Synapse-seq – Code and Vignettes

Synapse-seq is a suite of technologies designed to measure both single-neuron molecular expression profiles and long-range neuroanatomical projections. The pre-synaptic system enables detection of distal projections, making it possible to analyze how a neuron's transcriptomic identity relates to its projection behavior.

📘 Vignettes

This repository provides walkthroughs for key analytical components of Synapse-seq.
We recommend starting with the presynaptic vignette, which characterizes cortico-thalamic projections from the primary visual cortex (VISp).

Code

code/preprocessing: contains the scripts used to deduplicate VT dialouts and transform fastqs into cell(bead) / VT matrices.

code/CreateScoringDF.ipynb: walks through the creation of VT scoring for projection analysis.

code/run_analysis: reproduces Presynaptic main figures based on data accessed below.

code/postsynaptic_analysis/post_analysis.ipynb: contains notebook used to create Postsynaptic experiment figures.

---

Data Access

All required data is publicly available in a Google Cloud Storage bucket.

Programmatic Access

Data can be accessed using the Google Cloud SDK Google Cloud SDK at: gs://macosko_public/Synapseseq_data_2026

For executing vignettes, required files will be downloaded at beginning of the notebook.

Manual Access

You can also browse and download the data manually from the following link:
Google Cloud Console – Synapseseq Data (2025)

If one is unable to download the required data from the public google cloud bucket, we have also made a google drive link available with the data needed to run the VISp presynaptic vignette. https://drive.google.com/drive/u/0/folders/1U1V7Rei6ShXLDdc-2EoSN3UIrL0uW_OE

File Info:

Presynapse Analysis (gs://macosko_public/Synapseseq_data_2026/Presynaptic)

• projection_target_vts.pkl Pickle file containing VTs detected at different projection sites.

• vt_freq_objs.pkl Pickle file containing the number of instances each VT was observed in the deeply sequenced pool. Used for calculating library diversity and creating VT score.

VISp – Presynaptic Primary Visual Cortex Injection Experiment

• v1_adata.h5ad
  H5AD (Anndata) of snRNA-seq obtained from the VISp experiment.
• v1_vt_df.csv
  DataFrame containing relationships of cell-barcode to viral tags (VT).
• thalamus_combined_obj.pkl
  Pickle file encoding RNA expression obtained via Slide-seq at the thalamus.
• thalamus_bead_location.csv
  DataFrame containing the relationship between bead barcode and spatial position.
• thalamus_vt_df_bead_merged.csv
  DataFrame containing mappings of VTs found in the thalamus and the bead on which each molecule was identified.

---

AC – Presynaptic Anterior Cortex Injection Experiment

• aca_adata.h5ad
  H5AD (Anndata) of snRNA-seq obtained from the AC experiment.
• aca_vt_df.csv
  DataFrame containing relationships of cell-barcode to VTs.
• str_combined_obj.pkl
  Pickle file encoding RNA expression obtained via Slide-seq in the striatum.
• str_bead_locations.csv
  DataFrame containing the relationship between bead barcode and spatial position.
• str_vt_df_bead_merged.csv
  DataFrame containing mappings of VTs found in the striatum and the bead on which each molecule was identified.

---

Postsynaptic Analysis (gs://macosko_public/Synapseseq_data_2026/Postsynaptic)

Postsynaptic Hippocampus Experiment

• pyr_post_processed.csv
  DataFrame containing the DBSCAN cluster calls of VTs within the pyramidal layer (CA1, CA2, CA3), along with their spatial position (obtained from Slide-seq).
• vt_df_dgl.csv
  DataFrame containing the DBSCAN cluster calls of VTs within the infrapyramidal horn of the DG.
• vt_df_dgu.csv
  DataFrame containing the DBSCAN cluster calls of VTs within the suprapyramidal horn of the DG.
• processed_hippo_RNA.qs
  Seurat object with spatial transcriptomics measurements made by Slide-seq.
• rctd_results.csv
  DataFrame containing mappings of bead-barcode to cell-type proportions (RCTD results).

Create environment

The requirements.yaml file can be used to create a conda environment for which we can use to execute the notebooks.

Ensure a conda solver is installed. We recommend mamba. (https://mamba.readthedocs.io/en/latest/installation/mamba-installation.html)

The requirements.yaml file can be used to create a conda environment for running the notebooks.

 mamba env create -f requirements.yaml -n synapseseq_env
 mamba activate synapseseq_env

Next we need to export the environment so it is accessible via jupyter.

python -m ipykernel install --user --name synapseseq_env --display-name "synapseseq_env"

Then run jupyter-lab

jupyter-lab --ip='*' --port=8989