ISTDECO: In Situ Transcriptomics Decoding by Deconvolution

ISTDECO is a simple tool for deconvolving combinatorial image data generated by *FISH techniques into series of non-combinatorial images. This repository contains the implementation of ISTDECO and collection of notebooks demonstarting the tool.

Example

• The notebook synthetic_example.ipynb contains examples on simulated data.
• The notebook merfish_example.ipynb contains examples on MERFISH data. The data can be downloaded using Starfish python library [1].
• The notebook iss_exampl.ipynb contains examples on ISS images from [2]. The data can be downloaded from [3].

System Requirements

The source code have been developed and tested on a Linux machine running Ubuntu 16.04 OS with 16GB RAM, Intel® Core™ i7-8850H CPU @ 2.60GHz × 12, Nvidia Quadro P4200/PCIe/SSE2 GPU.

Python Library Requirements

The following python packages are required for ISTDECO

• torch
• numpy

The notebooks require the additional packages:

• pandas
• pillow
• scipy
• matplotlib
• starfish
• tqdm

To install we recommend using Anaconda enviroments. For instance:

conda create -n myenv python=3.6

conda activate myenv

pip install starfish[napari] torch pandas pillow scipy matplotlib tqdm

References

[1] starfish: scalable pipelines for image-based transcriptomics https://github.com/spacetx/starfish

[2] Qian, Xiaoyan, et al. "Probabilistic cell typing enables fine mapping of closely related cell types in situ." Nature methods 17.1 (2020): 101-106.

[3] Axel Andersson, Ferran Diego, Fred A. Hamprecht, & Carolina Wählby. (2021). In Situ Sequencing data used in "ISTDECO: In Situ Transcriptomics Decoding by Deconvolution" [Data set]. Zenodo. http://doi.org/10.5281/zenodo.4560687