. The term does not imply that the associated devices do not contain any wires, although in some embodiments they might not. Circuitry included in the interface device 706 for managing wireless communications may implement any of a number of wireless standards or protocols, including but not limited to Institute for Electrical and Electronic Engineers (IEEE) standards including Wi-Fi (IEEE 802.11 family), IEEE 802.16 standards (e.g., IEEE802.16-2005 Amendment), Long-Term Evolution (LTE) project along with any amendments, updates, and / or revisions (e.g., advanced LTE project, ultra mobile broadband (UMB) project (also referred to as "3GPP2"), etc.). In some embodiments, circuitry included in the interface device 4006 for managing wireless communications may operate in accordance with a Global System for Mobile Communication (GSM), General Packet Radio Service (GPRS), Universal Mobile Telecommunications System (UMTS), High Speed Packet Access (HSPA), Evolved HSPA (E-HSPA), or LTE network. In some embodiments, circuitry included in the interface device 4006 for managing wireless communications may operate in accordance with Enhanced Data for GSM Evolution (EDGE), GSM EDGE Radio Access Network (GERAN), Universal Terrestrial Radio Access Network (UTRAN), or Evolved UTRAN (E-UTRAN). In some embodiments, circuitry included in the interface device 4006 for managing wireless communications may operate in accordance with Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Digital Enhanced Cordless Telecommunications (DECT), Evolution-Data Optimized (EV-DO), and derivatives thereof, as well as any other wireless protocols that are designated as 3G, 4G, 5G, and beyond. In some embodiments, the interface device 706 may include one or more antennas (e.g., one or more antenna arrays) to receipt and / or transmission of wireless communications.

[0059] In some embodiments, the interface device 706 may include circuitry for managing wired communications, such as electrical, optical, or any other suitable communication protocols. For example, the interface device 706 may include circuitry to support communications in accordance with Ethernet technologies. In some embodiments, the interface device 706 may support both wireless and wired communication, and / or may support multiple wired communication protocols and / or multiple wireless communication protocols. For example, a first set of circuitry of the interface device 706 may be dedicated to shorter-range wireless communications such as Wi-Fi or Bluetooth, and a second set of circuitry of the interface device 706 may be dedicated to longer-range wireless communications such as global positioning system (GPS), EDGE, GPRS, CDMA, WiMAX, LTE, EV-DO, or others. In some embodiments, a first set of circuitry of the interface device 4006 may be dedicated to wireless communications, and a second set of circuitry of the interface device 4006 may be dedicated to wired communications.

[0060] The computing device 700 may include battery / power circuitry 708. The battery / power circuitry 708 may include one or more energy storage devices (e.g., batteries or capacitors) and / or circuitry for coupling components of the computing device 700 to an energy source separate from the computing device 700 (e.g., AC line power).

[0061] The computing device 700 may include a display device 710 (e.g., multiple display devices). The display device 710 may include any visual indicators, such as a heads-up display, a computer monitor, a projector, a touchscreen display, a liquid crystal display (LCD), a light-emitting diode display, or a flat panel display.

[0062] The computing device 700 may include other input / output (I / O) devices 712. The other I / O devices 712 may include one or more audio output devices (e.g., speakers, headsets, earbuds, alarms, etc.), one or more audio input devices (e.g., microphones or microphone arrays), location devices (e.g., GPS devices in communication with a satellite-based system to receive a location of the computing device 4000, as known in the art), audio codecs, video codecs, printers, sensors (e.g., thermocouplesor other temperature sensors, humidity sensors, pressure sensors, vibration sensors, accelerometers, gyroscopes, etc.), image capture devices such as cameras, keyboards, cursor control devices such as a mouse, a stylus, a trackball, or a touchpad, bar code readers, Quick Response (QR) code readers, or radio frequency identification (RFID) readers, for example.

[0063] The computing device 700 may have any suitable form factor for its application and setting, such as a handheld or mobile computing device (e.g., a cell phone, a smart phone, a mobile internet device, a tablet computer, a laptop computer, a netbook computer, an ultrabook computer, a personal digital assistant (PDA), an ultra-mobile personal computer, etc.), a desktop computing device, or a server computing device or other networked computing component.

