EP4229397A1

EP4229397A1 - Method and fluorescence microscope for determining the location of individual fluorescent dye molecules by means of adaptive scanning

Info

Publication number: EP4229397A1
Application number: EP21794800.9A
Authority: EP
Inventors: Benjamin HARKE; Christian Wurm; Lars Kastrup; Roman Schmidt
Original assignee: Abberior Instruments GmbH
Current assignee: Abberior Instruments GmbH
Priority date: 2020-10-16
Filing date: 2021-10-15
Publication date: 2023-08-23
Also published as: DE102020127320B3; CN116391143A; US20230384223A1; WO2022079260A1

Abstract

The invention relates to a method for determining the spatial location of individual dye molecules (4, 5, 6) of a fluorescent dye (9, 11) in a highly precise manner by means of a scanning process with a scanning light (26) intensity distribution that has a local minimum. The method is characterized in that the scanning process is not carried out in a uniform manner for all dye molecules (4, 5, 6) but rather is adapted individually to each dye molecule (4, 5, 6) to be scanned and optionally to the surroundings thereof in the sample (27) in order to determine the location as precisely as possible with the smallest number of fluorescence photons possible. In the process, a raster image (1) of the sample (27) or of a section of the sample (27) comprising the dye molecule (4, 5, 6) is captured by means of laser scanning microscopy prior to scanning the dye molecule (4, 5, 6), and a scanning instruction for scanning the dye molecule (4, 5, 6) is determined from a signal of a pixel of the raster image (1) at the location of the dye molecule (4, 5, 6) or from the signals of multiple pixels of the raster image (1) in the surroundings of the dye molecule (4, 5, 6).

Description

METHOD AND FLUORESCENCE MICROSCOPE FOR LOCATING INDIVIDUAL FLUORESCENT DYE MOLECULES BY ADAPTIVE SCANNING

Technical field of the invention

The invention relates to high-resolution localization microscopy based on the MINFLUX principle. In particular, it relates to a method for customizing the scanning parameters when scanning individual dye molecules of a fluorescent dye at multiple scanning positions for the purpose of determining the location of the dye molecules. The invention further relates to a fluorescence microscope carrying out the method.

State of the art

The publication WO 2015/097000 A1 describes a method for locating spatially isolated, fluorescent dye molecules, which is now known under the acronym MINFLUX, in which each of the individual dye molecules is scanned at different positions with an intensity distribution of excitation light having an intensity minimum. The fluorescence emission excited by the excitation light is registered for each of the scanning positions, and the position of the respective molecule is deduced from the profile of the intensity of the fluorescence light along the positions of the intensity minimum. Naturally, this location determination is subject to errors, but the error in the location determination can be reduced by using the method iteratively. For this purpose, the scanning positions are adjusted before each iteration step, ie arranged closer to the respectively assumed location of the molecule. At the same time, the intensity of the excitation light is increased, so that the intensity gradient increases near the intensity minimum. Alternatively, the measurement time can be increased, which corresponds to an increase in the intensity of the excitation light with respect to the amount of effective light. With the adjusted parameters, the molecule is sequentially illuminated at each of the adjusted scan positions and the intensity of the fluorescence emission is registered. The position of the molecule can now be determined with less error than before from the dependence of the fluorescence signal on the positions of the intensity minimum. These process steps can until the convergence of Location determination or until another termination criterion is reached, for example a predetermined maximum acceptable error, are repeated. With an achievable localization accuracy of approx. 1 nm, the MINFLUX method represents the most precise commercially available localization method for fluorescent molecules according to the current state of the art.

The document WO 2015/097000 A1 further discloses that a (high-resolution) image of the distribution of the molecules in the sample can be obtained from the location data of the individual molecules (“MINFLUX imaging”). This method corresponds to the well-known procedure from STORM and PALM microscopy for generating high-resolution images from a large number of localizations of individual fluorescent molecules, but in the case of MINFLUX microscopy it results in a further increased spatial resolution of the images of 5 nm.

DE 10 2017 104 736 B3 describes a variant of the MINFLUX method in which the isolated fluorescent dye molecules present are not scanned by illuminating them with an intensity distribution of excitation light having a local intensity minimum, but with two essentially complementary intensity distributions of an excitation light and a fluorescence prevention light . In this case, the intensity distribution of the excitation light has a local intensity maximum, while the intensity distribution of the fluorescence-preventing light has a local intensity minimum at the same point. The fluorescence prevention light can specifically be STED light, which prevents fluorescence dye molecules excited in the edge regions of the intensity distribution of the excitation light from emitting fluorescence photons by triggering stimulated emission. In this embodiment of the method, the excitation light and the fluorescence prevention light are superimposed with such intensity distributions as is also the case in RESOLFT and STED microscopy. This variant of the MINFLUX method makes use of the fact that the intensity of the fluorescent light that is registered for the respective fluorescent dye molecule depends on its distance from the local intensity minimum of the fluorescence prevention light, and that its position can be determined with high accuracy from the multiple positions of the intensity minimum of the Fluorescence prevention light registered intensities of the fluorescent light can be determined. In this variant of the MINFLUX method, too, the local intensity minimum can be positioned at a few positions in the sample and the intensities of the registered fluorescent light can be evaluated according to the same principles as in MINFLUX microscopy. The difference remains, however, that in MINFLUX microscopy, the intensity of the fluorescent light from the fluorescent marker increases with increasing distance from its position to the Position of the local intensity minimum increases, while it decreases with increasing distance in the embodiment of the method in which the further light is fluorescence prevention light.

In order to acquire a high-resolution image using one of the MINFLUX methods described, it is necessary to scan a large number of dye molecules and determine their locations. Compared to other methods of localization microscopy, such as STORM or PALM microscopy, in which many dye molecules are imaged simultaneously on a camera and their locations are determined and which are therefore inherently parallelized, image acquisition using the MINFLUX method is comparatively slow, since--at least in the currently known embodiments--a position determination must be carried out in accordance with the procedure described above for each luminous dye molecule in succession. As a result, the sample is irradiated with light for relatively long periods of time, which can result in bleaching of the fluorescent dye or, in living samples, damage to the sample due to phototoxic effects.

In order to minimize these weaknesses of the MINFLUX method, it is desirable to scan the dye molecules according to an optimized scheme, so that the location can be determined in the shortest possible time and with the greatest possible photon efficiency. In particular, it is to be avoided that unnecessary scans are carried out. A related approach was developed by J. Pape et al. in "Multicolor 3D MINFLUX nanoscopy of mitochondrial MICOS proteins", Proc. national Acad. May be. USA 117 (34), 20607 (2020) by restricting the scanning of dye molecules to regions identified in a fluorescent camera image of the sample. This procedure avoids the scanning of dye molecules that can be attributed to an unspecific and undesired background staining in the sample.

