Cytometry Part A最新文献

Staining panels have become increasingly complex in recent years, with 40 to 50 dye panels regularly reported, particularly on “spectral” flow cytometers. It is universal practice to include all dyes in the mixing matrix when unmixing the data, even though it is well-known that individual events within the sample only stain with a subset of dyes. Adding dyes to the mixing matrix increases the variance of the unmixed abundance distributions, even if those dyes are not present on particular events. This manuscript introduces a novel unmixing method called TRU-OLS. TRU-OLS utilizes a priori biological knowledge (i.e., unstained controls) to unmix each event with only the dyes present on that event. We show that TRU-OLS decreases the variance of the unmixed abundances both statistically and visually in simple (4–6 color) and complex (40 color) staining panels.

This manuscript is the first in a series that develops and realizes core ideas from metrology and uncertainty quantification (UQ) as applied to flow cytometry. The work herein is motivated by the problem of estimating the detection efficiency (Q) and background (B) of cytometers. Despite more than 30 years of study, canonical solutions to this problem make approximations that both ignore and amplify various sources of noise, thereby leading to unstable estimators of $�\; ��\; �Q��$ and negative values of $�\; ��\; �B��$. Moreover, it is not always clear how to compare instruments on the basis of such properties. To address these issues, we propose a global data analysis strategy that combines measurements taken with different gains while simultaneously accounting for gain-independent background effects, which are typically ignored but often dominant. Of note, this technique yields stable estimates of $�\; ��\; �Q��$ and $�\; ��\; �B��$ while also quantifying the relative impacts of other noise sources. Conceptually, our analysis also unifies and explains the shortcomings of existing data analysis methods. Most importantly, however, this work allows us to rigorously define concepts such as limits of detection and quantification associated with instrument performance alone and in a way that removes effects associated with sample preparation, operator effects, and so forth. Importantly, this allows for direct comparison of cytometers on the basis of sample-independent uncertainty metrics and yields information for optimizing cytometer performance in terms of instrument-induced uncertainties. Results are experimentally verified using both commercial instruments and a NIST-developed serial cytometer, with extensions considered in companion manuscripts of this series.

Cytometry enables simultaneous assessment of individual cellular characteristics, offering vital insights for diagnosis, prognosis, and monitoring various human diseases. Despite its significance, the process of manual cell clustering, or gating, remains labor-intensive, tedious, and highly subjective, which restricts its broader application in both research and clinical settings. Although automated clustering solutions have been developed, manual gating continues to be the clinical gold standard, possibly due to the suboptimal performance of automated solutions. We hypothesize that their performance can be improved via an appropriate representation of data from the clustering point of view. To this end, this work presents a novel unsupervised deep learning (DL) architecture wherein an efficient cytometry data representation is learned that helps discover cluster assignments. Specifically, we propose MuSARCyto, a multi-head self-attention-based representation learning network (RN) for the unsupervised clustering of cytometry data, utilizing a fully-connected representation network backbone. To benchmark MuSARCyto against the state-of-the-art cytometry clustering methods, we propose a cluster evaluation metric adjudicator score ($�\; ��\; �{\mathrm{Ad}}_n��$), which is an ensemble of prevalent cluster evaluation metrics. Extensive experimentation demonstrates the superior performance of MuSARCyto against the existing state-of-the-art cytometry clustering methods across six publicly available mass and flow cytometry datasets. The proposed DL achitectures are small and easily deployable for clinical settings. This work further suggests using DL methods for identifying meaningful clusters, particularly in the context of critical immunology applications.

Flow cytometers are powerful tools for bioanalytical applications, yet new systems that promise better measurements are continuously being introduced as sensors and other technologies advance. One such advancement by NIST was the recently demonstrated a serial microcytometer that enables unique capabilities for uncertainty quantification on a per-object basis. In an effort to benchmark and improve the measurement capabilities of the serial microcytometer, we found limitations to the quantitative comparison of instruments using conventional metrics and methods. To address these shortcomings, we recently developed an improved model that builds upon conventional models to improve comparability (Patrone et al. “Uncertainty Quantification of Fluorescence Signals in Flow Cytometry Part I: An Analytical Perspective Beyond Q and B” submitted in conjunction with this manuscript). In Part I, and continued here, our aim was to develop metrics that enable comparisons based on upper limit of linearity, limit of background, limit of detection, noise-to-signal ratio, and uncertainty decomposition thereof. We found that the NIST serial microcytometer has similar performance capabilities to a conventional analytical flow cytometer. This manuscript continues the development of uncertainty quantification (UQ) for flow cytometry by demonstrating how a serial microcytometer facilitates separation of the instrument-and population-dependent contributions to UQ. Component-level contributions to UQ can also be analyzed. Ultimately, these methods establish robust metrics for instrument performance and introduce per-object uncertainty as a mechanism facilitating better classification and utilization of cytometry data in research and clinical use.