April 11, 2018  |  Announcements, Conferences, Flow Cytometry  |  By  |  0 Comments

Announcing Cytobank’s Collaboration with Cytek Biosciences

We’rCytekLogoe excited to announce a new collaboration with Cytek, makers of the innovative CytekTM Aurora Flow Cytometer. Cytek’s Aurora and Cytobank’s next-generation analytics platform have united in their goal to make high-dimensional data and its subsequent machine learning analysis more accessible to more scientists.

  • Aurora’s unique optical design combined with spectral unmixing delivers quality, high-dimensional data where rare and dim populations can be easily resolved, regardless of sample complexity.
  • Cytobank’s algorithms simplify visualization and analysis of high-dimensional unmixed data and allow you to quickly share discrete insights from these data.

Here are two specific examples of how you can apply our combined solutions to enable faster, more highly-resolved discoveries:

1) Machine Learning tools in Cytobank accelerate discovery in high-dimensional Cytek data

With up to 48 fluorescent channels to detect the full spectrum of each fluorochrome across 3 lasers, Cytek’s Aurora brings the scientific community access to technology poised to detect high-dimensional data (up to 23 colors) more readily than before. For example, scientists can use Aurora’s full spectrum technology to run high-dimensional panels and to survey the immune system in their studies at the single cell level. However, the simultaneous measurement of multiple dimensions per single cell introduces new challenges for subsequent data analysis, such as creating intuitive summary figures to communicate results or to quickly perform exploratory data analyses.

Cytobank’s solution facilitates discoveries in these high-dimensional data by equipping scientists with a secure cloud-based analysis platform with easy to use machine learning tools to reduce dimensionality and make sense of the additional data complexity. Here, we partnered with Cytek to analyze healthy donor peripheral blood that was stained with 23 different markers using machine learning tools and we show you how these tools revealed biological differences across donors (Figs. 1 & 2)

Figure 1
Figure 1. Visualize single cell data on one plot with viSNE.
Intact single cells from three different healthy donor peripheral blood samples are shown in one plot on the tSNE1 and tSNE2 axes. All 23 measured markers were used for the viSNE analysis. Each dot represents a single cell.
A) Coloring the plots by a few of the measured markers – CD25, CD56, CD4, CD14, and CD19 – shows the phenotype across viSNE “islands.” Red represents high expression and blue represents low expression for each marker.
B) Immune cell populations were gated using SPADE, a clustering algorithm. These populations are overlaid onto the viSNE map highlighting regions in the map that correspond to unique immune cell subsets – APCs, T cells, NK cells, B cells.
Screen Shot 2018-04-08 at 10.21.28 AM
Figure 2. viSNE plots reveal biological differences across donors.
viSNE plots colored by indicated measured markers and cell density reveal a missing “island” in donor one that is present in donors 2 and 3 (pink arrows). Coloring viSNE plots by CD57, CD56, CD16, and CD3 shows that this missing island in donor 1 is a CD56+CD16+ NK cell population. Red represents high expression and blue represents low expression for each marker.

 

2) Cytek technology combined with Cytobank Analysis tools provide exceptional resolution into your data

Aurora’s optical design combined with spectral unmixing technology enables detection of rare populations and dim markers, which Cytobank tools deftly capture. Here, we highlight the CD4+ T cell compartment using both traditional and machine learning tools in Cytobank (Fig. 3). The Cytobank platform supports the complete analysis process post data acquisition from data QC to exploratory data analysis. Investigate the staining of your rare subsets, such as Tregs, during data QC with biaxial plots (Fig. 3A) and use viSNE as an exploratory data analysis tool to simultaneously look at multiple measured markers on your subset of interest (Fig. 3B).

Figure 3. Treg subset Resolved using Traditional and Machine-learning Tools. Healthy Donor 2 CD4+ T cell compartment is investigated using biaxial plots and viSNE. A) Biaxial plots used to gate for Treg and other CD4+ T cell subsets. B) viSNE plots show all intact single cells. Bottom-right plot shows Treg subset identified from biaxial plots in A overlaid onto the viSNE map in orange. All other plots show the expression of the indicated marker (eg. CD25, CD127). Red represents high expression and blue represents low expression for each marker.
Figure 3. Treg subset Resolved using Traditional and Machine Learning Tools. Healthy Donor 2 CD4+ T cell compartment is investigated using biaxial plots and viSNE. All 23 measured markers were used to generate the viSNE plots. A) Biaxial plots used to gate for Treg and other CD4+ T cell subsets. B) viSNE plots show all intact single cells. Bottom-right plot shows Treg subset identified from biaxial plots in A overlaid onto the viSNE map in orange. All other plots show the expression of the indicated marker (eg. CD25, CD127). Red represents high expression and blue represents low expression for each marker

See us in action at the AACR Meeting:

Get started setting up your own data analysis pipeline: