RU2782888C2 - Method for identification of microorganisms, using practical spectroscopy - Google Patents

Abstract

FIELD: laboratory diagnostics.

SUBSTANCE: method for identification of microorganisms and device for its implementation are proposed. The method includes the use of a Fourier transform IR method in an attenuated total reflection mode (FTIR-ATR), creation of precalculated models based on significant identification spectral characteristics for each species, subspecies, or subclass of the known microorganisms related to information about shapes, sizes, intensity, and peak absorption areas and correlations between them, obtaining a spectrum of an unknown sample and its analysis by comparison with precalculated models. At the same time, spectroscopic tracking of a level of sample drying is performed.

EFFECT: inventions provide the expansion of the arsenal of methods for identification of microorganisms in case of big databases with increase in the accuracy.

14 cl, 8 dwg, 5 ex

Description

The field of technology to which the invention belongs

The present invention relates to a method for identifying microorganisms using statistical methods, for example, multivariate spectral profile analysis or, alternatively, neural networks, to obtain data regarding the growth of microorganisms in a very short period of time compared to conventional methods.

In particular, the invention is based on the development of an analytical tool for identifying microorganisms by analyzing the ATR-FTIR spectral profile of an unknown sample and comparing the spectral profile with previously collected sample profiles and stored in a database.

State of the art

Fourier transform infrared spectroscopy (FTIR) is a non-destructive analysis method that provides information about the chemical composition of the analyzed sample. Since the early 1990s this method is used for the analysis of biological samples (Diem M. et al. Analyst (08.27.2005, 129(10): 880-885)).

At the same time, the ability of the FTIR method to identify and classify unknown microorganisms has been shown (Helm D. et al., Journal of General Microbiology (1991, 137, 69-79) and Marley L. et al., Vibrational Spectroscopy, 26, 2, 2001, 151-159).

Various species, subspecies, or subclasses of microorganisms are described by precise biochemical composition in terms of proteins, lipids, nucleic acids, and polysaccharides, as reflected in a specific vibrational spectrum (ed. Griggiths and Chalmers, 2001 Handbook of vibrational spectroscopy, John Wiley & sons, New York, volume 5).

Some potential applications proposed based on the FTIR method in microbiology include:

(i) identification of pathogens in the field of medicine and veterinary medicine, for example, in a clinical laboratory;

(ii) epidemiological studies, subject matter, pathogen screening, hygiene testing, explanation of chains of infection, control of therapy, and detection of recurrent infections;

(iii) characterization and screening of microorganisms from the environment;

(iv) monitoring of biotechnological processes;

(v) microbiological quality control in the food and pharmaceutical industries;

(vi) maintaining the collected strains.

Known methods for identifying microorganisms based on methods of comparative analysis of the spectrum of an unknown sample and the spectra of known types of microorganisms stored in the database.

Примеры известных способов идентификации такого типа изложены в документах US5660998A, CN103217398A, US6379920B1, WO2006002537A1, US9551654B2, US20170167973A1, WO2017/210783A1, Sousa et al., European Journal of Clinical Microbiology & Infectious Diseases (2014) 33: 1345-1353, Whittaker et al ., Journal of Microbiological Methods 55 (2003) 709-716, Wang et al., International Journal of Food Microbiology 167 (2013) 293-302.

However, in many known methods, the acquisition of spectra is performed in transmission, reflection or imaging modes.

For example, acquisition in transmission or reflection mode is set forth in US5660998 and US6379920 while imaging modes are set forth in WO2006002537A1, US9551654B2, US20170167973A1.

These methods have disadvantages in that the received signal is directly dependent on the thickness and structure of the sample, which may be too thick and cause saturation, or inhomogeneous and cause distortion due to scattering.

It is also known that methods based on imaging approaches, such as multi-pixel imaging, may have resolution limitations for individual spectra due to the lower signal to noise ratio obtained per pixel compared to single point detectors.

Another disadvantage of the current state of the art is that often the spectra can represent signals due to the presence of water or other contaminants that obscure the specific signals of microorganisms.

In some cases, for example, in CN 103217398A, drying procedures are used, but, as stated in WO 2017/210783A1, these procedures can have a significant and irreversible effect on microorganisms, as well as change their spectra.

In some cases, for example in WO 2017/210783 Al, it is possible to resort to multiple preparations to remove the water components present in the spectra. However, this solution has disadvantages related to the long analysis time and the greater complexity of the analytical procedures for comparing spectra.

Another disadvantage of the current state of the art is that often the reference database contains a limited number of microbial species and therefore does not cover the wide range of possible species.

For example, the database and the corresponding identification method described in CN103217398A refer to 13 kinds of bacteria.

In addition, with the increase in the size of reference databases, a number of problems arise related to the procedures for cultivation and cultivation of microorganism species, with the procedures for obtaining spectra and procedures for comparative analysis.

