visualization.Rmd

<br class="long"/>

# Visualization

<br>

To check the quality of your data at each step of the analysis, different levels of visualization is provided.

* Basic Plots

    + Cell count barplots for each sample
    + Density plots for each marker
    + PCA and MDS plots for each sample

<br>

* t-SNE 

    + with marker expressions
    + with marker expressions but 0 expression removed

<br>

* Clustering

    + FLOWSOM clustering on t-SNE plot
    + ClusterX clustering on t-SNE plot

<br>


### Basic Plots

A set of Basic plots is generated to let users understand their data. These include cell count plots, density plots for each marker and PCA plots. Cell count plots can be used to check if there are reasonable number of events in each sample. Density plots for markers give an idea about the distribution of each marker. Dimension reduction methods like PCA and MDS can be useful if different conditions are represented by samples and one expects to see difference for the specific markers in these condition group.

<br>

### t-SNE visualization

t-Distributed Stochastic Neighbor Embedding (t-SNE) is a well established method for dimensionality reduction and is well suited for the visualization of high-dimensional data like CyTOF. In this pipeline, t-SNE is implemented using Rtsne.multicore R package which allows multi-threading for faster computations. Two dimmensional t-SNE plots, colored by marker intensities are also plotted to highlight differnces amongst samples for markers.

<br>

### Clustering

Once clustering is performed by FLOWSOM or ClusterX, these clusters are projected on t-SNE plots for each sample.