It is known that the fluorescence emission of individual dye molecules occurs in so-called bursts, ie bursts of fluorescence photons that are limited in time and are interrupted by short pauses. The average fluorescence emission rate during a burst, the duration of a burst, the total number of fluorescence photons emitted in a burst, and the frequency and duration of the interruptions vary from dye to dye, but also depending on the binding state of a dye molecule or the composition of the medium ( e.g. the buffer) that surrounds the dye molecule. The intensity of the excitation light also has a significant influence on the characteristics of the bursts. Even fluorescent dyes that are structurally related and have similar spectral absorption and fluorescence properties (e.g. the commercially available cyanine dyes CF® 647 and AlexaFluor® 647) can differ significantly from each other in the characteristics of their bursts, whereby there is also a strong dependence on the composition of the medium.

The accuracy of localizing individual dye molecules using the MINFLUX principle is highly dependent on the total number of fluorescence photons a dye molecule emits before it bleaches or reverts to the non-fluorescent state. Ultimately, this number sets a limit for the achievable localization accuracy. However, not only the total number of fluorescence photons is decisive for the accuracy of the location determination, but also the distribution of the fluorescence photons over the different scanning positions. From the consideration of the limiting case that all fluorescence photons are emitted and detected at only one scanning location, it becomes immediately apparent that a location determination requires a (uniform) distribution of the fluorescence photons to the scanning positions. It follows from this that the scanning of the fluorescence molecules, in particular the number of scanning positions and the dwell time at each scanning position, must be adapted to the (average) total number of fluorescence photons as well as to the average duration of a burst. In any case, scanning all dye molecules according to a uniform scheme, regardless of their type and environmental conditions, is disadvantageous.

object of the invention

The object of the invention is to specify a method that enables photon-efficient and speed-optimized data acquisition according to the MINFLUX principle. The aim of the method is, on the one hand, to adapt the scanning of individual dye molecules individually in such a way that the location is determined as precisely as possible with as few fluorescence photons as possible. On the other hand, the method shortens the data acquisition in that the scanning is carried out with scanning parameters tailored to the respective dye and that unnecessary scanning steps are avoided. At the same time, the method can be used to reduce the illumination of the sample and thus the fading of the fluorescent dye at the scanning location, but also in neighboring areas of the sample. Finally, reducing the light dose also reduces the risk of damage occurring in the sample due to phototoxic effects.

solution

The objects of the invention are achieved by a method according to independent claim 1 and by a fluorescence microscope according to independent claim 26 . The dependent claims 2 to 25 relate to preferred embodiments of the method, while dependent claims 27 to 30 relate to preferred embodiments of the fluorescence microscope.

Description of the invention

The invention is based on the finding that for photon-efficient and speed-optimized data acquisition according to the MINFLUX principle, the scanning of the dye molecules must be individually adapted to the type of dye and, if necessary, to the environmental conditions.

To this end, the invention includes a method for spatially highly precise location determination of individual dye molecules of one or more fluorescent dyes in a sample, the fluorescent dyes usually being coupled to a structure to be examined in the sample and serving to visualize this structure. According to the invention, each of the fluorescent dyes has a first, fluorescent state which, when excited with light of a suitable wavelength, emits fluorescence in a fluorescence wavelength range, and a second state which, when the first state is excited, has no or only negligible fluorescence emission in the fluorescence wavelength range of the first state shows. The second state is often a non-fluorescent dark state, which can absorb the excitation light if necessary, but does not show any fluorescence, but thermally emits the excitation energy again; to this extent the second state (also in the following description) is also referred to as the non-fluorescent state. Strictly speaking, however, non-fluorescence of the second state only refers to the fluorescence wavelength range of the first state of the same dye and to a given excitation wavelength. It is therefore explicitly not excluded that the second state of a dye is also fluorescent, but only when excited with a wavelength that differs from the wavelength for exciting the first state, and / or with a fluorescence emission in a range other than the fluorescence wavelength of the first state. The wavelength for exciting the second state of a dye can also be the wavelength with which the first, fluorescent state of another dye is excited.

The method includes creating a distribution of individual dye molecules in the first, fluorescent state. In the extreme case, the distribution can also comprise only a single dye molecule in the fluorescent state; In any case, however, it must be ensured that the distance between neighboring dye molecules of the distribution is above the optical diffraction limit, so that neighboring fluorescent dye molecules can be resolved in an optical image, ie can be recognized as separate objects. This distance requirement applies also for dye molecules of different dyes, provided that the dye molecules can be excited to fluoresce in the same fluorescence wavelength range with excitation light of the same wavelength.

The generation of isolated, fluorescent dye molecules in a sample is known from the state of the art for localization microscopy (PALM, STORM, etc.). Spatially separated fluorescent dye molecules can be generated, for example, by thinning out an ensemble of initially fluorescent dye molecules, i. H. by converting the majority of the molecules from the first, fluorescent to the second (non-fluorescent) state. Conversely, distributions of spatially separated, fluorescent dye molecules can also be produced by converting a small number of dye molecules into the fluorescent state if the dye is initially in the second (non-fluorescent) state. In this case, it is not necessary for all the dye molecules to be scanned to be produced at once; it is also possible to photoactivate individual, spatially isolated dye molecules also gradually or even one after the other. A dye molecule can then be selected by scanning the sample with photoactivation light until the fluorescence of a (single) dye molecule is detected.

Alternatively, it is also possible to generate a distribution of individual, spatially isolated fluorescent dye molecules by adjusting the rates of spontaneous transitions between the first, fluorescent and second (non-fluorescent) state in such a way that the reactions always react in a very small manner Number of dye molecules in the first, fluorescent state. Reaction rates can be adjusted, for example, by adjusting the composition of a buffer in which the sample is embedded.

To carry out the method, individual dye molecules in the fluorescent state are selected and scanned with a scanning light at a number of scanning positions in accordance with a scanning specification. The intensity distribution of the scanning light in the sample has a local minimum. If a scanning position is mentioned below, this means the position of this local minimum. The scanning positions are chosen such that the minimum of the intensity distribution is located at various positions around the (assumed) location of the dye molecule, the distance from the respective dye molecule typically not exceeding 250 nm. At least in order to be able to define the first two or three scanning positions, an approximate knowledge of the location of the respective dye molecule to be scanned is required in advance. This initial location estimation can be done, for example, by scanning the sample with the excitation light or from a previously recorded fluorescence image take place; specific methods for this can be found in the prior art for MINFLUX microscopy.

On the one hand, the scanning light can be excitation light, which excites those dye molecules that are in the fluorescent state to fluoresce. On the other hand, the scanning light can also be a fluorescence prevention light, by which is meant any type of light that prevents, reduces or completely suppresses the fluorescence emission of the dye. In particular, the fluorescence-preventing light can be stimulating light, which induces a stimulated emission of electronically excited dye molecules, as a result of which the dye molecules are converted (back) into the electronic ground state and are thus prevented from spontaneous fluorescence emission. Importantly, the scanning light modulates the fluorescence emission of the dye molecules as a function of its intensity, with either an increase in emission (if the scanning light is excitation light) or a weakening of emission (if the scanning light is fluorescence prevention light/stimulation light) being possible. If the scanning light is not itself the excitation light, the dye molecule to be scanned is additionally illuminated with an excitation light.

