Improving accuracy of cytometers is challenging because optical configuration, flow control methods, and calibration issues make it difficult to characterize geometric factors associated with signal collection. State-of-the-art tools only collect a small solid angle of emitted light, so that minor variations in the aforementioned factors yield large uncertainties in fluorophore count, the primary measurand of interest. Moreover, such uncertainties are difficult to quantify because it has previously been impossible to repeat measurement at the same time and location. We have invented an amplitude modulation process that addresses both problems by multiplexing fluorescence signals in the frequency domain. This allows for a single, large photodetector to simultaneously detect many independent fluorescence signals originating at the same or different locations. The need for precise alignment or calibration is significantly reduced, thereby diminishing the corresponding uncertainties as well. Moreover, allowing larger detectors can increase the magnitude of any collected signal to nearly 100% of its theoretical maximum, further reducing the effects of noise. The AC and DC components of each signal can also be separated to yield multiple measurements of any given object. For comparison, state-of-the-art cytometers collect only a small solid angle (5% or less) of the available signal, with relative uncertainties of 25% or more, and no ability to repeat measurements. The invention builds on technology described NIST docket 18-017US1 (U.S. Patent Application serial number 15/967,966 “OPTICAL FLOW METER FOR DETERMINING A FLOW RATE OF A LIQUID,”), primarily as it describes method of exciting fluorophores in specific regions of microfluidic flow channel.
The invention, which we call multiplexed amplitude modulation fluorometry, is a method of signal generation, acquisition, and analysis that can simultaneously detect and distinguish fluorophores contained on or in many distinct samples separated in space and/or wavelength. The main idea of this approach is the encode information about the location of the measurements in the carrier frequency that generates the fluorescence, while using the amplitude to encode the actual measurement signal. Critically, this permits the use of one or more large optical detectors to collect signals from many separate laser interrogation regions at the same time and without interference. Thus, our technique can reduce or eliminate errors due to misalignment of photodetectors and unknown geometric factors. Moreover, the AC and DC components of any signal carry the same information, so that the method inherently performs repeat measurements at the same location and time. These advances enable reproducibility and reduce complexity of device design (both number of and geometric factors associated with detectors), which are critical to improving the quality and accuracy of biomedical research tools such as flow cytometers.
In more detail, the main operating principle of the technique is to multiplex fluorescence measurements by binding them to different carrier signals that can be superimposed to create a single signal still containing all the information. The carrier signals or “channels” are generated (for example) by oscillating the intensities of multiple light sources at different frequencies. Fluorescent dyes, beads, or cells exposed to any one such source will then be induced to fluoresce with an intensity that has the same frequency as the corresponding carrier signal and an amplitude proportional to the number of fluorophores illuminated. The resulting signal is thus a modulated version of the carrier signal transformed into the color of the fluorescent light, where the amplitude carries the information about the measurement. Critically, what distinguishes this invention from existing technology is the process of combining signals from multiple channels or interrogation regions that are separated in space-- e.g. by having them all impinge on the same detector or using waveguide combiners– to create a single superimposed signal. The latter can then be demodulated to separate the amplitudes associated with each channel, thereby recovering the individual measurements. Importantly, a multiplexing approach using a pair of large (e.g. disk-shaped) fluorescence detector sandwiching the interrogation region will collect almost 100% of the light from each sample irrespective of their exact locations, provided they are close enough to the detectors’ centers. This is especially useful, since it facilitates standardization and uniformity of measurements needed to achieve high accuracy fluorophore counts. A second distinguishing element of our technique is its unique ability to extract multiple measurements of the same object separately from the DC or AC component of the amplitude modulated signal. Critically, this first-of-its-kind ability to repeat measurements at the same time and location greatly facilitates uncertainty quantification.
High-throughput fluorescence measurements are at best an art-form. Cytometers, which are used in biomedical research, cancer detection, drug development, etc., are the only tool able to make such measurements, and their uncertainties on any sample (i.e. a bead, cell, etc.) for a single fluorescence signal are on the order of 25% to 50%. One of the fundamental engineering problems contributing to this uncertainty is variation in geometric factors, optics alignment, and fluorescence detector calibration. Moreover, given that cytometers typically measure up to 30 or more fluorescence signals on different detectors (corresponding to different cell properties), it is not unreasonable that these uncertainties are realized independently as many times for each cell or bead and may combine in a non-trivial way, depending on how the data is analyzed. As a result, the total effective uncertainty associated with characterizing a cell population may be significantly larger than that associated with a single measurement.
When using a single large detector to achieve multiplexing, the invention described herein eliminates geometric and alignment factors altogether. For example, a large circular detector with samples on top and near its center will, to very good approximation, collect half of the fluorescence emitted from each sample, irrespective of their exact locations. In this setup, the differences in fraction of collected light between samples can easily be controlled by the size of the detector and made negligible from a measurement perspective. Moreover, the use of a single detector for all measurements eliminates the need to calibrate many such detectors and thereby improves precision. From a commercial standpoint, this approach is also desirable, since it reduces complexity of the hardware and reduces costs by virtue of using fewer components. Similarly, sandwiching the device between 2 collectors would give approximately 100% fluorescence collection coverage, albeit with some increased cost and complexity.