
# DL-mo
## A benchmark study of deep learning based multi-omics data fusion methods for cancer
***
![Multi-Omics](https://easymed.ai/models/AlyssaS/DL-mo/git/ci/main/tree/img/3.png "Multi-Omics")  
We here compare the performances of 10 deep learning methods in three contexts: 
1. Simulated datasets
2. Cancer datasets
3. Single-cell datasets       

We use `python` and `R` to code the programs.  The python scripts are in `./python-scripts/` folder .The R scripts are in `./R-scripts/` folder .
***
## 16 deep learning methods
* [lfAE](./python-scripts/runCancerAE2.py)
* [efAE](./python-scripts/runCancerAE.py) 
* [lfDAE](./python-scripts/runCancerDAE2.py) 
* [efDAE](./python-scripts/runCancerDAE.py) 
* [lfVAE](./python-scripts/runCancerVAE2.py)
* [efVAE](./python-scripts/runCancerVAE.py)
* [lfSVAE](./python-scripts/runCancerSVAE2.py)
* [efSVAE](./python-scripts/runCancerSVAE.py) 
* [lfmmdVAE](./python-scripts/runCancerMMDVAE2.py) 
* [efmmdVAE](./python-scripts/runCancerMMDVAE.py) 
* [lfNN](./python-scripts/runCancerDNN.py) 
* [efNN](./python-scripts/runCancerDNN.py)
* [lfCNN](./python-scripts/runCancerCNN.py) 
* [efCNN](./python-scripts/runCancerCNN.py)
* [moGCN](./python-scripts/MOGONET/main_mogonet_zly.py)
* [moGAT](./python-scripts/MOGONET/main_mogonet_zyh.py)
***
## Input data
The data for python scripts is in `./python-scripts/data/` folder .The data for R scripts is in `./R-scripts/data/` folder .    
For python-scripts,Simulated datasets are in `./python-scripts/data/simulations`,Cancer datasets are in `./python-scripts/data/cancer` ,Single-cell datasets are in `./python-scripts/data/single-cell`.
***
## python scripts
Each of the three datasets above corresponds to a differnet python scripts in this repositiory:
1. `runSimulations*.py`
2. `runCancer*.py`
3. `runSingle*.py`
***
## R scripts
Each of the three datasets above corresponds to a differnet Jupyter notebook in this repositiory:
1. `simulated*.ipynb`
2. `cancer*.ipynb`
3. `single-cell*.ipynb`

***
## Install the R software environment
Use conda to create a new environment: `conda create -n momix -c conda-forge -c bioconda -c lcantini momix r-irkernel`
***
## Install the python software environment
You need to build a virtual environment for python.    
You need to install the following main libraries: `Python==3.7.0,Tensorflow==1.15.0, scikit-learn==0.20.0, Jupyter==1.0.0`.