During the scanning of a dye molecule, the fluorescence of the dye molecule is detected at each scanning position. In this case, the fluorescence can be detected with a light detector in the photon counting mode, in particular with an avalanche photodiode operated in the Geiger mode, which can have a particularly high sensitivity. However, the method can also be implemented with a light detector that generates an output signal proportional to the fluorescence intensity, for example with a photomultiplier.

Finally, the method includes a location determination of the scanned dye molecule from the numbers of photons or the intensities of the fluorescent light and the scanning positions. This location determination is considerably more accurate than the initial location estimate, on the basis of which the (first) sampling positions were determined. It is particularly advantageous to continue the scanning with further scanning positions, the further scanning positions being determined on the basis of the improved estimate of the location of the dye molecule as a result of the location determination. The number of photons or intensities of the fluorescence light detected at the further scanning positions now allow a renewed determination of the location with again improved precision. These steps can be repeated until the position of the dye molecule is determined with an acceptable uncertainty or the position determination has converged, ie the uncertainty of the position determination is no longer reduced. The invention is based on the idea that high accuracy and rapid convergence of the position determination can be achieved in a short time and with as few fluorescence photons as possible if the scanning process is individually adapted to the dye molecule to be scanned. The method according to the invention therefore differs from the MINFLUX methods known from the prior art in that the scanning does not take place uniformly for all dye molecules, but that the scanning is individually adapted to the type of dye molecule concerned and possibly its environment in the sample . For this purpose, a scanning rule is individually determined for each dye molecule before the start of the scanning, according to which the scanning for this dye molecule is carried out. In particular, the number and location of the scanning positions, scanning durations assigned to the scanning positions, minimum photon numbers or minimum intensities of the fluorescent light assigned to the scanning positions, waiting times inserted between the scanning positions and intensities and/or wavelengths of the scanning light applied to the scanning positions can be specified by the scanning specification. A simple scanning specification for dye molecules of the fluorescent dye AlexaFluor® 647, derived from the prior art, could therefore be, for example:

1 . Set the diameter L of a scanning circle to 70 nm, center the scanning circle at the assumed location of the dye molecule.

2. Position the scanning light (excitation light, 640 nm) sequentially at four scanning positions evenly distributed on the scanning circle.

3. Wait 14 ps at each scan position, then illuminate the sample with 400 pW optical power of the scan light for 14 ps. Detect a number of fluorescence photons at each scan position.

In order to determine the scanning specification, a raster image of the sample or of a section of the sample containing the dye molecule to be scanned is recorded before the dye molecule is scanned. According to the invention, the raster image is recorded using a scanning laser microscopy method, ie by scanning the sample with focused laser light. The scanning rule to be used for a molecule to be scanned is now determined from the signal of a pixel of the raster image at the location of the dye molecule or from the signals of several pixels of the raster image in the vicinity of the dye molecule.

From the signal of the raster image at the location of the dye molecule, the type of dye can be deduced (if the sample contains different dyes), or it can be distinguished whether a dye molecule is bound to a structure in the sample or as a single dye molecule. The scanning specification can also be used to tailor the scanning of a dye molecule to its context in the sample. For example it may be necessary to minimize the exposure of the sample to excitation light if another fluorescent dye is present in the immediate vicinity of a dye molecule that would be bleached out by excitation light that is too intense. The scanning with the scanning light can also be carried out in areas of the sample in which fast dynamic processes are taking place, with a scanning rule optimized for the speed of scanning and localization, while in less dynamic areas of the sample a scanning rule is applied that accepts a lower Scanning speed allows a more accurate localization.

It is also conceivable to determine a scanning specification on the basis of the fluorescence photons detected during the scanning; however, without recourse to the raster image, reliable adjustment of the scanning process can only be made after a significant number of photons have been detected. If unfavorable scanning parameters have been used up to this point in time, these photons contribute only insignificantly to an improvement in localization and are lost for the measurement. In addition, the photons emitted by a single dye molecule do not provide any information about the context in which the dye molecule is located in the sample. For example, a single dye molecule that is part of the (targeted) staining of a structure of interest in the sample cannot easily be distinguished by its fluorescence emission from a like dye molecule that happens to be in a background region of the sample. In contrast, the raster image of the sample also provides information about the context of the dye molecule to be scanned and thus allows the dye molecule to be classified as belonging to a structure or the background, for example.

In using a laser scanning raster image, the present invention also differs from that of J. Pape et al. in proc. national Acad. May be. USA 117 (34), 20607 (2020) in which regions of the sample of interest for MINFLUX imaging were determined from a wide-field epifluorescence image. In order to be able to use a wide-field image to select a sample region of interest or to define a scanning scheme for a given dye molecule, the positioning of the scanning light must be precisely coordinated with the (camera) image of the wide-field image. With regard to aberrations, in particular over large image areas, this can only be achieved with sufficient accuracy by means of complex calibration measurements. Above all, however, a (confocal) laser scanning image has improved resolution compared to a wide-field image and, in particular, depth discrimination, which is necessary in order to be able to reliably classify dye molecules even if the dye molecules are part of the coloring of a dense, especially three-dimensional extended structure and are parts of the Intersect structure in different planes of the sample. Due to a lack of resolution and, in particular, a lack of depth discrimination, these cannot be distinguished, or only insufficiently so, in a wide-field image.

In a preferred embodiment of the method according to the invention, a set of sampling rules is already defined in advance, i. H. before the scanning of individual dye molecules is started, from which the most suitable scanning rule is selected for each dye molecule to be scanned. The selection of the sampling specification from the set is usually automated, with rule-based selection algorithms being able to be used. In view of the considerable progress in the field of artificial intelligence and deep learning, however, a neural network trained for this classification task, which processes the raster image or an input vector calculated from the raster image as input, is increasingly available for the selection of a sampling rule.

The scanning instructions for the amount can be tailored, for example, to dye molecules of different fluorescent dyes or to different dye molecules of the same dye that are present in the sample in different binding states or in different contexts of the sample. The set of scanning rules can also include a scanning rule according to which a dye molecule is not scanned at all but is skipped if analysis of the raster image at the location of the dye molecule shows that the dye molecule is not part of a structure of interest. Such a scanning specification makes sense with regard to speed and light-efficient scanning, in particular when a large number of dye molecules are to be scanned.

The sampling specifications can consist of fixed sets of sampling parameters; however, it is also possible for individual or all of the scanning parameters of a set to be calculated using a function or an algorithm, in which case it is then not necessarily necessary to calculate the scanning parameters for all scanning positions in advance and at once; Rather, scanning parameters can also be calculated successively with the algorithm or the calculation specification, optionally also including the photons detected at the previous scanning positions or the intensities of the fluorescent light, in the course of the scanning.

