Skip to content


Plotting Spatial Data on the G4X

One of the first things you'll want to do with your spatial data is to visualize your data on a spatial plot. G4X outputs are designed to make this information easy to access and plot in a variety of ways. With the tables and data structures provided, you can create a variety of spatial visualizations using tools like Seurat in R or scanpy in Python.


Python Visualization

This brief tutorial uses the scanpy library and assume that you have a working installation of it on your local system. If you do not, please see the scanpy installation guide for instructions on how to install it.

1. Import libraries

# Import necessary libraries
import scanpy as sc
import matplotlib.pyplot as plt

2. Load your data

# Load your data
adata = sc.read_h5ad('path/to/your/data.h5ad')

3. Extract spatial coordinates into a scanpy obsm layer for scanpy compatibility

# Extract spatial coordinates into obsm layer
adata.obsm['X_spatial_cell'] = adata.obs[['cell_x', 'cell_y']].to_numpy()

Note

In order for scanpy to extract this information for default plotting, you have to put (x, y) coordinate pairs for each cell into the obsm layer and name it something starting with X_. You can do this with any pair of coordinates.

4. Plot the spatial coordinates

# Plot the spatial coordinates
plt.figure(figsize=(10,8))
ax = sc.pl.scatter(adata, basis='spatial_cell', color='log1p_total_counts', show=False)

# Flip the y-axis to match image orientation
ax.invert_yaxis()
plt.show()

Note

Here, we're plotting with the color='log1p_total_counts' to visualize the log total transcript counts per cell, but you can replace this with any other variable in your adata.obs or adata.var that is the correct dimensions (1xncells) to visualize different aspects of your data.

5. Save the plot as a PNG file

# Save the plot as PNG
plt.savefig("spatial_plot.png", dpi=300, bbox_inches='tight')


R Visualization

This brief tutorial uses the ggplot2 library and assumes that you have a working installation of it on your local system. If you do not, please see the ggplot2 installation guide for instructions on how to install it.

1. Load necessary libraries

# Load necessary libraries
library(ggplot2)

2. Load your data and create a Seurat object

# Load your data and create a Seurat object
run_base = c('/path/to/g4x_output/<sample_id>')
metadata_path = file.path(run_base, "single_cell_data/cell_metadata.csv.gz")
meta <- read.csv(gzfile(metadata_path), row.names = 1)

3. Plot spatial coordinates with ggplot2

# Ensure coordinates are numeric
meta$cell_x <- as.numeric(meta$cell_x)
meta$cell_y <- as.numeric(meta$cell_y)

p <- ggplot(meta, aes(x = cell_x, y = cell_y, color=log1p_total_counts)) +
  geom_point(size = 0.3, alpha = 0.6) +
  scale_y_reverse() +           # <-- invert Y
  coord_equal() +
  labs(x = "cell_x", y = "cell_y", color = "log1p(total_counts)") +
  theme_minimal()

print(p)

Note

Here, we're plotting with the color='log1p_total_counts' to visualize the log total transcript counts per cell, but you can replace this with any other variable in your meta dataframe that is the correct dimensions (1xncells) to visualize different aspects of your data.

4. Save the plot as a PNG file

# Save as Plot as PNG
ggplot2::ggsave(
  filename = "cell_xy_log1p_total_counts.png",
  plot     = p,
  width    = 8, height = 8, dpi = 300, units = "in",
  bg       = "white"   
)


βΈ»