Description
With IFC-Seq we couple two distinct modalities of single-cell data: (1) Single Cell Transcriptomics (SCT) and (2) Imaging Flow Cytometry (IFC). Both modalities are aligned using common surface markers (CD34 and FcgR in this example). By leveraging both views at the same time, IFC-Seq is able to predict gene expression at the single-cell level for data acquired from IFC experiments.
This is a supplementary tutorial to the methodology described in N.K. Chlis et al., 2019. A detailed step-by-step tutorial is presented in tutorial_ifcseq_mouse.ipynb. A concise demnostration of how to use ifcseq as a python package is demonstrated in tutorial_ifcseq_human.ipynb.
Instructions
- Clone the repository by using "git clone https://github.com/theislab/ifcseq.git"
- Download the mouse data from here and unzip into the repository, it should automatically unzip into ./ifcseq/ifcseq_mouse_data/
- Download the human data from here and unzip into the repository, it should automatically unzip into ./ifcseq/ifcseq_human_data/
- Open and run the tutorial notebook of choice
List of materials
- tutorial_ifcseq_mouse.ipynb: The notebook presenting IFC-Seq in detail, step by step on the mouse data
- tutorial_ifcseq_human.ipynb: The notebook demonstrating how to use the ifcseq pyhon package on the human data. Using ifcseq as a python package assumes that the surface markers are available for the IFC dataset. Either expermentally, or predicted in a label-free manner as shown in the previous notebook on the mouse data.
- CNN_train.py: Python code used to train the CNN keras model which is used in tutorial_ifcseq_mouse.ipynb. For convenience, a pre-trained model is availabe in ifcseq_mouse_data.zip (see below)
- ifcseq_mouse_data.zip: The mouse data and pre-trained models necessary to run the notebook
- ifcseq_human_data.zip: Additional human data used in the publication.