In an alternative embodiment of the method according to the invention, a closed set of scanning specifications is not defined in advance, but rather the scanning specification is determined individually for each dye molecule to be scanned. This can be done, for example, by calculating a function or by executing an algorithm, the intensities of one or more pixels of the raster image at the location or in the vicinity of the respective dye molecule being passed to the function or the algorithm as an argument. Even if the scanning instructions are defined initially, ie before the beginning of the scanning of a dye molecule, this does not mean that the scanning must be continued according to the same scanning instructions until a final position determination is available. Rather, it can be useful to check the scanning specification in the course of scanning a dye molecule or between the scanning of successive dye molecules and to adapt it if necessary. For this purpose, the selection of a scanning rule from a set of scanning rules or the calculation of a scanning rule according to a function or by an algorithm can be carried out again after a number of scanning positions, with the raster image also being recorded again and optionally also those detected at the scanning positions already scanned Numbers of photons or the intensities of fluorescence light can be taken into account.

The scanned raster image of the sample required to define the scanning specifications can be recorded in a particularly advantageous manner if a beam deflection unit provided for scanning the dye molecules and/or the scanning light is also used for the laser scanning or if a common beam path or partial beam path is used, whereby the Costs for a microscope implementing the method can be reduced. Above all, however, there is a fixed relationship between the pixels of the raster image and the scanning positions, so that there is no need to compensate for imaging errors in a separate optical imaging.

In the simplest case, the raster image recorded by means of laser scanning is a fluorescence image of the sample. In addition to the fluorescence intensity, the fluorescence image can also have a fluorescence parameter other than image contrast, for example a fluorescence lifetime. The fluorescence image can optionally be recorded using two- or multi-photon excitation, with depth discrimination also being provided in this case if the fluorescence signal is detected non-confocally, ie in particular without using a confocal pinhole in front of the detector. In order to obtain a high-contrast fluorescence image, it is preferred that the fluorescent dye for the recording of the raster image is present to a significant extent in the fluorescent state and that only after the recording of the raster image is a distribution of spatially separated dye molecules possible by converting a large part of the dye molecules into the non-fluorescent state is generated. This can be done, for example, by targeted photo-deactivation with the excitation light or a dedicated photo-deactivation light (possibly with a different wavelength). The photo deactivation can also already take place during the scanning image recording of the raster image by the excitation light. Alternatively, the raster image can also be the fluorescence image of another dye in the sample, in which case the fluorescence image can then be recorded with a different excitation wavelength and/or with a spectrally different detection range and thus independently of the dye molecules intended for scanning. Another dye can be used, for example, to counterstain and identify an area in the sample of interest for subsequent scanning, e.g. to label a cell or an organelle within a cell. Selectively binding fluorescent dyes are available for this purpose, for example DAPI for marking cell nuclei, so-called MitoT racker™ for marking mitochondria or lipophilic carbocyanines (Dil, DiO, DiR) for marking cell membranes. Such a counterstain allows, for example, a decision as to whether a dye molecule is in a region of interest in the sample. If appropriate, a dye molecule can also be assigned to a specific part of the sample (eg a cell organelle) and a suitable or optimized scanning specification can be defined on the basis of this assignment.

A particularly advantageous variant of the method can be implemented if the second state of the dye is also fluorescent, but can be excited (selectively) with a different wavelength or fluoresces in a different wavelength range. Such properties are exhibited, for example, by photoconvertible proteins known from the prior art, such as EosFP, mMaple or Dendra, which, with activation light in the blue-green spectral range (typically 400 nm - 490 nm), change from a green fluorescent state in the range around 510 nm emits can be converted to a red fluorescent state emitting around 580 nm. In terms of the present invention, the red state can be understood as the first, fluorescent state, while the green state represents the second, non-fluorescent state with respect to the “red” detection area. A raster image in the green state can first be recorded from a structure in the sample marked with such a fluorescent protein. Since the fluorescent protein is initially exclusively or predominantly in this state, a high-signal, high-contrast raster image of the marked structure can be obtained. Individual protein molecules of the fluorescent protein can then be converted to the red state, which is the fluorescent state for the purposes of the present invention, by illumination with photoactivation light. These converted protein molecules can now be scanned according to the method according to the invention and their locations can be determined, with the scanning rule used in each case being based on the previously recorded raster image. Because it is just a two-state dye, the stains for taking the raster image and scanning the protein molecules are inherently in the same place, which is not possible with two separate stains reach. There is also no need to convert a fluorescent dye between a dense distribution of the fluorescent state (to capture the grid image) and a highly sparse distribution (to scan individual dye molecules), nor is it necessary to second stain the structure with a fluorescent dye other than Counterstain to take the halftone image. Because the majority of the fluorescent protein remains in the green state during scanning, it is also possible to use laser scanning to take a raster image of the structure again in between, i.e. between scanning successive protein molecules, and thus adapt the scanning specifications to dynamic changes in the adapt to the sample.

The fluorescence image can also have a number of channels which differ from one another in terms of the excitation wavelength, the spectral fluorescence detection range and/or the direction of polarization of the excitation light or the fluorescence light. In addition to the obvious possibility of detecting several spectrally different counterstains in the sample, it can also make sense to detect the fluorescence of just one dye separately in two or more spectral wavelength ranges. Such a detection scheme is particularly suitable for the fluorescence detection of ratiometric indicator dyes with which parameters such as ion concentrations (e.g. Ca ²⁺ , Mg ²⁺ , Zn ²⁺ , Na ⁺ ), pH values or membrane potentials from the ratio of the fluorescence intensities in the separately detected Wavelength ranges can be determined. These parameters provide spatially resolved functional information about the sample that goes beyond the structural information, for example about the activity of synapses between neurons or about the opening state of ion channels in a cell membrane. The parameters can be determined individually for each pixel or also for groups of pixels and, like the fluorescence intensity, can be used to define individual scanning specifications for the dye molecules to be scanned. Likewise, dye molecules may be excluded from scanning if analysis of the parameters at or around a dye molecule's location reveals that that dye molecule cannot be assigned to a region or context of interest in the sample. In this sense, a functional selection can also be carried out in such a way that only dye molecules are scanned in the vicinity of which the parameters indicate an interesting state, for example of a cell or a cell organelle.

The fluorescence image can also show the fluorescence emission as a function of time, for example in order to depict the kinetics of the fading of the dye. From these kinetics, the photostability of the dye and - if the local dye concentration is known - the average number of emitted fluorescence photons per Dye molecule are closed, which is an important variable for determining the number and arrangement of the scanning positions and the scanning time per scanning position.

Although the raster image is in many cases a fluorescence image, the method according to the invention can also be carried out with raster images that have a different image contrast. In particular, second harmonic generation contrast (SHG contrast), third harmonic generation contrast (THG contrast), scattered light contrast, reflected light contrast, differential interference contrast (DIC contrast) and polarization contrast come into consideration. These contrast modes also enable a scanning specification to be assigned or determined on the basis of structures in the sample that do not have any fluorescence staining.

Various image operations can be applied to the raster image in order to determine a scanning rule from the raster image. The analysis can be done locally, i. H. only taking into account the pixel in which the dye molecule is located for which a scanning rule is to be derived. A local analysis is usually associated with very little computational effort; In the simplest case, only a threshold value is generated for each pixel of the raster image in order to determine whether or not a dye is present at the location of the respective image pixel. A decision can then be made, for example, whether a dye molecule present at this location is to be scanned or to continue with another dye molecule. Alternatively or additionally, a threshold value can also be formed from intensity ratios or correlation amplitudes, which are formed from a plurality of channels of the raster image or from corresponding pixels or image areas of the raster image and another raster image. An intensity ratio formed from two color channels provides information about the identity of one of several spectrally different fluorescent dyes, and a suitable scanning rule can be selected from a set of alternative scanning rules by thresholding the intensity ratio.

