Authors: Xiaoxin Ye and Joshua W. K. Ho
Contact: [email protected]
Copyright © 2018, Victor Chang Cardiac Research Institute
Our FlowGrid algorithm could be applied into many format data set but the sample code only accept csv format. In the csv file, the first row is feature name and each columns is seperated by ",". If you have true label file , you could use --l filename to input label file for testing the ARI of FlowGrid result.
Before using the package, we need to install the dependent package sklearn and numpy.
pip install -r requirements.txt --useror
pip install sklearn numpy scipy --userA summary of the argument of sample code is included in the table below.
| Argument | Usage | Required? |
|---|---|---|
| --f | the input file name | required |
| --n | number of bins | required |
| --eps | maximun distance between two bins | required |
| --t | threshold for high density bin | optional (default:40) |
| --o | the output file name | optional (default: out.csv) |
| --l | the true label file name | optional |
After installing all the dependent packages, you could try to use the sample code to run FlowGrid on the sample data.
python sample_code.py --f sample_data.csv --n 4 --eps 1.1 --l sample_label.csvThe predicted label is saved at out.csv and the sample result is as follow.
The number of cells is: 23377
The number of dimensions is: 4
runing time: 0.027
ARI:0.9816
Authors: Xiunan Fang and Joshua W. K. Ho
Contact: [email protected]
Copyright © 2019, Joshua W. K. Ho's Lab
FlowGrid implementation in Scanpy can be found in github PAGE: https://github.com/holab-hku/FlowGrid
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent