PropaNet discovers the dynamics of TF networks against external stress such as heat stress or cold stress.
Given time-series gene expression profile data and template network, PropaNet retrieves master regulatory TFs in each time-point.
To download all the examples, simply clone this repository with the following command:
git clone https://github.com/bhi-kimlab/PropaNet.git
To run them, you will need the following dependencies
Python version 2.7 needs to be installed(The code may not work with Python3).
In addition, a number of libraries are used in the analysis.
Some of the non-default packages include scipy.stats, networkx and multiprocssing.
There are three types of input files, gene expression profile data, DEG binary data and template network.
Each data should take the following format:
Time-series gene expression data has to be stored in a single tab-delimited matrix form. There should be a header as the first row and it should take the following format:
| Gene_id | Condition-time1 | Condition-time2 | Condition-time3 | Condition-time4 | ... |
|---|---|---|---|---|---|
| Gene1 | expression level | expression level | expression level | expression level | ... |
| Gene2 | expression level | expression level | expression level | expression level | ... |
| Gene3 | expression level | expression level | expression level | expression level | ... |
| ... | ... | ... | ... | ... | ... |
gene Heat-Shoots-0.25h Heat-Shoots-0.5h Heat-Shoots-1h Heat-Shoots-3h
AT1G01010 4.91798979107909 4.81768901700114 5.67991373730826 6.02273932823551
AT1G01030 4.45806149779191 4.53272006731038 5.36916581768974 5.51771002546865
AT1G01040 3.8552645118378 3.87640592858555 4.42911094757011 4.83351626001844
AT1G01050 -2.84787294715136 -2.64691204175781 -3.05377896002984 -3.87740186446588
AT1G01060 -0.899222494190051 0.609781349298916 -0.985720858842949 1.57868637136184
AT1G01070 -0.899064779308411 0.609662493243061 0.985561143281074 1.57841350780699
AT1G01080 -0.89887322954867 -0.609596150621448 0.98554170116417 1.57833846949708
AT1G01090 0.898799238909556 0.609558025275848 0.985367888674293 -1.57830555053209
AT1G01100 -0.898776072706743 -0.609434153689444 -0.985195207884564 1.57828766511001
DEG binary indication data is a matrix that indicates which genes in our expression data are differentially expressed genes(DEGs).
If a value in the DEG binary indication data matrix is "1", it would mean that the corresponding gene at the same location in our expression profile data is differentially expressed.
The following is our example DEG binary data.
gene Heat-Shoots-0.25h Heat-Shoots-0.5h Heat-Shoots-1h Heat-Shoots-3h
AT1G01010 1 1 -1 -1
AT1G01030 0 0 0 -1
AT1G01040 0 0 0 0
AT1G01050 0 0 0 0
AT1G01060 0 0 -1 -1
AT1G01070 0 0 0 -1
AT1G01080 0 0 0 -1
AT1G01090 0 0 0 -1
AT1G01100 0 0 0 0
...
The template network file should be comprised of 2 columns : One for source nodes and one for target nodes.
There should be no header in the first row.
| Source gene | Target gene |
|---|---|
| Source gene1 | Target gene1 |
| Source gene2 | Target gene2 |
| Source gene2 | Target gene2 |
| ... | ... |
AT1G01060 AT1G01470
AT1G01060 AT1G01500
AT1G01060 AT1G06460
AT1G01060 AT1G06470
AT1G01060 AT1G10200
AT1G01060 AT1G10350
AT1G01060 AT1G13590
AT1G01060 AT1G16840
AT1G01060 AT1G19630
AT1G01060 AT1G20030
...
The output is a network(in the form of an edge list) comprised of the resulting TFs/TGs that are found by PropaNet.
- To run PropaNet with our example data, simply run the
run.shscript in command line as follows.
bash run.sh
- In order to run PropaNet with another dataset,
- Prepare the data in the format that is described above.
- run the
run.shby feeding in the arguments as follows
bash run.sh <Gene expression profile data> <DEG binary indication data> <Template network> <(optional)additional gene list>
For example, if we explicitly specify the arguments for our example data, it would look like this
bash run.sh data/DEG.AtGenExpress.signed_zstats.heat_shoots data/DEG.AtGenExpress.signed_binary.heat_shoots data/Ath_TF_list.gene