For example, as the size of the database increases, the time to analyze an unknown sample also increases given that the spectrum of the unknown sample needs to be compared with the large number of spectra contained in the database.

Moreover, the same species of microorganisms can have a large variability in spectral characteristics, which makes it difficult to compare them with the spectra contained in the database.

This variability may be due to the presence of different strains for each species or differences between different cultures of the same species due to possible variations in the chemical composition of the nutrient medium and/or growth conditions.

In this regard, there is a need to develop new methods for identifying microorganisms that can work for a large number of species, subspecies or subclasses of microorganisms, and with a large variability of the reference database.

Therefore, one of the objectives of the present invention is to provide a method for identifying microorganisms that is sufficiently general to be able to identify a large number of different species, subspecies or subclasses of microorganisms with high discrimination and accuracy.

Another object of the present invention is to provide a microorganism identification method that allows the identification of different strains of the same species, subspecies or subclass of microorganism possibly grown in different cultures on different media or under different environmental conditions in the presence or absence of biological fluids.

Another object of the present invention is to provide an accurate method for the identification of microorganisms, while requiring little time for each individual analysis, even in the case of large databases.

Another object of the present invention is to provide a device that can easily implement a method for identifying microorganisms by integrating all the various assay steps and instrumental components.

Applicant has developed, tested and carried out the present invention in order to overcome the shortcomings of the prior art and to achieve these and other objects and advantages.

The essence of the invention

The present invention is set forth and characterized in independent claims, while dependent claims describe other characteristics of the invention or variants of the basic inventive idea.

The present invention relates to a method for identifying microorganisms, and in particular to a method for identifying several species, subspecies and subclasses of microorganisms present in an unknown sample by comparing the infrared spectrum of said unknown sample with the infrared spectra of known samples that were previously stored in a database.

The method for identifying microorganisms in accordance with the present invention contains the following:

- the stage of preparing a database of reference spectra associated with known samples of species, subspecies and subclasses of known microorganisms;

- the step of creating pre-calculated models or reference libraries, starting from said database, based on the most significant identifying spectral characteristics, for example associated with information about shapes, sizes, intensity and peak absorption areas, as well as correlations and ratios between intensities or between different peak areas;

- one or more steps of taking an unknown sample taken from a patient, possibly subsequently grown on solid or liquid media and possibly also containing biological fluids;

- one or more stages of obtaining the spectrum of an unknown sample;

- one or more stages of processing the spectrum of an unknown sample;

- one or more steps of analyzing the spectrum of the unknown sample by comparing it with pre-calculated models, each of which gives at least some indicator of belonging, possibly in the form of percentages, of the unknown sample to one or more of the mentioned species, subspecies or subclasses of known microorganisms, as well as the reliability parameter of the mentioned indicators of belonging;

- one or more control steps, in which, based on the mentioned membership indicators and the mentioned confidence parameter, a new step of obtaining the spectrum of an unknown sample is requested or the final result is provided, which may contain an unsuccessful identification or a successful identification.

In some embodiments, the step of preparing a database of reference spectra associated with known species, subspecies, or subclasses of known microorganisms for each sample also includes the following:

- the step of selecting a known sample, possibly taken from patients, possibly subsequently grown on solid or liquid media, and possibly containing biological fluids;

- one or more stages of obtaining the spectrum of a known sample;

- one or more stages of processing the spectrum of a known sample.

In some embodiments, the database preparation step may be performed only once to create a database that can then be used for each subsequent analysis of unknown samples.

In some embodiments, the step of preparing the database may also refer to updating the database, which should be performed whenever new species, subspecies, or subclasses of microorganisms are desired to be included, thereby expanding the range of applicability of the method.

Similarly, pre-calculated models, once created, can be used for each subsequent analysis of unknown samples, or can also be updated each time new species, subspecies, or subclasses of micro-organisms need to be added.

In accordance with the present invention, spectra are obtained using an apparatus containing an FTIR-ATR spectrophotometer.

This acquisition mode has the advantage that it acquires spectra regardless of sample thickness and morphology, guaranteeing high reproducibility compared to methods based on transmittance or reflectance.

In addition, it allows you to work on an extremely small sample, as well as using very simple and fast working procedures.

In some embodiments of the present invention, the FTIR-ATR spectrophotometer comprises a single point detector that provides a better signal to noise ratio than, for example, multi-pixel and/or imaging-based methods.

This feature also makes it possible to limit or prevent scatter and signal saturation phenomena and to guarantee easier sample preparation.

In some embodiments, the spectrophotometer has a continuous acquisition mode that facilitates cleaning of the instrument as well as real-time diagnosis of the presence of water and/or other contaminants, thereby limiting a disadvantage of the state of the art where the presence of water and/or contaminants can override the specific signals of microorganisms.