[0064] One or more computing devices implementing any of the scientific instrument support modules or methods disclosed herein may be part of a scientific instrument support system. FIG. 8 is a block diagram of an example scientific instrument support system 800 in which some or all of the scientific instrument support methods disclosed herein may be performed, in accordance with various embodiments. The scientific instrument support modules and methods disclosed herein (e.g., the scientific instrument 100 of FIG. 1 and the method 400 of FIG. 4) may be implemented by one or more of the scientific instrument 810, the user local computing device 820, the service local computing device 830, or the remote computing device 840 of the scientific instrument support system 800.

[0065] Any of the scientific instrument 810, the user local computing device 820, the service local computing device 830, or the remote computing device 840 may include any of the embodiments of the computing device 700 discussed herein with reference to FIG. 7, and any of the scientific instrument 810, the user local computing device 820, the service local computing device 830, or the remote computing device 840 may take the form of any appropriate ones of the embodiments of the computing device 700 discussed herein with reference to FIG. 7.

[0066] The scientific instrument 810, the user local computing device 820, the service local computing device 830, or the remote computing device 840 may each include a processing device 802, a storage device 804, and an interface device 806. The processing device 802 may take any suitable form, including the form of any of the processing devices 802 discussed herein with reference to FIG. 4, and the processing devices 802 included in different ones of the scientific instrument 810, the user local computing device 820, the service local computing device 830, or the remote computing device 840 may take the same form or different forms. The storage device 804 may take any suitable form, including the form of any of the storage devices 804 discussed herein with reference to FIG. 7 and the storage devices 804 included in different ones of the scientific instrument 810, the user local computing device 820, the service local computing device 830, or the remote computing device 840 may take the same form or different forms. The interface device 806 may take any suitable form, including the form of any of the interface devices 806 discussed herein with reference to FIG. 4, and the interface devices 806 included in different ones of the scientific instrument 810, the user local computing device 820, the service local computing device 830, or the remote computing device 840 may take the same form or different forms.

[0067] The scientific instrument 810, the user local computing device 820, the service local computing device 830, and the remote computing device 840 may be in communication with other elements of thescientific instrument support system 800 via communication pathways 808. The communication pathways 808 may communicatively couple the interface devices 806 of different ones of the elements of the scientific instrument support system 800, as shown, and may be wired or wireless communication pathways (e.g., in accordance with any of the communication techniques discussed herein with reference to the interface devices 706 of the computing device 700 of FIG. 7). The particular scientific instrument support system 800 depicted in FIG. 8 includes communication pathways between each pair of the scientific instrument 810, the user local computing device 820, the service local computing device 830, and the remote computing device 840, but this “fully connected” implementation is simply illustrative, and in various embodiments, various ones of the communication pathways 808 may be absent. For example, in some embodiments, a service local computing device 830 may not have a direct communication pathway 808 between its interface device 806 and the interface device 806 of the scientific instrument 810, but may instead communicate with the scientific instrument 810 via the communication pathway 808 between the service local computing device 830 and the user local computing device 820 and the communication pathway 808 between the user local computing device 820 and the scientific instrument 810.

[0068] The scientific instrument 810 may include any appropriate scientific instrument, such as a inductively coupled plasma optical emission spectrometer (ICP-OES), inductively coupled plasma mass spectrometer (ICP-MS) or total ion chromatography for gas chromatography-mass spectrometer (TIC for GC-MS).

[0069] The user local computing device 820 may be a computing device (e.g., in accordance with any of the embodiments of the computing device 700 discussed herein) that is local to a user of the scientific instrument 810. In some embodiments, the user local computing device 820 may also be local to the scientific instrument 810, but this need not be the case; for example, a user local computing device 820 that is in a user’s home or office may be remote from, but in communication with, the scientific instrument 810 so that the user may use the user local computing device 820 to control and / or access data from the scientific instrument 810. In some embodiments, the user local computing device 820 may be a laptop, smartphone, or tablet device. In some embodiments the user local computing device 820 may be a portable computing device. In some embodiments, the user local computing device 820 may perform the methods described herein with reference to FIG. 2, FIG. 3 and FIG. 4.

