Spyne

Description

Spyne is a Python package designed for advanced analysis and visualization of neuronal morphologies and dendritic spines calcium imaging data. 🧠🔬

spyne parses, processes, visualizes and analyzes calcium event data. It is tailored for working with comprehensive datasets from 2-photon microscopy collected using Vidrio ScanImage software.

---

Installation

Note to GPU users

To use the GPU environment, you must have an NVIDIA GPU and the appropriate CUDA drivers installed. See the NVIDIA CUDA Installation Guide for instructions. Make sure your CUDA version is compatible with the versions of deep learning libraries specified in environment.yml.

How to install:

To install spyne, clone the repository and create a conda environment using the appropriate YAML file for your hardware:

git clone https://github.com/dcupolillo/spyne.git
cd spyne

# For CPU-only users:
conda env create -f environment_CPU.yml

# For GPU users (NVIDIA CUDA installed):
conda env create -f environment.yml

# Activate env
conda activate spyne-env  # or spyne-env-cpu for CPU-only

CPU vs GPU environments:

• environment_CPU.yml is for users without a dedicated GPU. It installs CPU-compatible versions of all dependencies.
• environment.yml is for users with an NVIDIA GPU and CUDA drivers. It installs GPU-accelerated packages (e.g., tensorflow-gpu).

If you only want to use plotting and data loading features (not deep learning/segmentation), the CPU environment is sufficient. To perform analysis such as spine segmentation and calcium event classification, GPU is required.

---

Features

spyne is composed of 3 main elements:

1. ImagingDataset : manages the actual imaging data and metadata;
2. SpineDataset : performs spine segmentation and calcium imaging analysis.
3. EphyDataset: manages and analyzes patch-clamp data.

Imaging dataset

The ImagingDataset class is the primary handler for calcium imaging data. It integrates data from 2-photon microscopy experiments, organizing data hierarchically as follows:

Dataset
└── ROIs
    └── Sweeps
        └── Channels
            └── Frames

Spine dataset

spine dataset

Ephy dataset

ephy dataset

---

Dependencies

ROIpy

spyne depends and relies on structural data generated using ROIpy (find repository here).

Example Usage:

import spyne
from pathlib import Path

data_folder = Path("data")  # where your data is stored
date = "250101"  # YYMMDD format
cell_n = "cell0001"  #cell000n format 
imaging_folder = data_folder / date / cell_n / "neuron"

# Initialize the dataseta
dataset = spyne.ImagingDataset(imaging_folder)
spine_dataset = spyne.SpineDataset(dataset)
ephy_dataset = spyne.EphyDataset(dataset)

---

Example usage

ImagingDataset

SpineDataset

Key Features

• Calcium Imaging Data Processing:

  • Handle dF/F traces and z-scores.
  • Detect and binarize calcium events with customizable thresholds.
  • Segment calcium traces into spines and neurites.

• Region of Interest (ROI) Analysis:

  • Automatically generate and optimize rectangular ROIs for scanning.
  • Assign ROIs to specific neuronal branches and compute branch attributes.
  • Handle transformation matrices for pixel-to-reference coordinate mapping.

• Visualization:

  • Plot neuronal structures with color-coded branches.
  • Visualize calcium traces for individual spines and neurites.
  • Generate ROI layouts for scan fields.

---