Imaging of fast-moving single cells with adaptive single pixel detection
Lead Research Organisation:
Imperial College London
Department Name: Metabolism, Digestion and Reproduction
Abstract
Development of an untargeted technique capable of rapidly detecting cell populations at a single cell level is a sought after next-generation technology that faces various challenges.
It is thus our vision that a mass spectrometric system that facilitates unlabelled, multi-omic analysis can be developed with the following design criteria:
1. the biochemical coverage to match the power of low throughput methods (~500 proteins, 200 metabolites and 300 lipid species)
2. analysis of up to 10,000 mammalian cells (or other single-cell organisms)/second
We have thus proposed a new design addressing these challenges by using cell-containing liquid droplets produced by a flow focusing mechanism, and the in-droplet chemical digestion is followed by picosecond resonant infrared laser evaporation prior to mass spectrometric analysis of the produced gaseous organic ions.
A major research problem to be solved is the accurate evaporation of the incoming cells/organisms which is expected to travel at a speed of at least 1m/s and the content of the droplets needs to be verified to avoid 'false positives', hence spatial information in real-time is needed.
An ultrafast imaging set-up capable of capturing the spatial information of the moving cells at a rate that is at least on the order of 10^4 frames per second will thus be constructed to facilitate the succeeding mass spectrometric analysis and to provide additional biophysical information at a single cell level. Instead of costly array detectors that generally have limited availability and differing performance characteristics across the application spectrum, single pixel detection (SPD) offers a robust alternative solution for the problem at hand.
It is thus our vision that a mass spectrometric system that facilitates unlabelled, multi-omic analysis can be developed with the following design criteria:
1. the biochemical coverage to match the power of low throughput methods (~500 proteins, 200 metabolites and 300 lipid species)
2. analysis of up to 10,000 mammalian cells (or other single-cell organisms)/second
We have thus proposed a new design addressing these challenges by using cell-containing liquid droplets produced by a flow focusing mechanism, and the in-droplet chemical digestion is followed by picosecond resonant infrared laser evaporation prior to mass spectrometric analysis of the produced gaseous organic ions.
A major research problem to be solved is the accurate evaporation of the incoming cells/organisms which is expected to travel at a speed of at least 1m/s and the content of the droplets needs to be verified to avoid 'false positives', hence spatial information in real-time is needed.
An ultrafast imaging set-up capable of capturing the spatial information of the moving cells at a rate that is at least on the order of 10^4 frames per second will thus be constructed to facilitate the succeeding mass spectrometric analysis and to provide additional biophysical information at a single cell level. Instead of costly array detectors that generally have limited availability and differing performance characteristics across the application spectrum, single pixel detection (SPD) offers a robust alternative solution for the problem at hand.
Technical Summary
The development of our single platform multi-omic device targets two main design criteria: the biochemical coverage is expected to match the power of low throughput methods; and an equally significant challenge, the analysis speed, should be comparable to standard cytometry. While proven achievable with our pilot data, it is expected, however, that the spectrometer would be operating at or close to the limit of detection when receiving single cells and in order to attain good signal-to-noise ratio (SNR) and hence spectral quality, sufficient ionisation of every cell is vital. At the same time, it is anticipated that some droplets will not contain any cell/organism, and consist of only the liquid solvent utilised (e.g. isopropanol). While the mass spectral characteristics are different for these droplets, excessive ingestion of these empty droplets over time will provide a comparatively strong background affecting the sensitivity and requiring more samples to be used. As such, to achieve the desired throughput, it is necessary to accurately evaporate the incoming cells/organisms, which are travelling at a speed of >=1 m/s, and verify the content of the droplets to minimise the false positive rate, thus requiring spatial information in real-time.
To achieve this, the proposed system rapidly takes 'snapshots' of the fast-moving sample by encoding the spatial information with light masks and then reconstructing computationally after signal acquisition with a single photodetector. Calibration particles of various sizes, different cell lines and bacterial and fungal species (e.g. yeast) will be tested with the proposed system, which in turn provides the feedback to a machine learning approach which will adaptively optimise the masking strategy for different samples and operating conditions.
To achieve this, the proposed system rapidly takes 'snapshots' of the fast-moving sample by encoding the spatial information with light masks and then reconstructing computationally after signal acquisition with a single photodetector. Calibration particles of various sizes, different cell lines and bacterial and fungal species (e.g. yeast) will be tested with the proposed system, which in turn provides the feedback to a machine learning approach which will adaptively optimise the masking strategy for different samples and operating conditions.