Preferably, in accordance with the present invention, spectroscopic monitoring of the drying level of the sample is provided, so that the spectrum is obtained only when a predetermined standard drying level is reached.

Thus, the spectra obtained using this mode refer to samples that have almost the same drying levels and are therefore more comparable to each other.

This feature overcomes or at least limits some of the disadvantages of the state of the art, since no oven drying steps are required to remove water from the sample, thus avoiding irreversible effects on the microorganisms and therefore on the corresponding spectra.

In some embodiments, the present invention provides for the use of algorithms that automatically identify the most significant identifying spectral characteristics of samples in predetermined spectral ranges.

This feature can significantly reduce the time required to complete the analysis stage, which allows you to speed up the time and improve the accuracy of identification.

Preferably, the use of said pre-calculated models makes it possible to speed up the identification time for each unknown sample compared to methods known in the art.

These characteristics make it possible to increase the size of the database, allowing the identification of many species, subspecies or subclasses of different microorganisms, different strains of the same species, subspecies or subfamily, also possibly grown in different cultures on different media or in different environmental conditions, possibly in the presence of biological fluids.

Brief description of the drawings

These and other aspects, characteristics and advantages of the present invention will be understood with reference to the following description, drawings and the appended claims. The drawings, which form an integral part of the description, show some embodiments of the present invention and, together with the description, provide a description of the principles of the invention.

The present invention is better described and shown with the help of the following drawings, in which:

in fig. 1 is a schematic representation of a possible device according to the method of the present invention;

in fig. 2 is a flow chart showing, by way of example, the steps of one embodiment of a method in accordance with the present invention;

in fig. 3a and 3b show experimental data collected using the FTIR-ATR method, where FIG. 3a shows the collected raw data in which each species is represented by a different color, while FIG. 3b shows the converted data;

in fig. 4 schematically shows the zones of clusters obtained using PCA in the space of two principal components corresponding to the spectra of various types of microorganisms;

in fig. 5 schematically shows an example of cluster zones obtained using LDA in the space of the first two components;

in fig. 6 schematically shows an example of cluster zones obtained using PCA in the space of two principal components, corresponding to the spectra of various types of microorganisms;

in fig. 7 shows: panel a) a representation of the PCA space colored as Gram (+) (light grey) and Gram (-) (dark grey); panel b) is a confusion matrix showing the "calculated" data in relation to the "actual" data obtained in this way; panel c) - the results of the mathematical method applied to the embodiments described in this document when calculating the Gram-type;

in fig. 8 shows the confusion matrix obtained by cross-validating the method according to the options described in this document with database spectra.

To facilitate understanding, where possible, the same reference numbers have been used to indicate identical common elements throughout the drawings. It is clear that the elements and characteristics of one option can be included in other options for implementation without further clarification.

Detailed description of some embodiments of the invention

Let us now consider in detail the various embodiments of the present invention, one or more examples of which are shown in the accompanying drawings. Each example is given as an illustration of the invention and should not be understood as limiting it. For example, features shown or described as part of one embodiment may be adapted in or in combination with other embodiments to provide another embodiment. The present invention is intended to include all such modifications and variations.

Before describing these embodiments, we should also clarify that the present description is not limited in its application to the details of construction and arrangement of components as described in the following description using the accompanying drawings. The present description may contain other embodiments and may be obtained or performed in various other ways. We should also clarify that the phraseology and terminology used in this document is for descriptive purposes only and should not be considered restrictive.

Unless otherwise defined, all technical and scientific terms used hereinafter have the same meaning as generally understood by a person skilled in the art to which the present invention pertains.

While methods and materials similar or equivalent to those described herein may be used in practice or in tests to test this invention, the methods and materials are described below by way of example. In the event of a conflict, the present application containing the definitions shall prevail. The materials, methods, and examples are purely illustrative and are not to be understood as limiting.

In the following, by a sample we mean the minimum amount taken for the purposes of analysis from a large homogeneous part of a biological substance containing microbial organisms, for example, microorganisms, but not only.

Sometimes, if necessary, we will indicate an unknown sample or a known sample, in cases where the composition of the microorganisms of the biological substance is respectively unknown or known.

In some embodiments, species, subspecies, and subclasses of microorganisms may be of clinical interest.

In addition, the term "spectrum" will be used in relation to the set of electromagnetic emissions compatible with the acquisition tool used, and therefore the expressions "infrared spectrum", "vibrational spectrum", or references to vibrational transitions or certain specific spectral ranges should not be understood in a restrictive sense in regarding the applicability of the method in accordance with the present invention.

In FIG. 1 schematically shows a device 10 for identifying microorganisms in accordance with the present invention.

For example, the device contains a spectrophotometer for FTIR spectroscopy (Fourier transform infrared spectroscopy) in the ATR (frustrated total internal reflection - ATR) mode, that is, an FTIR-ATR spectrophotometer associated with a processing system and a data display system.