In particularly advantageous embodiments of the method, however, the raster image is analyzed taking into account the surroundings of the dye molecule for which a scanning specification is to be determined. Morphological operations, ie proximity operators, are then also used to analyze the raster image, of which only the basic operations erosion, dilatation, opening, closing are mentioned here as examples. For advanced methods of image segmentation and pattern or object recognition, the person skilled in the art can fall back on extensive prior art in the field of image processing. Even if the raster image has to be analyzed once before the scanning of a dye molecule begins, in order to define a scanning specification with which the scanning is started, the analysis can be repeated in the course of the scanning and the scanning specification can be updated. If, for example, it is not possible to unambiguously select a scanning rule from a set of scanning rules from the initially recorded raster image, it makes sense to first calculate probability values for each scanning rule in the set so that the dye molecule is optimally scanned with the respective scanning rule. The scanning can then be started with the scanning specification for which the highest probability was determined, while the probability values are updated in the course of the scanning and taking into account the fluorescence detected in the process, and the scanning is optionally continued with a different scanning specification. Such a situation can occur, for example, in a sample stained with two spectrally similar dyes, so that individual molecules of a dye can only be identified with certainty after a sufficient number of fluorescence photons have been detected.

While the core of the method according to the invention is a method for determining the locations of individual dye molecules in a sample, the method can be further developed such that a spatially high-resolution image is reconstructed from the location determinations of many dye molecules. For this purpose, the locations of many dye molecules obtained from the location determinations can be visualized in a two-dimensional representation, for example in the form of a two-dimensional histogram. The generation of such high-resolution representations is known to the person skilled in the art from the state of the art for localization microscopy or PALM/STROM microscopy and is particularly useful if the dye molecules form a fluorescent coloration of a structure in the sample. In a preferred form of representation, the high-resolution image is inserted into the raster image, so that the raster image forms a suitable context for the structures represented in high resolution.

In another development of the method according to the invention, the location of an individual dye molecule can be determined repeatedly over longer periods of time in order to generate a time-related representation of the location of the dye molecule in the form of a (movement) trajectory. Here, too, the trajectory can advantageously be displayed in the context of the raster image. The types of visualization mentioned can also be combined, for example trajectories can be displayed as overlays in a high-resolution image of the sample in order to identify possible interactions of structures in the sample that are reflected in a change in the mobility of the traced dye molecules. Another aspect of the invention is that the recording of the raster image of a region of the sample and the generation of a high-resolution image from the localization of the individual dye molecules provides information on different resolution scales and over different large areas of the sample and that the inventive method is inherently a multi-scale Allows imaging of the sample.

This aspect can be further developed such that the halftone image is selected by selecting an area in a preview image that represents a larger area of the sample than the halftone image. This preview image can be scanned, for example, by fast laser scanning of the sample, in particular with a large increment and/or a short integration time per pixel. Alternatively, however, the preview image can also be recorded using a different image recording method, in particular by direct imaging onto a camera. The image field of the preview image can in turn be selected by selection in an overview image that represents an (even) larger section of the sample or the entire sample. Such an overview image can also be recorded by quickly scanning the sample or by camera-based imaging, with the overview image optionally also being able to be combined from a number of partially overlapping individual images (stitching). By displaying the preview image in the context of the overview image, the raster image in the context of the preview image and the high-resolution image in the context of the preview image, a multi-scale representation of even larger objects, e.g. of cell clusters or entire organisms, can be generated, with the resolution of the image recording being locally adjusted the structures in the sample can be adjusted. The use of different color scales is particularly useful for this type of display in order to identify the high-resolution information in the context of the overview, preview and/or raster image. For example, the image components recorded by determining the location of individual dye molecules can be displayed in a false color display in a gray scale display of the overview image.

The invention further relates to a fluorescence microscope that is set up to carry out the method according to the invention. For this purpose, the fluorescence microscope comprises at least one light source for excitation light, with which the fluorescent dye in the sample can be excited to emit fluorescence, and a detector for detecting the fluorescence of the fluorescent dye. An avalanche photodiode operated in the photon counting mode, which can have a particularly high sensitivity, is particularly suitable as a detector. However, an analogue photomultiplier can also be used as a detector as long as it has sufficient sensitivity. The fluorescence microscope also has beam shaping means for forming an intensity distribution of a scanning light in the sample that has a local intensity minimum, wherein this scanning light can be either the excitation light or a fluorescence prevention light from another light source. Fluorescence prevention light is to be understood here as meaning any type of light that is suitable for preventing, reducing or completely suppressing the fluorescence emission of the fluorescent dye. In particular, the fluorescence preventing light may be stimulating light that induces stimulated emission of electronically excited dye molecules. If the scanning light is not itself the excitation light, the fluorescence microscope according to the invention thus has light sources for the excitation light and for the scanning or fluorescence prevention light.

Phase filters or programmable phase modulators (spatial light modulator, SLM) can be used, inter alia, as are familiar from STED microscopy, for example.

The fluorescence microscope also includes a scanning image recording unit, with which a raster image of the sample can be recorded by scanning the sample with focused excitation light along a (regular) raster. In this regard, it is therefore a conventional laser scanning microscope, which is preferably designed as a confocal or STED microscope with a fluorescence contrast.

According to the invention, the fluorescence microscope also has a scanning device with which the scanning light can be positioned in the sample and individual dye molecules can be scanned at a sequence of scanning positions. In particular, the scanning of the dye molecules requires the scanning light to be positioned in the sample with an accuracy of 1 nm or less. At the same time, positioning times in the microsecond range are preferred in order to be able to complete the scanning of a dye molecule at a sufficient number of scanning positions within the duration of a burst of fluorescence photons.

These requirements cannot be met, or only insufficiently, with mechanical beam deflection units such as galvo mirrors alone. Beam deflection units that do not require moving parts, for example electro-optical or electro-acoustic deflectors, are therefore suitable for the scanning device. The desired positioning times can be easily achieved with these, but the maximum deflection angles are very limited. For this reason, a preferred embodiment of the fluorescence microscope according to the invention In the beam path, there is both a galvo scanner, with which the beam can be positioned over larger image fields, and an electro-optical deflector, with which the (rapid) scanning of the individual dye molecules is carried out. In this case, it makes sense for the scanning device and the scanning image acquisition unit to share the galvano scanner, so that a further beam deflection device can be saved. As a result, there is automatically also a fixed spatial relationship between the pixels of the raster image and the scanning positions, and there is no need to compare a number of beam deflection devices with one another.

In a development, the fluorescence microscope also has means for merging a plurality of spatially partially overlapping raster images of the scanning image recording unit. These means include at least one image processing unit which can align and cross-fade the individual raster images into the combined raster image. The means can also include an adjustable sample table, with which the sample can be shifted between the recording of individual raster images, so that the combined raster image can also have an image field that exceeds the scanning area of the scanning image recording unit.

