Standardization and quality control for high‐dimensional mass cytometry studies of human samples

Abstract

Mass cytometry (CyTOF), a mass spectrometry-based single cell phenotyping technology, allows utilization of over 35 antibodies in a single sample and is a promising tool for translational human immunology studies. Although several analysis tools are available to interpret the complex data sets generated, a robust method for standardization and quality control within and across studies is needed. Here we report an efficient and easily adaptable method to monitor quality of individual samples in human immunology studies and to facilitate reproducible data analysis. Samples to be assessed are spiked with a defined amount of reference peripheral blood mononuclear cells from a healthy donor, derived from a single large blood draw. The presence of known standardized numbers and phenotypic profiles of these reference cells greatly facilitates sample analysis by allowing for: 1) quality control for consistent staining of each antibody in the panel, 2) identification of potential batch effects, and 3) implementation of a robust gating strategy. We demonstrate the utility of this method using peripheral blood and bronchoalveolar lavage samples from HIV+ patients by characterizing their CD8+ T-cell phenotypes and cytokine expression, respectively. Our results indicate that this method allows quality control of experimental conditions and results in highly reproducible population frequencies through a robust gating strategy.

Automation of sample preparation for mass cytometry barcoding in support of clinical research: protocol optimization

“…Sample preparation is an important and critical step in the entire workflow [14, 18] and represents the largest bottleneck in the modern research lab. Most sample preparation techniques for CyTOF are still non-automated and require substantial labor-intensive and time-consuming operations, which leads to unwanted levels of variation in bioanalytical measurements.…”

“…If particular attention is not paid to all of sources of error such as sample processing, contamination, sample stability, instrument stability, storage conditions, and correct data processing, the sample integrity may be sacrificed and the analysis data affected, compromised, or rendered invalid [14, 18]. Sample preparation and the number of cells per sample are critical factors in determining cell recovery from shipped samples.…”

CytoNorm: A Normalization Algorithm for Cytometry Data

“…For each control sample, we used the original quantiles as the x-values and the corresponding quantiles from the goal distribution as the y-values to define the interpolation points. To avoid spurious extrapolation, we also added (0,0) and (8,8) to the interpolation points. The resulting spline function was then used to translate all original marker values to the new values, causing the data to become close to the goal distribution (Step 4).…”

The anatomy of single cell mass cytometry data

“…Like any assay type, batch effects arising from such things as different lots of reagent, runs on different days , or runs on different CyTOF instruments do exist and must be considered when making comparative analyses. These effects can be minimized with appropriate experimental design, such as having enough of a reagent lot for an entire experiment, running all samples for a project in a shorter amount of time, appropriate distribution of sample types such as cases and controls across each plate, including all timepoints for a given donor on the same plate, and the inclusion of a well‐characterized healthy control with each batch to help identify problematic plates . Batch effects are a known problem for most omics data types, and has in many cases been addressed by statistically modeling the noise and removing it.…”

“…Another issue that remains largely unaddressed is batch effects when comparing data from runs performed at different centers , on different instruments or versions, by different operators, or even simply on different dates. In some cases, a well‐characterized control sample is stained on the same plate, but primarily as a qualitative quality‐control measure . This lack of a robust model is somewhat related to the lack of a comprehensive error model, but the need is likely to increase as more and more public data becomes available.…”