	a/README.md		b/README.md
1	# DL-mo	1	# DL-mo
2	## A benchmark study of deep learning based multi-omics data fusion methods for cancer	2	## A benchmark study of deep learning based multi-omics data fusion methods for cancer
3	***	3	***
4	![Multi-Omics](./img/3.png "Multi-Omics")	4	![Multi-Omics](https://easymed.ai/models/AlyssaS/DL-mo/git/ci/main/tree/img/3.png "Multi-Omics")
5	We here compare the performances of 10 deep learning methods in three contexts:	5	We here compare the performances of 10 deep learning methods in three contexts:
6	1. Simulated datasets	6	1. Simulated datasets
7	2. Cancer datasets	7	2. Cancer datasets
8	3. Single-cell datasets	8	3. Single-cell datasets
9		9
10	We use `python` and `R` to code the programs. The python scripts are in `./python-scripts/` folder .The R scripts are in `./R-scripts/` folder .	10	We use `python` and `R` to code the programs. The python scripts are in `./python-scripts/` folder .The R scripts are in `./R-scripts/` folder .
11	***	11	***
12	## 16 deep learning methods	12	## 16 deep learning methods
13	* [lfAE](./python-scripts/runCancerAE2.py)	13	* [lfAE](./python-scripts/runCancerAE2.py)
14	* [efAE](./python-scripts/runCancerAE.py)	14	* [efAE](./python-scripts/runCancerAE.py)
15	* [lfDAE](./python-scripts/runCancerDAE2.py)	15	* [lfDAE](./python-scripts/runCancerDAE2.py)
16	* [efDAE](./python-scripts/runCancerDAE.py)	16	* [efDAE](./python-scripts/runCancerDAE.py)
17	* [lfVAE](./python-scripts/runCancerVAE2.py)	17	* [lfVAE](./python-scripts/runCancerVAE2.py)
18	* [efVAE](./python-scripts/runCancerVAE.py)	18	* [efVAE](./python-scripts/runCancerVAE.py)
19	* [lfSVAE](./python-scripts/runCancerSVAE2.py)	19	* [lfSVAE](./python-scripts/runCancerSVAE2.py)
20	* [efSVAE](./python-scripts/runCancerSVAE.py)	20	* [efSVAE](./python-scripts/runCancerSVAE.py)
21	* [lfmmdVAE](./python-scripts/runCancerMMDVAE2.py)	21	* [lfmmdVAE](./python-scripts/runCancerMMDVAE2.py)
22	* [efmmdVAE](./python-scripts/runCancerMMDVAE.py)	22	* [efmmdVAE](./python-scripts/runCancerMMDVAE.py)
23	* [lfNN](./python-scripts/runCancerDNN.py)	23	* [lfNN](./python-scripts/runCancerDNN.py)
24	* [efNN](./python-scripts/runCancerDNN.py)	24	* [efNN](./python-scripts/runCancerDNN.py)
25	* [lfCNN](./python-scripts/runCancerCNN.py)	25	* [lfCNN](./python-scripts/runCancerCNN.py)
26	* [efCNN](./python-scripts/runCancerCNN.py)	26	* [efCNN](./python-scripts/runCancerCNN.py)
27	* [moGCN](./python-scripts/MOGONET/main_mogonet_zly.py)	27	* [moGCN](./python-scripts/MOGONET/main_mogonet_zly.py)
28	* [moGAT](./python-scripts/MOGONET/main_mogonet_zyh.py)	28	* [moGAT](./python-scripts/MOGONET/main_mogonet_zyh.py)
29	***	29	***
30	## Input data	30	## Input data
31	The data for python scripts is in `./python-scripts/data/` folder .The data for R scripts is in `./R-scripts/data/` folder .	31	The data for python scripts is in `./python-scripts/data/` folder .The data for R scripts is in `./R-scripts/data/` folder .
32	For python-scripts,Simulated datasets are in `./python-scripts/data/simulations`,Cancer datasets are in `./python-scripts/data/cancer` ,Single-cell datasets are in `./python-scripts/data/single-cell`.	32	For python-scripts,Simulated datasets are in `./python-scripts/data/simulations`,Cancer datasets are in `./python-scripts/data/cancer` ,Single-cell datasets are in `./python-scripts/data/single-cell`.
33	***	33	***
34	## python scripts	34	## python scripts
35	Each of the three datasets above corresponds to a differnet python scripts in this repositiory:	35	Each of the three datasets above corresponds to a differnet python scripts in this repositiory:
36	1. `runSimulations*.py`	36	1. `runSimulations*.py`
37	2. `runCancer*.py`	37	2. `runCancer*.py`
38	3. `runSingle*.py`	38	3. `runSingle*.py`
39	***	39	***
40	## R scripts	40	## R scripts
41	Each of the three datasets above corresponds to a differnet Jupyter notebook in this repositiory:	41	Each of the three datasets above corresponds to a differnet Jupyter notebook in this repositiory:
42	1. `simulated*.ipynb`	42	1. `simulated*.ipynb`
43	2. `cancer*.ipynb`	43	2. `cancer*.ipynb`
44	3. `single-cell*.ipynb`	44	3. `single-cell*.ipynb`
45		45
46	***	46	***
47	## Install the R software environment	47	## Install the R software environment
48	Use conda to create a new environment: `conda create -n momix -c conda-forge -c bioconda -c lcantini momix r-irkernel`	48	Use conda to create a new environment: `conda create -n momix -c conda-forge -c bioconda -c lcantini momix r-irkernel`
49	***	49	***
50	## Install the python software environment	50	## Install the python software environment
51	You need to build a virtual environment for python.	51	You need to build a virtual environment for python.
52	You need to install the following main libraries: `Python==3.7.0,Tensorflow==1.15.0, scikit-learn==0.20.0, Jupyter==1.0.0`.	52	You need to install the following main libraries: `Python==3.7.0,Tensorflow==1.15.0, scikit-learn==0.20.0, Jupyter==1.0.0`.
53		53