Further advantageous developments of the invention result from the patent claims, the description and the drawings. The advantages of features and combinations of several features of the invention described in the description are merely examples and can have an effect alternatively or cumulatively without the advantages necessarily having to be achieved by embodiments according to the invention. The following applies with regard to the disclosure content of the original application documents and the patent, without the subject-matter of the appended claims being changed as a result: further features can be found in the drawings. The combination of features of different embodiments of the invention or of features of different patent claims is also possible, deviating from the selected dependencies of the patent claims and is hereby suggested. This also applies to features that are shown in separate figures or are mentioned in their description. These features can also be combined with features of different patent claims. Likewise, features listed in the patent claims can be omitted for further embodiments of the invention.

The indefinite article "a" used in the patent claims and the description for a feature is to be understood in such a way that the number of versions of this feature can be exactly one or more, without there being an explicit one Use of the adverb "at least" requires. The features listed in the patent claims can optionally be supplemented by further features

Brief description of the figures

1 shows a flow chart of the method according to the invention.

2 shows the method according to the invention schematically.

3 shows a fluorescence microscope according to the invention.

4 shows a type of representation of the data collected using the method according to the invention.

Description of the figures

1 shows the method according to the invention in the form of a flow chart. Therein solid lines indicate the sequence of procedural steps, dotted lines an alternative/optional sequence of procedural steps and dashed lines the flow of data. In the first step, a raster image 1 of the sample is recorded by means of scanning image recording, on the basis of which later scanning instructions for scanning individual dye molecules are determined. As a prerequisite for the subsequent scanning of individual dye molecules, the dye molecules are separated in the second step, i. H. a thinning out of the dye in the fluorescent state, for example by illumination with photodeactivation light, until only isolated fluorescent dye molecules are present, the distance between which is above the optical diffraction limit. In particular, if the raster image 1 is a fluorescence image of another dye or if it uses a different image contrast (e.g. an SHG contrast), the acquisition of the raster image 1 and the separation of the dye molecules do not necessarily have to be carried out in the order given, but can also be done be done in reverse.

After the dye molecules of the fluorescent dye are present individually, a dye molecule in the fluorescent state is selected and a suitable scanning specification is determined for this dye molecule, with the raster image 1 being analyzed at the location or in the vicinity of the respective dye molecule to determine this. In particular, the number and the arrangement of the scanning positions and the scanning duration at the scanning positions are defined by the scanning specification. The dye molecules are then scanned according to the previously determined scanning specification, ie they are scanned with an intensity distribution of excitation light having a local minimum or with excitation light and an intensity distribution of fluorescence prevention light (e.g. stimulation light) is illuminated at each scanning position, in each case a number of fluorescence photons or alternatively a fluorescence intensity being detected. The number of detected fluorescence photons and the respective scanning position are stored as pairs of values 2 in a data memory 3 . Optionally, the scanning rule can be updated or re-determined after a dye molecule has been scanned at some of the scanning positions and the data recorded in the process indicate that the scanning rule used is not optimal. Finally, the location of the dye molecule is determined using a method related to triangulation, with the value pairs 2 of scanning positions and detected photon numbers from the data memory s being accessed for location determination. The process can optionally be repeated with additional dye molecules in the fluorescent state.

The step of determining the position can also be carried out downstream after the scanning of all dye molecules has been completed, in particular if the algorithm used for determining the position is computationally intensive and cannot be executed sufficiently quickly between the scanning of successive dye molecules.

2 shows the method according to the invention for determining the location of three dye molecules 4 , 5 , 6 within a cell 7 . Within the cell 7, filaments 8 (these may be, for example, actin, vimentin or another filamentous protein) are stained with a fluorescent dye 9, while a structure of the cell nucleus 10 (e.g. the nuclear pore complex) is labeled with another fluorescent dye 11. The fluorescent dyes 9 and 11 are different here, but have similar spectral properties, so that they can be excited with excitation light of the same wavelength and detected in the same detection range. The use of spectrally similar dyes has the advantage that a common scanning light beam and a common detector can be used for the scanning. In addition to the economic advantage, both color channels are intrinsically aligned with one another due to the joint use of the scanning means and do not have to be adjusted relative to one another.

Both dyes are initially in a fluorescent state 12 but can be converted to a non-fluorescent state. First, a raster image 1 is recorded from the cell 7 or here only a part of the cell 7, the raster image 1 preferably being a fluorescence image 13 detected confocally. Fluorescence image 13 shows filaments 8 in area 14 and cell nucleus 10 in another area 15. Areas 14 and 15, which represent binary masks here, are determined by image processing (not shown) from the raw data of the raster image. The image processing typically includes i) Dilation steps that ensure complete coverage of the respective structures, ii) C/os/ng steps, with which the areas are closed so that they are gap-free, and iii) thresholding, by which binary masks are generated from the gray levels.

After the raster image 1 has been recorded, the dye molecules of the fluorescent dyes 9, 11 are separated. For this purpose, the initially dense distribution of the dyes must be thinned out to such an extent that only individual, spatially separated and optically resolvable dye molecules 4, 5, 6 remain in the fluorescent state 12. This individualization can take place, for example, by illuminating with deactivation light, as a result of which the majority of the dye molecules are converted to the non-fluorescent state.

After separation, the dye molecules 4 , 5 , 6 can be scanned with scanning light to determine their location, a number of photons or an intensity of fluorescent light being detected at each scanning position 16 . According to the invention, the intensity distribution of the scanning light, which is excitation light in the illustrated embodiment of the method, has a local minimum in the sample. The scanning is shown schematically in the figure for the three dye molecules 4, 5, 6. The dye molecule 4 is in the area 14 and can therefore be assigned to the filaments 8 that are marked with the fluorescent dye 9 . A scanning specification 17 is selected for the fluorescent dye 9, which provides six scanning positions 16 of the intensity minimum on an equilateral hexagon 18. A location is determined from the number of fluorescence photons detected at these six scanning positions 16, as a result of which the location of the dye molecule 4 is determined with improved accuracy. The dye molecule 4 is now scanned again at six scanning positions 16 on a hexagon 19 that is now smaller and arranged more closely around the dye molecule 4, and a new location determination is carried out. This process can continue until the location has been determined with the desired accuracy or until the location estimate has converged.

The dye molecule 5 cannot be assigned to any of the areas 14, 15 and is therefore to be regarded as part of an undesired, non-specific staining of the sample. A scan of dye molecule 5 is therefore not undertaken and the scan continues with dye molecule 6. The dye molecule 6 is in the region 15 and can therefore be assigned to the cell nucleus 10, which is marked with the fluorescent dye 11. For this fluorescent dye 11, a scanning specification 20 is selected which provides three scanning positions 21 of the intensity minimum on an equilateral triangle 22. The scanning is carried out according to the same principle as for the dye molecule 4, but only successively at three scanning positions smaller triangles. In addition, the scanning time per scanning position for the dye molecule 6 can deviate from that of the dye molecule 4 .