The processing system and the display system may, for example, be integrated into a single processing and display system and/or may be included in the computer 15, such as, for example, a personal computer equipped with a screen.

In some embodiments, the processing system includes a computer program that can be written to a computer-readable medium and that contains instructions that can be executed in each case by the device 10.

Further, for the sake of simplicity, reference will be made to the computer 15 to indicate the set of hardware and software systems capable of controlling the device 10, as well as processing and displaying data.

The FTIR-ATR spectrophotometer, known per se in its main components, comprises an infrared source 13 17, reflective elements 16, a total internal reflection element, and a detector 14.

For simplicity, other components of the spectrophotometer, such as monochromator, modulator, interferometer, are not shown in FIG. one.

In some embodiments, the source 13 may be a black body source of the globary type.

For example, the radiation 17 emitted by the source 13, ie the incident radiation 17a, is guided by the reflective elements 16 so that it hits the total internal reflection element.

In some embodiments, the element of total internal reflection may be, for example, a crystal 12.

In some embodiments, crystal 12 may be a high refractive index crystal 12.

In some embodiments, the implementation of the crystal 12 may be a crystal 12 of diamond, ZnSe, silicon or germanium.

Preferably, if the crystal 12 is a diamond crystal 12, it has the advantage that it is stronger than the ZnSe, silicon or germanium used in comparable measurements and has a larger transparency range.

The reflection of radiation 17 inside the crystal 12 creates a rapidly decaying field on the surface of the crystal 12, on which the zone 18 for obtaining information can be defined, on which the analyzed biological sample is located.

The rapidly decaying field can penetrate to various depths, which, depending on the case, can reach several microns.

In particular, the depth range depends on the angle of incidence and wavelength of the incident radiation 17a, as well as the refractive index of the material used for the crystal 12, and therefore can be considered almost constant for all analyzed samples.

For this reason, the preferred ATR mode produces spectra that are independent of the optical path through the sample, and therefore independent of the sample thickness, providing higher reproducibility compared to either transmission or reflection approaches.

In addition, this mode allows you to limit or prevent the phenomena of scattering and/or saturation of the signal, which can adversely affect the quality of the spectrum.

Thus, this mode guarantees greater measurement repeatability and easier sample preparation.

The radiation 17 emerging from the crystal 12 after interaction with the sample, that is, the outgoing radiation 17b, is then directed by means of the reflective elements 16 towards the detector 14.

In some embodiments, detector 14 may be a DLaTGS detector (Deuterated L-alanine doped with triglycine sulfate).

In alternative embodiments, detector 14 may be a DTGS detector.

For example, detector 14 converts optical information contained in outgoing radiation 17b into electrical signals that are transmitted to computer 15.

In some embodiments, the device is designed to acquire spectra in the NIR (Near Infrared) and/or MIR (Mid Infrared) and/or FIR (Far Infrared) region.

In some embodiments, the device 10 can operate in a continuous reception mode, that is, display on the screen of the computer 15 in real time what is received in each case in the zone 18 of obtaining information.

The present invention also relates to a method for identifying microorganisms, shown schematically in FIG. 2 and containing:

- the stage of preparing a database (A1, A2, A3) of reference spectra associated with known samples of species, subspecies and subclasses of known microorganisms;

- step A4 of creating pre-calculated models, starting from said database, based on the identification of the spectral characteristics most significant for each species, subspecies or subclass of known microorganisms;

- step B of the selection of an unknown sample taken from the patient, possibly grown subsequently on solid or liquid media and possibly containing biological fluids;

- one or more steps C to obtain the spectrum of an unknown sample;

- one or more stages D of processing the spectrum of an unknown sample;

- one or more steps of analyzing the spectrum of the unknown sample by comparing it with pre-calculated models, each step of the analysis giving at least some indication, possibly as a percentage, that the unknown sample belongs to one or more of the mentioned species, subspecies or subclasses of known microorganisms, as well as the reliability parameter of the mentioned belonging indicators;

- one or more control steps F, in which, starting from the mentioned membership indicators and the mentioned confidence parameter, a new stage of obtaining the spectrum of the unknown sample is requested or the end result of the method is provided, which may contain an unsuccessful identification of J or a successful identification of G, H, I, according to with different criteria.

In some embodiments of the present invention, step A of preparing the database may, in turn, comprise:

- step A1 of the selection of a known sample, possibly taken from patients, possibly subsequently grown on solid or liquid media, and possibly containing biological fluids;

- one or more steps A2 of obtaining the spectrum of a known sample;

- one or more stages A3 of processing the spectrum of a known sample.

In some embodiments, said steps A1, A2, A3 may be repeated each time a new pattern known in the database needs to be inserted.

In some embodiments, the sampling steps B, A1 of the sample may include that the sample is taken from a patient.