[0070] The service local computing device 830 may be a computing device (e.g., in accordance with any of the embodiments of the computing device 4000 discussed herein) that is local to an entity that services the scientific instrument 810. For example, the service local computing device 830 may be local to a manufacturer of the scientific instrument 810 or to a third-party service company. In some embodiments, the service local computing device 830 may communicate with the scientific instrument 810, the user local computing device 820, and / or the remote computing device 840 (e.g., via a direct communication pathway 808 or via multiple “indirect” communication pathways 808, as discussed above) to receive data regarding the operation of the scientific instrument 810, the user local computing device 820, and / or the remote computing device 840 (e.g., the results of self-tests of the scientific instrument 810, calibration coefficients used by the scientific instrument 810, the measurements of sensors associated with the scientific instrument 810, etc.). In some embodiments, the service localRECTIFIED SHEET (RULE 91) ISA / EPcomputing device 830 may communicate with the scientific instrument 810, the user local computing device 820, and / or the remote computing device 840 (e.g., via a direct communication pathway 808 or via multiple “indirect” communication pathways 808, as discussed above) to transmit data to the scientific instrument 810, the user local computing device 820, and / or the remote computing device 840 (e.g., to update programmed instructions, such as firmware, in the scientific instrument 810, to initiate the performance of test or calibration sequences in the scientific instrument 810, to update programmed instructions, such as software, in the user local computing device 820 or the remote computing device 840, etc.). A user of the scientific instrument 810 may utilize the scientific instrument 810 or the user local computing device 820 to communicate with the service local computing device 830 to report a problem with the scientific instrument 810 or the user local computing device 820, to request a visit from a technician to improve the operation of the scientific instrument 810, to order consumables or replacement parts associated with the scientific instrument 810, or for other purposes.

[0071] The remote computing device 840 may be a computing device (e.g., in accordance with any of the embodiments of the computing device 4000 discussed herein) that is remote from the scientific instrument 810 and / or from the user local computing device 820. In some embodiments, the remote computing device 840 may be included in a datacenter or other large-scale server environment. In some embodiments, the remote computing device 840 may include network-attached storage (e.g., as part of the storage device 804). The remote computing device 840 may store data generated by the scientific instrument 810, perform analyses of the data generated by the scientific instrument 810 (e.g., in accordance with programmed instructions), facilitate communication between the user local computing device 820 and the scientific instrument 810, and / or facilitate communication between the service local computing device 830 and the scientific instrument 810.

[0072] In some embodiments, one or more of the elements of the scientific instrument support system 800 illustrated in FIG. 8 may not be present. Further, in some embodiments, multiple ones of various ones of the elements of the scientific instrument support system 800 of FIG. 8 may be present. For example, a scientific instrument support system 800 may include multiple user local computing devices 820 (e.g., different user local computing devices 820 associated with different users or in different locations). In another example, a scientific instrument support system 800 may include multiple scientific instruments 810, all in communication with service local computing device 830 and / or a remote computing device 840; in such an embodiment, the service local computing device 830 may monitor these multiple scientific instruments 810, and the service local computing device 830 may cause updates or other information may be “broadcast” to multiple scientific instruments 810 at the same time. Different ones of the scientific instruments 810 in a scientific instrument support system 800 may be located close to one another (e.g., in the same room) or farther from one another (e.g., on different floors of a building, in different buildings, in different cities, etc.). In some embodiments, a scientific instrument 810 may be connected to an Internet-of-Things (loT) stack that allows for command and control of the scientific instrument 810 through a web-based application, a virtual or augmented reality application, a mobile application, and / or a desktop application. Any of these applications may be accessed by a user operating the user local computing device 820 in communication with the scientific instrument 810 by the intervening remote computing device 840. In some embodiments, a scientific instrument 810 mayRECTIFIED SHEET (RULE 91) ISA / EPbe sold by the manufacturer along with one or more associated user local computing devices 820 as part of a local scientific instrument computing unit 812.

[0073] In some embodiments, different ones of the scientific instruments 810 included in a scientific instrument support system 800 may be different types of scientific instruments 810. In some such embodiments, the remote computing device 840 and / or the user local computing device 820 may combine data from different types of scientific instruments 810 included in a scientific instrument support system 800.

