Skip Ribbon Commands
Skip to main content

 Technologies & Bioinformatics Projects

Bioinformatics

By Martin Wasser (Team Lead), Pavanish Kumar, Yeo Joo Guan
Mass cytometry (CyTOF) measures the expression of multiple proteins in single cells (currently up to 40) and is used to characterise cellular diversity of the immune system in peripheral blood and other tissues. TII’s immune atlas is a growing database of CyTOF data acquired from blood of clinically and demographically diverse human subjects. Data of samples labeled with identical antibody panels are grouped and jointly clustered to identify distinct populations of immune cells. After determining their expression patterns, we assign cell types to these clusters. To analyse these complex multi-dimensional data, we developed a web application using the Shiny R programming environment that has two main objectives. First, users can explore the immune landscape at different levels of detail using a variety of interactive visualisation methods, such as bar charts, tSNE or UMAP scatter plots, heat maps or histograms. For instance, the abundance of immune cell types can be compared between different age groups or between patients suffering from immune-diseases and healthy subjects. Second, users can upload their own cytometry data and compare them with the immune atlas. For instance, machine learning is applied to classify expression patterns in uploaded data and provide estimates about the abundance of selected immune cell populations.
 
The application was implemented in Shiny R and runs in all web browsers. The bar chart below compares the abundance of immune cell population between atlas (blue) and uploaded data. The heat maps at the bottom compare the expression patterns of 37 genes in 100 clusters between atlas (left) and uploaded data (right).
 
Web-based data analytics (high res).png 

Healthy Immune Atlas

​By Yeo Joo Guan (Team Lead), Martin Wasser, Pavanish Kumar, Poh Su Li, Yeo Kee Thai

There is a critical unmet need for standardised datasets depicting at single cell level and with high dimensionality in the entire healthy human immune system. The advent of high-dimensional mass cytometry, in contrast to conventional oligo-dimensional approach with fluorescence-based flow cytometry, has provided the tool to create a holistic depiction of the immune system. However, the power available from these big data sets has not been yet fully harnessed due to the absence of clinically relevant and standardised datasets that can be effectively interrogated. Other hindrances are data fragmentation as well as the focus on individual cell subsets or ages which limits the ability to interrogate holistically with the Immunome and the entire developmental gradient from neonatal to adult age. Collectively, these limitations hamper mechanistic, translational and clinical research.

To address these issues, EPIC (https://epicimmuneatlas.org), a free web-based immune atlas platform, has been created. EPIC contains a high dimensional mass cytometry-based healthy human Immunome database spanning from cord blood to adult ages with the characterisation of 63 non-redundant phenotypical and functional immune markers at the single cell level. Concurrently, to enable this dataset to be used effectively as a reference atlas, we have built an analysis and visualisation pipeline to allow its dynamic interrogation in many dimensions and made this entire database and framework available to the research community.

EPIC Immune Landscape_developing healthy immune system.png

EPIC Immune Landscape: A depiction of the developing healthy immune system