In some embodiments, a patient sample may be subjected to a preliminary analysis procedure that includes the addition of microbial nutrients.

In some embodiments, the implementation of the sample can be grown on a solid nutrient medium.

In some embodiments, the implementation of the sample can be grown in Petri dishes.

In some embodiments, the sample can be grown on a liquid nutrient medium and then centrifuged to obtain a concentrated bead.

In some embodiments, the sample may be grown in liquid growth media or liquid growth broth and then filtered.

In some embodiments, the sample may be grown on liquid growth media or liquid growth broth, or it may be subjected to concentration or enrichment procedures to increase the concentration of microorganisms present.

In some embodiments, the sample may contain body fluids of human or animal origin.

In some embodiments, spectrum acquisition steps C, A2 may include a preliminary step of cleaning device surfaces in contact with the sample, such as the acquisition zone 18 shown in FIG. one.

In some embodiments, background spectra can be obtained, for example, using a continuous acquisition mode, to check the effectiveness of such cleaning, for example, by observing the disappearance of absorption bands associated with impurities deposited on the acquisition zone 18.

In some embodiments, steps C, A2 of obtaining a spectrum also provide that a sample of the sample is taken and deposited on the crystal 12 of the spectrophotometer, for example, on the zone 18 of the acquisition, in order to obtain a spectrum.

In some embodiments, the sample is deposited on the receiving zone 18 in solid form, for example, by removing and depositing it with a disposable rod.

In some embodiments, the implementation of the sample can be pressed against the zone 18 receiving, for example, using a torque press.

In some embodiments, the level of drying of the sample is monitored spectroscopically.

In some embodiments, it is possible to estimate the content of water leaving the device in a continuous mode of production, and control the reduction, up to the possible complete disappearance, of the absorption bands related to the spectral characteristics of water in the respective ranges of wave numbers.

In accordance with the present invention, a spectrum is obtained upon reaching a predetermined standard level of sample drying.

Advantageously, this feature overcomes or at least limits the problem of the prior art whereby spectra can have signals due to the presence of water that override or interfere with the specific signals of microorganisms.

This feature also avoids sample drying procedures that could irreversibly modify the organisms being analyzed.

Spectrum acquisition steps C, A2 also record the spectrum of the sample taken from the sample.

In some embodiments, the implementation of the spectrum is recorded for a predetermined period of time.

In some embodiments, the implementation of the spectrum is recorded for a period of time less than 30 seconds.

In some embodiments, recording a spectrum provides a predetermined number of consecutive spectra, which are then averaged to improve the signal-to-noise ratio.

In some embodiments, 8 to 512 spectra per sample can be acquired and averaged, preferably 32 to 256 per sample, even more preferably 64 to 128 per sample.

In some embodiments, the spectrum, or possibly the average of the spectra, is displayed on a screen.

Steps D, A3 for processing the spectrum or averaging the spectra may include the following:

- use of linear and/or non-linear interpolation and/or fitting algorithms (spectral profile);

- calculation of the first and/or second derivative of the spectral profile;

- normalization of derivatives using the vector normalization algorithm over the entire spectral range;

- selection of the most useful spectral zones for the classification of species, subspecies or subclasses of microorganisms.

For example, the spectral bands may include: a range of 950-1280 cm ^-1 for nucleic acids, carbohydrates and polysaccharides; 1280-1480 cm ^-1 for starches, such as proteins, methyls and methylenes, such as lipids; 1700-1800 cm ^-1 for carbonyl groups, such as lipids; 2800-3000 cm ^-1 for aliphatic chains, such as lipids.

In some embodiments, artificial learning (machine learning) algorithms may be used to improve the spectral band selection process.

For example, in FIG. 3a shows the spectra that can be collected from various samples, while FIG. 3b shows the processed spectra.

In FIG. 3 also shows, by way of example, some identification spectral characteristics that are relevant for the purposes of the method according to the present invention.

In some embodiments, the implementation of the spectral regions and identification spectral characteristics are then used to classify and subsequently identify individual species, subspecies or subclasses of microorganisms.

In some embodiments, an automatic recognition system is provided to identify spectral ranges and identify spectral features of greatest interest.

Advantageously, the presence of different chemical classes, such as, for example, nucleic acids, lipids, proteins, carbohydrates and other microorganism constituents, makes it possible to obtain characteristic signals for each microbial species, subspecies or subclass, so that a calculation method can be obtained.

By repeating steps A1, A2, A3 for several known samples, a database of reference spectra of species, subspecies or subclasses of known microorganisms can be prepared.

However, some embodiments of the present method also provide that the database may contain spectra of species, subspecies or subclasses of known microorganisms obtained using other methods.

In some embodiments, the database contains spectra related to monomicrobial cultures belonging to various species, subspecies, or subclasses of known microorganisms.