Claims

CLAIMS1 . A computer-implemented method of determining a specific analytical protocol for a sample, the sample being of a type of a plurality of types of samples and each type of sample being associated with a corresponding specific analytical protocol, the method comprising: obtaining a baseline spectrum of a sample using a baseline analytical protocol, the baseline analytical protocol being the same for the plurality of types of samples; providing a machine learning model trained to output, in response to a spectrum of a sample, output data indicating a specific analytical protocol to use for the sample; applying the obtained spectrum of the sample, as an input to the machine learning model and obtaining an output of the machine learning model; and determining, based on the output, a specific analytical protocol for the sample.

2. The method of claim 1 wherein the baseline spectrum is one obtained using one of the following techniques: optical emission spectrometry, such as inductively coupled plasma optical emission spectrometry; mass spectrometry, such as inductively coupled plasma mass spectrometry; gas chromatography - mass spectrometry; or liquid chromatography - mass spectrometry, and wherein the machine learning model is trained on spectra obtained using the same technique.

3. The method of claims 1 or 2 further comprising: recording a further spectrum of the sample using the determined specific analytical protocol.

4. The method of any preceding claim, wherein obtaining the baseline spectrum of a sample comprises obtaining a spectrum of a sample in a solvent and removing peaks that correspond to the solvent from the spectrum.

5. The method of any preceding claim wherein the machine learning model comprises a Convolutional Neural Network.

6. The method of any preceding claim wherein the specific analytical protocol comprises: a method of acquisition of the sample.

7. The method of any preceding claim wherein the specific analytical protocol for the sample comprises one or more of:one or more analytical techniques for analysis of the sample; one or more preferred settings for analytic instruments used for analysis of the sample; one or more wavelength ranges of functional groups of the sample; a concentration range for external standard preparation; one or more calibration factors; one or more sample introduction settings, the one or more sample introduction settings comprising a peristaltic pump speed, a tubing diameter and material, a nebulizer type, nebulizer gas flow, auxiliary gas flow, cooling gas flow and plasma power; one or more optical system settings, the one or more optical system settings comprising plasma view direction (axial or radial), viewing height (if radial view is appropriate), exposure duration, subarray size; dilution settings for calibration curve; flags / print limits; and a recommended number of repeats per sample.

8. The method of claim 7, wherein the analytical protocol comprises one or more preferred settings for analytic instruments used for analysis of the sample, the method further comprising: adjusting one or more settings on one or more analytical instruments based on the one or more preferred settings.

9. The method of any preceding claim wherein the analytical protocol comprises: one or more criteria for storing the sample.

10. A method of training a machine learning model to output, in response to an input of a spectrum of a sample, output data indicating a specific analytical protocol for the sample from a plurality of specific analytical protocols, each specific analytical protocol being adapted for a corresponding sample type, the method comprising: obtaining a training data set comprising training data pairs, each training data pair comprising a spectrum obtained from a sample using a baseline analytical protocol and an indication of one of the specific analytical protocols to use for the sample, the training data comprising corresponding spectra for each one of the sample types obtained from samples of each one of the sample types, wherein the baseline analytical protocol is the same for all types of samples; and adjusting parameters of the machine learning model to reduce a discrepancy between the indications of the training data pairs and the indications output by the machine learning model in response to the respective spectra of the training data pairs.

11. A method of obtaining training data for training a machine learning model, the method comprising: for each of a plurality of specific analytical protocols, obtaining a baseline spectrum for each of a plurality of samples of a type for which the specific analytical protocol is applicable, the baseline spectra being obtained using a baseline analytical protocol, the baseline analytical protocol being the same for all of the plurality of types of samples; storing pairs of input and output data for the machine learning model, the input data of each pair comprising an obtained baseline spectrum and the output data of each pair comprising an indication of the analytical protocol for the respective sample.

12. One or more non-transitory computer readable media comprising instructions thereon that, when executed by one or more processing devices of a scientific instrument support system, cause the scientific instrument support apparatus to perform the method of any preceding claim.

13. A scientific instrument support system comprising: one or more processors; one or more memories having stored thereon computer readable instructions configured to cause the one or more processors to perform operations comprising the method of any of claims 1-11.

14. A scientific instrument comprising the scientific support system of claim 13 and further comprising a spectrometer for recording a spectrum.

Images