3 schematically shows the structure of a fluorescence microscope 23 according to the invention. A light source 24 provides excitation light 25, which is also the scanning light 26 in the embodiment shown. In order to form the intensity distribution of the excitation or scanning light in the sample 27, which has a local minimum, a spatial light modulator (SLM) 29 is arranged in the beam path 28 of the excitation light 25, with which the wavefront of the excitation light 25 is modified in such a way that during focusing of the excitation light 25 through the microscope lens 30 results in the intensity distribution having an intensity minimum. In the beam path 28 of the excitation light 25 there is also an electro-optical deflector (EOD) 31 for rapid positioning of the intensity distribution of the excitation light in two directions in the sample 27.

The excitation light 25 reflected by the spatial light modulator (SLM) 29 is coupled into a main beam path 33 of the fluorescence microscope 23 using a beam splitter 32 . The beam splitter 32 is advantageously designed as a narrow-band reflecting dielectric notch filter whose reflection range overlaps as little as possible with the emission spectrum of the fluorescent dye, so that only small portions of the fluorescent light 34 propagating in the main beam path 33 in the opposite direction to the excitation light 25 are reflected out of the main beam path 33. The excitation light 25 is directed into the rear aperture of the microscope objective 30 with a scanning lens 35, a scanner 36 shown here by way of example in a quad configuration for only one scanning direction, and a tube lens 37. The fluorescent light 34 received from the sample 27 by the microscope objective 30 propagates along the main beam path 33 in the opposite direction to the excitation light 25 , being transmitted by the beam splitter 32 . The fluorescent light 34 is separated by a filter 38 from reflected excitation light 25 and from scattered light and is focused with a lens 39 through a confocal pinhole 40 onto a detector 41 .

In the configuration shown, the scanner 43 is configured to scan individual dye molecules jointly by the electro-optic deflector (EOD) 31 and the scanner 36 . The scanner 36 is used for a comparatively slow pre-positioning of the focused excitation light 25, which is possible over a large image field, on a dye molecule in the sample 27 that is in the fluorescent state, while the EOD 31 is used for positioning the intensity minimum at the scanning positions arranged closely around a dye molecule. The EOD 31 allows positioning at high speed, but with a positioning range that is limited to a few microns. At the same time is the Scanner 36 is also part of the scanning image acquisition unit which, together with the detector 41, enables confocal image acquisition of the sample 27 in order to acquire raster images, on the basis of which the scanning rules for scanning individual dye molecules in the sample 27 can be defined.

Finally, the fluorescence microscope 23 shown has a control and image processing unit 42, which on the one hand controls the activation of the scanner 36 and the reading out of the detector 41 for a confocal image recording and on the other hand a scanning of individual molecules at a sequence of scanning positions by a corresponding coordinated control of the EOD 31 and the scanner 36 allowed. The control and image processing unit 42 also includes a computing unit for image analysis of the raster image and for calculating scanning rules from the raster image. The control and image processing unit is usually designed as a programmable computer, as a programmable integrated circuit or as a microcontroller with appropriate input and output interfaces.

4 shows a preferred representation of locations of dye molecules determined using the method according to the invention in the form of reconstructed, high-resolution images 50 and trajectories 51 in the context of a larger area of the sample. The method according to the invention is embedded here in a workflow of several image recording steps, starting with the recording of an overview image 44 of the sample, the overview image 44 being composed of several overlapping individual images 45, for example several transmitted light images 46 recorded with low magnification. The overview image 44 shows a large area of the sample with many cells 7, 47, and thus represents a context on a large scale, but with low spatial resolution closer examination is selected, from which a preview image 48, for example a confocal fluorescence image 49, is recorded in the next step. In this preview image 48, for example, the fluorescence staining of the selected cells 47 can be checked. The locations of individual dye molecules are now determined in the selected cells 47 in order to determine high-resolution images 50 of structures and trajectories 51 of individual dye molecules in the selected cells 47 . The locations of individual dye molecules are determined according to the invention by recording raster images 1 of sections of the sample, each of which contains a cell 47 or parts of a cell 47, and by scanning individual dye molecules according to scanning specifications that are determined from the raster images 1. The high-resolution images 50 or trajectories 51 generated from the location determinations of many dye molecules or from the repeated location determinations of a single dye molecule are displayed in the context of the raster image 1, the preview image 48 and the overview image 44. Reference List

1 raster image

2 pairs of values

3 data storage

4 dye molecule

5 dye molecule, unselected

6 dye molecule

7 cell

8 filaments

9 fluorescent dye

10 cell nucleus

11 fluorescent dye

12 fluorescent state

13 Fluorescence image

14 area

15 area

16 scanning position

17 Sampling Rule

18 hexagon

19 smaller hexagon

20 sampling rule

21 scanning position

22 triangle

23 fluorescence microscope

24 light source

25 excitation light

26 scanning light

27 sample

28 beam path

29 Spatial Light Modulator (SLM)

30 microscope lens

31 electro-optic deflector (EOD)

32 beamsplitter

33 main beam path

34 fluorescent light scan lens

scanner

tube lens

filter

lens

pinhole

detector

Control and image processing unit

scanning device

overview screen

single frame

transmitted light image

cell

thumbnail

Fluorescence image high-resolution image

trajectory

Claims

26

patent claims

1. A method for spatially highly precise localization of individual dye molecules (4, 5, 6) of one or more fluorescent dyes (9, 11) in a sample (27), wherein the fluorescent dyes (9, 11) between a first, fluorescent state (12), which can be excited to emit fluorescence in a fluorescence wavelength range, and a second state which, when the first state is excited, shows no fluorescence emission in the fluorescence wavelength range of the first state, can be converted in at least one direction, comprising the method steps:

- Generating a distribution of individual dye molecules (4, 5, 6) in the first, fluorescent state (12), the distance between adjacent fluorescent dye molecules (4, 5, 6) which, when excited with an excitation light (25), produce a fluorescence in emit the same fluorescence wavelength range, is above the optical diffraction limit;

- selecting a fluorescent dye molecule (4, 5, 6) from the distribution, the location of the selected dye molecule (4, 5, 6) being known approximately, preferably with a location uncertainty of less than 250 nm;

- Exciting the selected dye molecule (4, 5, 6) with the excitation light (25);

- Scanning the dye molecule (4, 5, 6) with an intensity distribution of a scanning light (26) having a local minimum at a plurality of scanning positions (16) arranged at a distance of no more than 250 nm from the location of the dye molecule (4, 5, 6), 21) according to a scanning specification, wherein the scanning light can be the excitation light (25) or a fluorescence prevention light;

- detecting a number of photons or an intensity of fluorescent light (34) at each scanning position (21);

- Determining the location of the dye molecule (4, 5, 6) from the number of photons or intensities of the fluorescent light (34) and the scanning positions (16, 21) with a location uncertainty reduced by at least a factor of 10, characterized in that before scanning of the dye molecule (4, 5, 6) a raster image (1) of the sample (27) or of a section of the sample (27) comprising the dye molecule (4, 5, 6) is recorded by means of scanning laser microscopy and that the scanning specification for scanning the Dye molecule (4, 5, 6) from a signal of a pixel of the raster image (1) at the location of the dye molecule (4, 5, 6) or from the signals of several pixels of the raster image (1) in an environment of the dye molecule (4, 5, 6) is determined.

