Pegasus Plotting Tutorial

Author: Hui Ma, Yiming Yang, Rimte Rocher
Date: 2022-03-09
Notebook Source: plotting_tutorial.ipynb

import pegasus as pg

For this plotting tutorial, we provide an analysis result of gene-count matrix dataset on Human Bone Marrow with 8 donors. You can get the data from https://storage.googleapis.com/terra-featured-workspaces/Cumulus/MantonBM_result.zarr.zip, or use gsutil to download via its Google bucket URL (gs://terra-featured-workspace/Cumulus/MantonBM_result.zarr.zip):

After downloading, load the file using Pegasus read_input function:

data = pg.read_input("MantonBM_result.zarr.zip")
data

2022-03-09 00:07:00,912 - pegasusio.readwrite - INFO - zarr file 'MantonBM_result.zarr.zip' is loaded.
2022-03-09 00:07:00,913 - pegasusio.readwrite - INFO - Function 'read_input' finished in 0.85s.

MultimodalData object with 1 UnimodalData: 'GRCh38-rna'
    It currently binds to UnimodalData object GRCh38-rna

UnimodalData object with n_obs x n_vars = 35465 x 25653
    Genome: GRCh38; Modality: rna
    It contains 2 matrices: 'X', 'raw.X'
    It currently binds to matrix 'X' as X

    obs: 'n_genes', 'Channel', 'gender', 'n_counts', 'percent_mito', 'scale', 'louvain_labels'(cluster), 'anno'
    var: 'featureid', 'n_cells', 'percent_cells', 'robust', 'highly_variable_features', 'mean', 'var', 'hvf_loess', 'hvf_rank'
    obsm: 'X_diffmap', 'X_fle'(basis), 'X_pca', 'X_pca_harmony', 'X_phi', 'X_tsne'(basis), 'X_umap'(basis), 'diffmap_knn_distances'(knn), 'diffmap_knn_indices'(knn), 'pca_harmony_knn_distances'(knn), 'pca_harmony_knn_indices'(knn)
    varm: 'means', 'partial_sum', 'de_res'
    obsp: 'W_diffmap', 'W_pca_harmony'
    uns: 'genome', 'louvain_resolution', 'modality', 'norm_count', 'pca_features', 'stdzn_max_value', 'PCs', 'diffmap_evals', 'ncells', 'stdzn_mean', 'stdzn_std', '_attr2type', 'df_qcplot', 'pca'

In the following sections, we'll cover Pegasus plotting functions using this dataset. Moreover, for gene plots, the canonical gene markers below will be used:

• B cells and Plasma cells: CD38, JCHAIN, CD79A, MS4A1.
• T cells: TRAC, CD3D, CCR7.
• Cytotoxic T cells: CD8A, CD8B.
• NK cells: NKG7.
• Monocytes: CD14, FCGR3A.
• Erythroid cells: GYPA.
• Hematopoietic Stem cells: CD34, SELL.
• Dendritic cells: HLA-DPA1, CD4.

marker_genes = ['CD38', 'JCHAIN', 'FCGR3A', 'HLA-DPA1', 'CD14', 'CD79A', 'MS4A1', 'CD34', 'TRAC', 'CD3D', 'CD8A',
                'CD8B', 'GYPA', 'NKG7', 'CD4', 'SELL', 'CCR7']

Sections

• QC Violin Plot
• Highly Variable Feature Plot
• Composition Plot
• Scatter Plot
• Violin Plot
• Heat Map
• Dot Plot
• Dendrogram Plot
• Volcano Plot
• More Reference

QC Violin Plot

The first step in preprocessing is to perform the quality control analysis, and remove cells and genes of low quality.

pg.qcviolin shows the effect of quality control more intuitively by presenting the violin plot of cell distribution before and after filtration.

plot_type='gene' shows the number of expressed cells before and after filtration.

pg.qcviolin(data, plot_type='gene', dpi=100)

Quality control stats on number of percentage of mitochondrial genes:

pg.qcviolin(data, plot_type='mito', dpi=100)

The number of UMIs before and after filtration is also an important aspect of quality control.

pg.qcviolin(data, plot_type='count', dpi=100)

Highly Variable Feature Plot

Highly Variable Genes (HVG) are more likely to convey information discriminating different cell types and states. Thus, rather than considering all genes, people usually focus on selected HVGs for downstream analyses.

Pegasus provides hvfplot function to generate a scatterplot of genes upon HVG selection. This plot only works for Pegasus-flavor HVGs (i.e. flavor='pegasus' in Pegasus highly_variable_features function).

After selecting 2000 HVGs using the Pegasus selection method, the plot below is generated. Each point stands for one gene. Blue points are selected to be HVGs, which account for the majority of variation of the dataset. By default, it prints labels of 20 top HVGs. You can change this number in top_n parameter.

pg.hvfplot(data, dpi=200)

Composition Plot

Composition plot is a bar plot showing the cell compositions (under different conditions) in each cluster. Below is to show the composition of different samples in each Louvain cluster:

fig = pg.compo_plot(data, 'louvain_labels', 'Channel', style = 'frequency')

Composition plot is useful to fast assess library quality and batch effects.

Scatter Plot

Scatter plot requires at least 2 parameters

• data – Gene-count matrix to show.
• basis – Cell embedding to show. Can be either 'umap', 'tsne', ‘fle’, ‘net_umap’ or ‘net_fle’.

For this demonstration, we select annotation and channel as data attributes, and tsne as basis.

Cell Visualization

To visualize the data, users need to first run visualization functions to calcualte the corresponding cell embeddings. See here for all visualization functions in Pegasus.

tSNE Plot. Pegasus uses FIt-SNE algorithm for tSNE plot, which works better on large-scale datasets:

pg.scatter(data, attrs=['anno','Channel'], basis='tsne')

You can show both cluster-specific cell types (anno) and samples (Channel) in one function.

Note here, the legend of left plot is not fully dislplayed. We need to adjust the horizon distance between two plots by changing "wspace"

"wspace" is the ratio of horizon distance to width of plots. For example ,the default value of wspace is 0.4, which means the gap is 0.4 times the width of plot.

pg.scatter(data, attrs=['anno','Channel'], basis='tsne', wspace=1.2)