SPRINT-seq

SPRINTseq (Spatially Resolved and signal-diluted Next-generation Targeted sequencing) is an innovative in situ sequencing strategy that combines hybrid block coding and molecular dilution strategies, allowing an accurate RNA detection in crowded environment. This method can profile a mouse brain coronal slice within 1 day, and generate gene expression architecture in a sub-micron resolution.

For more information, please read the article. Chang et.al (2023) bioRxiv

Code Preview

Code for SPRINT-seq consists of five parts, barcode_design, image_processing, gene_calling cell_segmentation and expression_matrix_analysis. Data will be processed in this order.

Data Architecture

Raw data base directory and processed data output directory can be whatever place you need. But its subdirectory should be like this:

Raw data root

|---RUN_ID1

|---RUN_ID2

|...

|---RUN_IDN

Output root

|---RUN_ID1_processed (automate created)

| |---focal_stacked (automate created)

| |---background_corrected (automate created)

| |---registered (automate created)

| |---stitched (automate created)

| |---readout (automate created)

| |---segmented (automate created)

Your raw data should be in folder RUN_ID.

Start

This code should be run under python 3.8. Later version will bring some environment problems.

To run this code, packages must be installed with command:

pip install -r requirements.txt

And MATLAB engine should be installed, according to your local MATLAB version, you can follow official guideline.

For example, MATLAB R2021b：python -m pip install matlabengine==9.11.21

And pip setuptools needs a correct version.

For MATLAB R2021b: pip install --upgrade pip setuptools==57.5.0

Remind! There are lots of paths or directories need to be edited in files mentioned below.

Barcode Design

Code in this part will generate a suitable barcode space for you to choose.

Caution! These codes may cost a lot of memory.

Step 1: Run generate_graph.py. An in-between file will be generated.

Step 2: Run MIS_variance_trail_qsub.py using the in-between file as input. Suitable barcodes will be generated in a txt file.

We have provided example in-between file test_graph_210809_G58.txt and example final file min_var_mis_0809_G58.txt.

Image Processing

Step 1: Edit the directory in python file scan_fstack.py as the directory you wish. Run the code:

python scan_fstack.py Raw_data_root

We have provided a preprocessed example data for Step 2 and pipeline after, which path is ./example_dataset. You should create a folder output_root/whatever_name_processed/focal_stacked/ and put all example data in it.

Step 2: Edit the directory in python file image_process_after_stack.py the same as the directory before. Run the code:

python image_proess_after_stack.py

Results will be large stitched images, which will be used in next part.

Gene Calling

Step 1: Before running code, a padlock sequence file should be provided. It should be a csv file and formatted like:

Name,Padlock
Gene_name1,Padlock_sequence1
Gene_name2,Padlock_sequence2
...,...
Gene_nameN,Padlock_sequenceN

Step 2: Run the jupyter notebook file generate_ref_files.ipynb. A reference barcode book will be produced named Your_Given_Name.csv.

Step 3: Edit the directory in Gene_calling.py and run it. A csv file containing spots coordinate and gene name will be generated, named output_root/whatever_name_processed/readout/mapped_genes.csv.

We have provided example csv file 108_plex_10base_barcodes_padlock_sequence.csv and example barcode book plex_map_filtered_108plex_10base_barcode.csv.

Cell Segmentation

Edit the directory in segment.py and run it. This code will segment cell nucleus according to DAPI channel. A csv file containing the coordinate of nucleus centroid will be generated named output_root/whatever_name_processed/segmented/centroids_all.csv.

Expression Matrix Analysis

Edit the directory and run the jupyter notebook file integrated_analysis_SPRINTseq.ipynb, important post processing results will be shown inside the jupyter.

The provided dataset is just an example data, so we provided another true experimental data for post-analysis in folder example_dataset_whole_brain. Run merge.py to build dataset from splited dataset, and replace mapped_genes.csv and centroids_all.csv mentioned above with file in this folder to use this dataset for Matrix Analysis.