2. The method according to claim 1, characterized in that the sampling rule (17, 20) defines one or more of the following parameters:

- Number and position of the scanning positions (16, 21) in the sample,

- the scanning positions (16, 21) assigned scanning times,

- Minimum photon numbers or minimum intensities of the fluorescent light (34) assigned to the scanning positions (16, 21),

- waiting times inserted between the scanning positions (16, 21),

- the intensities of the excitation and/or scanning light assigned to the scanning positions (16, 21),

- Wavelengths of the excitation and/or the scanning light.

3. The method according to claim 1 or 2, characterized in that determining the sampling rule (17, 20) comprises selecting a sampling rule (17, 20) from a set of previously defined sampling rules (17, 20).

4. The method according to claim 3, characterized in that the scanning specifications (17, 20) are adapted to the quantity of the properties of different fluorescent dyes (9, 11).

5. The method according to claim 3 or 4, characterized in that the sampling rule (17, 20) is selected by an artificial neural network trained for this task with the raster image (1) or an input vector calculated from the raster image (1) as input .

6. The method according to claim 1 or 2, characterized in that the determination of the sampling rule (17, 20) comprises calculating a function or executing an algorithm, the function or the algorithm receiving the signals of one or more pixels of the raster image (1) be passed as an argument.

7. The method according to any one of claims 1 to 6, characterized in that the scanning rule (17, 20) after scanning the dye molecule (4, 5, 6) at a part of the scanning positions (16, 21) is updated or newly determined and that the sampling is continued according to the updated or newly determined sampling rule (17, 20).

8. The method as claimed in claim 7, characterized in that the updating or redetermination of the scanning specification (17, 20) takes place using the number of photons detected at the already scanned scanning positions (16, 21) or the intensities of the fluorescent light (34).

9. The method according to any one of claims 1 to 8, characterized in that for the scanning of the dye molecules (4, 5, 6) and for the scanning recording of the raster image (1) common scanning means, the same excitation light (25), a common beam path and/or a common partial beam path can be used.

10. The method as claimed in one of claims 1 to 9, characterized in that the raster image (1) is a raster image (1) with a fluorescence intensity signal or a fluorescence lifetime signal that is detected confocally or recorded with multi-photon excitation, in particular two-photon excitation.

11. The method according to claim 10, characterized in that the second state is also a fluorescent state, which can be selectively excited with a wavelength different from the wavelength of the excitation light (25) and/or whose fluorescence emission occurs in a range other than the fluorescence wavelength range, and that the fluorescence emission of the second state forms the fluorescence intensity signal or the fluorescence lifetime signal of the halftone image (1).

12. The method according to claim 10 or 11, characterized in that the raster image (1) with the fluorescence intensity signal or the fluorescence lifetime signal has a plurality of channels which are in the excitation wavelength, a detection wavelength range and/or a polarization direction of the excitation light (25) or the detected fluorescence light (34) distinguish.

13. The method according to any one of claims 10 to 12, characterized in that the determination of the scanning specification (17, 20) includes a calculation of a parameter derived from the fluorescence intensity signal or from the fluorescence lifetime signal of the raster image (1), in particular a pH value, an ion concentration or a fluorescence anisotropy in individual pixels or in groups of pixels of the raster image (1).

14. The method according to any one of claims 1 to 9, characterized in that the raster image (1) has an image contrast from the following group: second harmonic generation contrast (SHG contrast), third harmonic generation correction (THG contrast), scattered light contrast, reflected light contrast, differential interference contrast (DIC contrast), polarization contrast.

15. The method according to any one of claims 1 to 14, characterized in that the determination of the sampling rule (17, 20) includes a calculation of signal ratios or of 29

Correlation amplitudes between corresponding pixels or areas of pixels in and in a further raster image (1) or between several channels of the raster image (1). Method according to one of Claims 1 to 15, characterized in that the determination of the sampling rule (17, 20) comprises a thresholding of the signal, a parameter derived from the signal, a signal ratio or a correlation amplitude. Method according to one of Claims 1 to 16, characterized in that the determination of the sampling rule (17, 20) comprises segmenting the raster image (1) on the basis of the signal, a parameter derived from the signal, an intensity ratio or a correlation amplitude. Method according to one of Claims 1 to 17, characterized in that dye molecules (5) are not selected if the signal of the raster image (1) at the location of the respective dye molecule (5) or in the vicinity of the respective dye molecule (5) falls below a minimum value or exceeds a maximum value. The method according to any one of claims 1 to 18, characterized in that generating the distribution of individual dye molecules (4, 5, 6) of the fluorescent dye (9, 11) in the first, fluorescent state (12) by illuminating the sample (27). Photo activation light or photo deactivation light takes place and that the intensity of the photo activation light or the photo deactivation light is adjusted on the basis of the raster image (1). Method according to one of Claims 1 to 19, characterized in that a high-resolution image of the sample (27) is reconstructed from the locations of a number of dye molecules (4, 5, 6). Method according to Claim 20, characterized in that the high-resolution image is displayed in the raster image (1). Method according to one of Claims 1 to 19, characterized in that a dye molecule (4, 5, 6) is repeatedly scanned and its location is repeatedly determined and that a trajectory of the dye molecule (4, 5, 6) is generated from the repeatedly determined locations . 30

23. The method according to claim 22, characterized in that the trajectory is represented in the raster image (1).

24. The method according to any one of claims 1 to 23, characterized in that the section of the sample (27) in which the raster image (1) is recorded is selected by marking an area in a larger preview image.

25. The method as claimed in claim 24, characterized in that the image area of the preview image is selected by marking an area in an overview image, the overview image comprising an even larger image area than the preview image.

26. Fluorescence microscope (23) with

- a light source (24) for excitation light (25),

- Beam shaping means for forming an intensity distribution of a scanning light (26) in a sample, which has a local intensity minimum, wherein the scanning light (26) can be the excitation light (25) or a fluorescence prevention light of a further light source,

- a scanning device for positioning the scanning light (26) in a sample (27),

- a detector (41) for detecting fluorescent light (34) from the sample (27),

- A scanning image recording unit, characterized in that the fluorescence microscope (23) is set up to carry out a method according to any one of claims 1 to 25.

27. Fluorescence microscope according to claim 26, characterized in that the scanning device and the scanning image recording unit jointly use a beam deflection unit.

28. Fluorescence microscope according to claim 26 or 27, characterized in that the scanning device comprises an acousto-optical or an electro-optical deflector (31).

29. Fluorescence microscope according to one of claims 26 to 28, characterized in that the scanning image acquisition unit is a confocal laser scanning or a STED image acquisition unit.

30. Fluorescence microscope according to one of claims 26 to 29, characterized in that the fluorescence microscope has means for merging a plurality of spatially partially overlapping raster images of the scanning image recording unit.