MERSCOPE mouse liver

In this notebook, we will use ovrlpy to investigate the Vizgen MERSCOPE’s mouse liver dataset.

We want to create a signal embedding of the transcriptome, and a vertical signal incoherence map to identify locations with a high risk of containing spatial doublets.

Settings and Imports

First, let’s define settings and input files.

from pathlib import Path

import matplotlib.pyplot as plt

import ovrlpy

sample_nr = 1
slice_nr = 1

# TODO: adjust path
data_path = Path("path/to/downloaded_data") / f"Liver{sample_nr}Slice{slice_nr}"

Loading the data

Next, we want to load the data.

coordinate_df = ovrlpy.io.read_MERSCOPE(data_path / "detected_transcripts.csv")

print(f"Number of transcripts: {len(coordinate_df):,}")

Number of transcripts: 417,243,171

coordinate_df.head()

shape: (5, 4)

x	y	z	gene
f64	f64	f64	cat
2506.407	-95.45148	0.0	"Comt"
2531.8447	-95.18702	0.0	"Comt"
2483.7969	-91.360115	0.0	"Comt"
2505.7693	-84.08165	0.0	"Comt"
2501.394	-81.38709	0.0	"Comt"

Tissue overview

# only show every 5,000th transcript
n = 5_000

fig, ax = plt.subplots()
ax.scatter(coordinate_df[::n, "x"], coordinate_df[::n, "y"], s=0.1)
_ = ax.set(aspect="equal")

Running the ovrlpy pipeline

ovrlpy provides a convenience function run to run the entire pipeline. The function creates a signal integrity map, a signal strength map and a Visualizer obejcet to visualize the results.

# ensure reproducibility by setting random_state
# min_distance???
liver = ovrlpy.Ovrlp(coordinate_df, n_components=10, n_workers=8, random_state=42)
liver.analyse()

Running vertical adjustment
Creating gene expression embeddings for visualization
determining pseudocells
found 94105 pseudocells
sampling expression:

100%|██████████| 72/72 [03:33<00:00,  2.96s/it]

Modeling 10 pseudo-celltype clusters;
Creating signal integrity map

100%|██████████| 420/420 [14:50<00:00,  2.12s/it]

_ = ovrlpy.plot_pseudocells(liver)

Signal integrity of the liver sample

fig = ovrlpy.plot_signal_integrity(liver, signal_threshold=3)

Doublet probability

doublets = liver.detect_doublets(min_signal=3, integrity_sigma=3)

fig, ax = plt.subplots()
_scatter = ax.scatter(
    doublets["x"], doublets["y"], c=doublets["integrity"], s=0.2, cmap="viridis"
)
_ = ax.set_aspect("equal")
_ = fig.colorbar(_scatter, ax=ax)

Visualize a specific doublet event like so.

doublet_case = 1

x, y = doublets["x", "y"].row(doublet_case)

_ = ovrlpy.plot_region_of_interest(liver, x, y, window_size=40)