Belayer: Modeling discrete and continuous spatial variation in gene expression from spatially resolved transcriptomics. Cong Ma*, Uthsav Chitra*, Shirley Zhang, Ben Raphael
Belayer depends on the following python packages: numpy, scipy, pandas, sklearn, networkx, glmpca. Further installation TBD.
See tutorial.ipynb for a complete tutorial of how to run Belayer on three different datasets. Note that the tutorial requires downloading two files from here - one file for the DLPFC tutorial and one for the mouse wound tutorial - and placing them in their respective folders.
python belayer.py (-i <10x directory> | -s <count matrix file> <spatial coordinate file>) -m <running mode> -L <number layers> [options]
Details of required and optional input arguments:
| Argument | Data type | Description |
|---|---|---|
| -i (--indir) | str | Input 10X directory for ST data. |
| -s (--stfiles) | list of str | Input count matrix file followed by spatial coordinate file for ST data. Count matrix and spatial coordinate must have the same number of spots. Only one of -i and -s is allowed. |
| -m (--mode) | char | Running mode. A: axis-aligned layered tissue. R:rotated axis-aligned layered tissue. S:arbitrarily curved tissue supervised by annotated layers. L:layered tissue with linear layer boundaries. |
| -L (--nlayers) | int | Number of layers to infer. |
| -a (--annotation) | str | File of annotated layers for each spot when using S mode. |
| -o (--outprefix) | str | Output prefix. |
| -p (--platform) | str | Platform for spatial transcriptomics data. Only used when running mode is S. |
- <outprefix>_layer.csv. This file contains the identified layers for each spot.
- estimated piecewise function coefficients file TBD, see tutorial for details.