In some embodiments, the database contains spectra related to different strains of known microorganisms for each species, subspecies, or subclass of known microorganisms.

In some embodiments, the database contains spectra related to samples grown on various nutrient media, for example, but not limited to, agar-agar, CNA agar, CLED agar, blood agar, chromogenic agar.

In some embodiments, the database contains spectra relating to samples grown in liquid phase growth broth, which are then centrifuged, filtered, or enriched to obtain pellets or concentrated samples.

Preferably, spectra obtained by broth growth and subsequent granulation or filtration can be used as standard reference spectra, since they do not contain interference from solid phase nutrient media matrices.

Advantageously, measurements of spectra obtained from broth culture and subsequent granulation, filtration or enrichment allow the spectra to be identified directly in the presence of biological fluids, such as body fluids.

Such heterogeneity and variability of the spectra contained in the database allows for a general analysis that allows the identification of many species, subspecies or subclasses of microorganisms, many different strains for each individual species, subspecies or subclass, including also the effects that occur when growing on different nutrient media.

In embodiments that use multiple acquisitions of each spectrum, individual repeated spectra and/or an average spectrum calculated over repetitions can be inserted into the database.

The pre-calculated models or reference libraries created or updated in step A4 may be based on the most significant identifying spectral features for analysis purposes in the selected spectral ranges.

In some embodiments, the most significant identifying spectral characteristics may be related to, for example, information about the shapes, sizes, intensities, and areas of absorption peaks, as well as correlations and ratios between intensities or between areas of different peaks.

In some embodiments, the pre-calculated models may contain a list of identifying spectral characteristics that are significant for each known sample of species, subspecies, or subclasses of known microorganisms.

In some embodiments, the implementation of the most significant identification spectral characteristics can be identified using the mentioned methods of statistical analysis and/or approaches based on neural networks and/or artificial learning.

These pre-calculated models allow you to quickly and automatically identify the most significant identification spectral features for comparison purposes.

Thus, each time two spectra are compared, the comparison can be made only in the spectral ranges and/or in relation to the spectral features of greatest interest, and not in the entire spectral range and/or across all spectral features.

For example, if one wanted to compare, over the entire spectral range, the spectrum of an unknown sample obtained at a resolution of 1 cm ^-1 with a database containing, for example, 18,000 reference spectra obtained at the same resolution, then this comparison is performed using prior art methods would require an evaluation of 65 million points.

The Applicant has found that when using the methods of the present invention, operating on selected spectral bands and with pre-calculated models, this comparison can require 20 to 80 times fewer point estimates than prior art methods.

Thus, the use of pre-calculated models can significantly reduce the analysis time for each unknown sample, even in the presence of large databases.

Thus, this feature allows the size of databases to be substantially increased to contain a large number of species, subspecies or subclasses of known microorganisms, thereby improving and expanding the predictive and identification capabilities of the method of the present invention.

This feature also makes it possible to work with spectra even at high resolution, increasing the reliability and accuracy of the method.

Moreover, for example, the presence of a list may make it redundant to apply least squares and/or possible calculations of standard deviations over the entire range of wavenumbers for all spectra.

In other embodiments, it is proposed to use artificial learning (machine learning) and artificial intelligence models to improve the automatic recognition system and pre-calculated models gradually as the method is used.

In some embodiments, analysis step E involves comparing the spectrum of the unknown sample with reference spectra contained in a reference database and/or with pre-calculated models.

In some embodiments of the present invention, statistical methods may be used for such comparison.

In some embodiments of the present invention, statistical methods may include multivariate analysis, such as principal component analysis (PCA) or linear discriminant analysis (LDA), as well as linear discriminant partial least squares (LDPLS), quadratic discriminant analysis (QD).

In some embodiments, methods, techniques, or algorithms that implement approaches based on neural networks may be used for said comparison.

In some embodiments, methods, techniques, or algorithms that implement approaches based on artificial learning (machine learning) may be used for said comparison.

In some embodiments of the present invention, said comparison may also provide a comparison of first and/or second derived spectra.

In some embodiments, statistical weights can be used to weight different regions of the spectra in different ways.

In some embodiments of the present invention, a comparison between two spectra can be performed by determining the distance between the two spectra, such as using least squares based methods.

In some embodiments, distance may be used as a metric for statistical and/or chemometric analysis.

In some embodiments, the implementation of the dispersion between the spectra can be used for PCA analysis.

In some embodiments, analysis step E provides an indication, such as a percentage, of the unknown specimen belonging to one or more of the species, subspecies, or subclasses represented in the database.

In some embodiments, analysis step E provides a confidence parameter that evaluates the confidence of the analysis and, in particular, the membership score.

In some embodiments, the control step F provides for the processing of the membership score and the confidence parameter.

In accordance with some embodiments, a membership score is automatically considered significant if the assay confidence parameter is above a predetermined acceptance level. For example, if percentage scores are used, then the acceptance level can be selected in the range of 80% to 100% (box G in FIG. 2).

