Quantum-Enabled Flow Cytometer

patent description

We invented a flow cytometer that can self-calibrate in an absolute way that can reach the sensitivity for quantification and enumeration of a single and small number of biological markers per a biological entity. The cytometer uses a photon-number resolving detector (or a detector arrangement capable of resolving photon numbers, such as Hanbury-Brown Twiss arrangement), accumulating times of arrival and photon number statistics and using first principles of quantum optics that provide a number of participating emitters. There are two aspects to this invention. One describes a hardware modification, and the other describes measurement methods. Specifically:

1. A photon-number resolving detector or a Hanbury Brown and Twiss-type arrangement of two or more single-photon detectors should be used to detect fluorescent signal. This is a change in hardware, and it is NOT done to merely improve sensitivity. Rather, it is done to access photon number information and detection timestamps. This hardware modification provides access to new information, not considered in prior art. Without this modification our measurement methods cannot be applied.
2. We teach a measurement method that consists of two modes of operation.--The absolute calibration. With our setup, at a given concentration of fluorophores (per cell), we accumulate the optical signal from many exposures (many cells) and statistically find second-order correlation functions and their dependence on various time delays. We also measure average photon flux as a function of time and time delays. This process is repeated for several concentrations. The second-order correlation functions relate to concentrations via laws of quantum optics, this dependence does not require the prior knowledge of transmittance losses and detection inefficiencies, even with an unknown optical background present. This establishes an optical flux to moles calibration from first principles. The dependence of second-order correlation functions on time gives the most advantageous signal processing algorithm for a "sorting" mode of operation. This signal processing algorithm improves signal to noise.

The implications of this method are twofold. This invention solves the problem of calibration of flow cytometers for the purposes of intercomparing and quantification. Note that intercomparison with flow cytometers has not yet been achieved. Second, this invention teaches how to use the calibration information to optimize a sensitivity of a real-time cell-to-cell measurement to a level of a single fluorophore.

Light produced by a system of few optically active biomarkers (fluorescent molecules, quantum dots) exhibits quantum behavior. Therefore, quantum measurement methods can be applied. A measurement of photon number statistics can unambiguously resolve the number of emitters in the system and it is resilient to loss. This unambiguous measurement therefore enables self-calibration of a flow cytometer. We start with a sample of an unknown concentration of fluorophores, and dilute it by known factors to reach a range of concentrations with an average of fluorophores ranging from approximately .1 (or lower) to approximately 10 per interrogation volume. Then, we acquire photon number statistics for each of diluted samples. Finally, we combine the measurements and fit them to theory. The initial unknown concentration is then extracted from the fit. Although a cw-laser can be used to optically excite the fluorophores, we believe that using a pulsed laser will significantly improve sensitivity. In addition to an absolute calibration, because the complete list of all photoelectronic detections is retained, the time-resolving statistical analysis identifies temporal properties of signal (nonclassical) and noise (classical).

During a normal operation, a flow cytometer collects signal from each biological entity for only a short time (lO's of microseconds to a millisecond). Therefore, the number of collected photons from trace concentrations of biomarkers is limited. Therefore, shot noise is norma11y the limit for separating signal from background.  Because accurate  time of single photon detections are available, and because temporal properties of signal and noise are independently identified, the separation of signal from background can be made sub-shot noise.