Use Built-In Visualization Tools

We have a variety of built-in visualization tools to get further insight into the results of various pipelines. All these visualizations are accessible via advanced search. Currently the following is available.

scRNA-seq

For scRNA-seq, we allow the labeled scRNA-seq data of a single sample to be loaded in a heatmap. 

At present this heatmap is restricted to the “observation item” Seurat PBMC Type with a value set to “CD8 effector” and only a subset of gene markers and unique cells can be loaded. The sample is also limited by “feature name” using a set of features including ['ACTB', 'HLA-A', 'MALAT1', 'OR4F5', 'OR4F29', 'CD3E', 'PTPRC', 'CD8A', 'CD4', 'CD79A', 'CD14', 'NCAM1', 'CD1C', 'FCGR3A', 'FLT3', 'GZMA', 'GZMB', 'ILR2A', 'RORC', 'TCL1A'].

Flow Cytometry

For flow cytometry there are interactive visualizations for displaying cell population stats and MFI results produced by supervised gating. In a per sample and per panel visualization, cell subsets can be selected, dropping the non-selected cells from view.

In addition, it is possible to select results of the same type (including panel) of multiple samples, and do a side-by-side comparison of the results. There are multiple ways in which the data can be segmented, such as subject demographics, sample metadata and CBC data, depending on the nature of the samples that were chosen. 

Finally, when CBC data is available, HISE offers a tight integration of CBC data into the data visualization.

Integration of CBC data

When CBC data accompanies a sample's flow data, absolute counts in whole blood can be used to derive an absolute number of cells per volume of whole blood for any supervised reportable.

To do this, ‘Derived Absolute Counts’ are calculated by relating ‘Total leukocytes’ in cytometry to its equivalent from CBC, accounting for the absence of most neutrophils and eosinophils in PBMC flow cytometry data.

Caveats:
Some fraction of basophils and a small percentage of neutrophils are captured by ficoll PBMC isolation, though variably sample-to-sample. Both populations will introduce errors in this calculation.

Calculation:
Flow Cytometry (Population Count / Leukocytes Counts) X CBC(White blood cell WBC (#/ul) - Neutrophil ANC (#/ul) - Eosinophil AEC (#/ul))

Using MFI for analysis across panels and batches
Median fluorescence intensity (MFI) is influenced by the fluor, antibody lot variability, staining variability, instrument variability, unmixing or compensation, compensation adjustments, and gating.These sources of variability are considered to be constant for any given batch of samples but prohibit direct batch-to-batch MFI comparisons. Therefore, to analyze MFI across more than one batch a normalization must be applied.

As a technical replicate sample (the bridging control leukopak) is included with every batch, this bridging control can be used for cross-batch normalization of MFI to permit cross-batch analysis. For this purpose, we consider the bridging control as biologically invariant and relate the sample MFI to the bridging control MFI, per batch, per panel, per analyte.

Caveats:
MFI “normalization” requires a significant number of events in the bridging control target gate; some populations (eg, plasmablasts) may not be amenable to this process

Calculation:
     Sample MFI (panelPxx/populationpop/targettar)
--------------------------------------------------------------------
Bridging control MFI (panelPxx/populationpop/targettar)

Statistics

There are three statistical functions that are automatically calculated if the prerequisites are available.

  1. Percent CD45
  2. MFI Bridging Control
  3. CBC DAC (Derived Absolute Count)

Human Metadata 

In Advanced Search for samples and for subjects, human metadata data (EMR, CBC, and survey results) can now be visualized, including basic controls for grouping data by demographics.

Visualizations can be saved to the Collaboration Space and CSV data can be downloaded for further analysis.

How to get started:

  1. Start a new query, and for the query output, select:
    1. "Subject metadata and EMR" for subject demographics and patient history (EMR Data) (Note: this option is only available if your project contains EMR data.)
    2. "Sample metadata, survey and lab data" for subject demographics, questionnaire data, and CBC results.
  2. Filter down your search results by various subject metadata and EMR filters.
  3. Select one or more results from the results set.
  4. When you click the ‘Visualize’ button, you will be presented with a menu that will allow you to select the human metadata of interest.
  5. Clicking the Visualize button once more will generate the visualizations of that data.
    1. For each data set, the most appropriate visualizations are pre-selected based on the nature of the data (data type and scale of measurement).
    2. You will be presented with overall values and shapes of the data, and have the possibility to break down the data along various axes.

Note: EMR data (patient history) is not tied to research visits and cannot therefore not be directly tied to a sample. However, each EMR data set contains both the clinic visit date and the relative timing between the clinic visit data and one of the research visits, measured in number of days. Using this parameter, the separate data sets can be overlaid in time by an analyst using the IDE, if needed.

How do I… Visualizations?

How do I zoom in on the displayed plot?

  1. Hover over the plot
  2. Click and Drag to zoom in 
  3. Double click with in the plot to zoom out

What other types of plots are available?
The available plot types for Flow Cytometry data are:
Bar
Box
Histogram2d
Histogram2dcontour
Scatter (Line and Marker)
Violin


The available plot type for scRNAseq H5 data is:
Heatmap


How do I download my plot as a png?
From the Plotly Toolbar

  1. Hover over the Plotly Toolbar
  2. Click the camera icon
  3. The plot will be downloaded as a png to your default download folder on your local machine.