In some embodiments, if it is not possible to uniquely identify unknown species, subspecies, or subclasses of microorganisms on the first try, then steps C, D, E, and F can be repeated to add new data and refine the result of the analysis.

In some embodiments, this may be done if the confidence parameter is below a second predetermined acceptance level. For example, in embodiments using percentages, this second level may be selected in the range of 70% to 80%.

In this case, the method includes a second preparation step C, a second processing step D and a second analysis step E to be performed on a second sample taken from an unknown sample.

For example, the second sample may be a different colony of microorganisms taken from the same Petri dish in which the unknown sample was cultured and grown.

If the sample is in the form of granules, then a second sample can be taken, for example, using a disposable rod from the same concentrated granule from which the first sample was taken.

Thereafter, if the reliability of the result of the second analysis is again below the third predetermined acceptance level, for example numerically equal to the second acceptance level, then the method involves comparing the similarity between the two spectra obtained in the first acquisition step C and in the second acquisition step C.

If it is established that the spectra are similar (block II in Fig. 2), then the method as the result of the analysis gives indicators of belonging of the unknown sample to one or more species, subspecies or subclasses of microorganisms in accordance with two preparations.

If the spectra are found to be inconsistent, then a third analysis is performed and a third acquisition step C, a third processing step D and a third analysis step E are performed.

If at least two of the three spectra turned out to be similar (block I in Fig. 2), then the method as a result of the analysis gives indicators of the sample belonging to the species, subspecies or subclasses of microorganisms from the database, in accordance with two of the three spectra.

Otherwise (block J in Fig. 2), a message about unsuccessful identification is displayed along with the spectra and three membership indicators related to the three acquisitions.

In one embodiment, a computer program or software is provided that is used to collect data and analyze it against a database that can be recorded on a computer-readable medium and that contains instructions that, when executed by an analysis device for classifying and identifying microorganisms, the method is executed in in accordance with this description.

Example 1

For example, in FIG. 4 schematically shows the results of computer analysis A of the spectra contained in the database, in accordance with the present invention. In particular, each point in FIG. 4 corresponds to the spectrum or the average of the obtained spectra related to species, subspecies or subclasses of microorganisms, in accordance with the first two main components.

The spectra were also grouped into zones, the expansion of which is associated with the variability of the spectra of each species, subspecies or subclass of microorganisms, based on different strains and/or different nutrient media for each sample.

Example 2

For example, in FIG. 5 schematically shows the results of an LDA analysis of the spectra contained in the database, in accordance with the present invention.

In particular, each point in FIG. 5 corresponds to the spectrum or the average of the obtained spectra, related to species, subspecies or subclasses of microorganisms, in accordance with the first two components.

The spectra were also grouped into zones, the extent of which is associated with the variability of the spectra of each species, subspecies or subclass of microorganisms, based on different strains and/or different nutrient media for each sample.

Example 3

In one embodiment, a database in accordance with the present invention was obtained by inoculating selected and certified microorganisms of the NNCCLS Atlanta type in Petri dishes or liquid media.

In this embodiment, specific spectra were obtained for each individual microorganism seeded on a Petri dish or liquid medium; in addition, each individual strain was inoculated into different culture media using Petri dishes from different liquid media manufacturers to check the variability of the spectra with respect to growth conditions.

The media used were selected from the following:

- CLED agar

- MacConkey agar

- CNA agar

- blood agar.

For each microorganism, 500 microorganism cultures were obtained in order to have a wide and complete case study in relation to obtaining a numerical cluster capable of checking the correlation of data compared to other methods used.

In FIG. Figure 6 shows the data obtained in the space of the first two principal components, where each species is represented by its own shade of gray.

Example 4

In one embodiment, the method in accordance with the invention can be used to predict the type of microorganism using the Gram method. An example of Gram + Gram classification results is shown in FIG. 7.

On panel a) in Fig. 7 shows a representation of the PCA space colored as Gram+ (light grey) and Gram- (dark grey).

In panel b) in FIG. 7 is a confusion matrix showing "calculated" data compared to "actual" data obtained by the method.

In panel c) in FIG. 7 shows the results of the mathematical method when calculating the Gram-type.

Example 5

In one embodiment, the method can be used to identify species, subspecies or subclasses of microorganisms. See fig. 8, which shows the confusion matrix resulting from testing the effectiveness of the method with database spectra in accordance with the present invention, a 99.9% to 96% agreement between the actual and calculated value can be observed in the matrix.

It is clear that changes and/or additions to parts or steps can be made to the method and/or apparatus described above without deviating from the scope and scope of the present invention. It is also clear that, although the present invention has been described with reference to some specific examples, a person skilled in the art will undoubtedly be able to obtain many other equivalent forms of a method and/or device having the characteristics set forth in the claims, and, therefore, all they fall under the scope of protection thus defined.

Claims (25)

1. A method for identifying microorganisms in a biological sample using Fourier transform IR spectroscopy in total internal reflection violation mode (ATR-FTIR), containing: - step (A1, A2, A3), which prepare a database of reference spectra associated with known samples of species, subspecies and subclasses of known microorganisms; - one or more steps (A4) in which pre-calculated models are created, based on a specified database, based on the most significant identification spectral characteristics for each species, subspecies or subclass of known microorganisms, while these characteristics are associated with information about shapes, sizes , intensities and peak absorption areas, as well as correlations and ratios between intensities or between different peak areas; - step (B), in which an unknown sample is taken, obtained from a culture grown on solid or liquid media, possibly with biological fluids; - one or more steps (C) where the spectrum of the unknown sample is obtained; - one or more steps (D) in which the spectrum of the unknown sample is processed; - one or more steps (E) in which the spectrum of the unknown sample is analyzed by comparing it with pre-calculated models, each of which gives at least some indicator of belonging, in the form of percentages, of the unknown sample to one or more of the indicated species, subspecies or subclasses of known microorganisms, as well as the reliability parameter of the indicated indicators of belonging; - one or more stages (F) of control, in which, based on the specified indicators of membership and the specified parameter of reliability, a new stage of obtaining the spectrum (C) of an unknown sample is requested or the final result of the method is provided, which contains an unsuccessful identification (J) or a successful identification ( G, H, I); wherein the acquisition steps (C, A2) perform spectroscopic monitoring of the drying level of the sample so that a spectrum is obtained only when a predetermined standard drying level is reached. 2. The method according to claim 1, characterized in that the steps (A1, A2, A3) of preparing a database of reference spectra associated with known species, subspecies or subclasses of known microorganisms for each known sample contain: - step (A1), which select a known sample obtained from a culture grown on solid or liquid media in the presence or absence of biological fluids; - one or more steps (A2) where the spectrum of a known sample is obtained; - one or more steps (A3) in which the spectrum of a known sample is processed. 3. The method according to claim 1 or 2, characterized in that the reference spectra contain spectra associated with samples of different species, subspecies or subclasses of microorganisms and / or different strains of the same species, subspecies or subclasses of microorganisms and / or grown on different nutrient environments with or without biological fluids. 4. The method according to any of the previous paragraphs, characterized in that the unknown sample is grown on a solid nutrient medium. 5. The method according to any of the preceding paragraphs, characterized in that at least one of these known samples is grown on a solid nutrient medium. 6. The method according to any one of the preceding claims, characterized in that said unknown sample is grown in at least one liquid growth broth and then centrifuged or filtered or enriched to obtain granules or a concentrated sample, wherein samples are isolated from said spectrum to obtain said spectrum. granules or concentrated sample. 7. The method according to any of the preceding claims, characterized in that at least one of said known samples is grown in at least one liquid growth broth, and then centrifuged or filtered or enriched to obtain granules or a concentrated sample, while to obtain of the specified spectrum, samples are isolated from the specified granule or concentrated sample. 8. A method according to any one of the preceding claims, characterized in that the processing step (D, A3) uses algorithms that automatically identify the most significant identifying spectral characteristics of the samples in predetermined spectral ranges. 9. Method according to any one of the preceding claims, characterized in that step (E) of the analysis uses statistical methods such as multivariate analysis. 10. A method according to any one of the preceding claims, characterized in that in step (E) of the analysis methods based on principal component analysis (PCA) are used. 11. A method according to any one of the preceding claims, characterized in that the analysis step (E) uses methods based on neural networks. 12. A method according to any one of the preceding claims, characterized in that in step (E) of the analysis methods are applied based on the analysis of clusters of spectra. 13. A method according to any one of the preceding claims, characterized in that said pre-calculated models contain a list of said identification spectral characteristics for each sample, both known and unknown. 14. A device for performing a method for identifying microorganisms in a biological sample according to any one of paragraphs. 1-13, containing a spectrophotometer for Fourier transform IR spectroscopy in the total internal reflection violation mode, associated with a processing system and a data display system in a computer (15), and the specified spectrophotometer for Fourier transform IR spectroscopy in the total internal reflection violation mode contains a source (13) infrared radiation (17), reflective elements (16), a total internal reflection element such as a crystal (12), and a detector (14); said device has an information acquisition zone (18) defined on the surface of said crystal (12), on which the analyzed sample is placed, and said computer (15) stores said pre-calculated models associated with known samples of species, subspecies and subclasses of known microorganisms, from a database, based on the most significant identifying spectral characteristics for each species, subspecies or subclass of known microorganisms, these characteristics being associated with information about shapes, sizes, intensities and peak absorption areas, as well as correlations and ratios between intensities or between different peak areas.

Images