NeuralCVD / Git / Diff of /README.md

Models:

RaymondKing/

NeuralCVD

Downloads: 1

Diff of /README.md [aca2dc] .. [5f51c3]

Switch to unified view


<div align="center">    



**Neural network-based integration of polygenic and clinical information: Development and validation of a prediction model for 10 year risk of major adverse cardiac events in the UK Biobank cohort**

[![Paper](https://img.shields.io/badge/TheLancet-DigitalHealth-informational)](https://www.thelancet.com/journals/landig/article/PIIS2589-7500(21)00249-1/fulltext)

</div>
 
## Description   
Code related to the paper "Neural network-based integration of polygenic and clinical information: Development and validation of a prediction model for 10 year risk of major adverse cardiac events in the UK Biobank cohort". 
This repo is a python package for preprocessing UK Biobank data and preprocessing, training and evaluating the NeuralCVD score.



## Methods
**NeuralCVD** is based on the fantastic [Deep Survival Machines](https://arxiv.org/abs/2003.01176) Paper, the original implementation can be found [here](https://github.com/autonlab/DeepSurvivalMachines).

## Assets
This repo contains code to preprocess [UK Biobank](https://www.ukbiobank.ac.uk/) data, train the NeuralCVD score and analyze/evaluate its performance.

- Preprocessing involves: parsing primary care records for desired diagnosis, aggregating the cardiovascular risk factors analyzed in the study and calculating predefined polygenic risk scores.
- Training involves Model specification via pytorch-lightning and hydra.
- Postprocessing involve extensive benchmarks with linear Models, and calculation of bootstrapped metrics.


## How to train the NeuralCVD Model  
1. First, install dependencies   
```bash
# clone project   
git clone https://github.com/thbuerg/NeuralCVD

# install project   
cd NeuralCVD
pip install -e .   
pip install -r requirements.txt
 ```   

2. Download UK Biobank data. Execute preprocessing notebooks on the downloaded data.

3. Edit the `.yaml` config files in `neuralcvd/experiments/config/`:
```yaml
setup:
  project_name: <YourNeptuneSpace>/<YourProject>
  root_dir: absolute/path/to/this/repo/
experiment:
  tabular_filepath: path/to/processed/data
```

4. Set up [Neptune.ai](https://www.neptune.ai)

5. Train the NeuralCVD Model (make sure you are on a machine w/ GPU)
 ```bash
# module folder
cd neuralcvd

# run training
bash experiments/run_NeuralCVD_S.sh
```

## Citation   
```
@article{steinfeldt2022neural,
  title={Neural network-based integration of polygenic and clinical information: development and validation of a prediction model for 10-year risk of major adverse cardiac events in the UK Biobank cohort},
  author={Steinfeldt, Jakob and Buergel, Thore and Loock, Lukas and Kittner, Paul and Ruyoga, Greg and zu Belzen, Julius Upmeier and Sasse, Simon and Strangalies, Henrik and Christmann, Lara and Hollmann, Noah and others},
  journal={The Lancet Digital Health},
  volume={4},
  number={2},
  pages={e84--e94},
  year={2022},
  publisher={Elsevier}
}
```  

	a/README.md		b/README.md
1	<div align="center">	1	<div align="center">
2		2
3	![Logo](./src/neuralcvd_logo.png?raw=true "Logo")	3
4
5	Neural network-based integration of polygenic and clinical information: Development and validation of a prediction model for 10 year risk of major adverse cardiac events in the UK Biobank cohort	4	Neural network-based integration of polygenic and clinical information: Development and validation of a prediction model for 10 year risk of major adverse cardiac events in the UK Biobank cohort
6		5
7	[![Paper](https://img.shields.io/badge/TheLancet-DigitalHealth-informational)](https://www.thelancet.com/journals/landig/article/PIIS2589-7500(21)00249-1/fulltext)	6	[![Paper](https://img.shields.io/badge/TheLancet-DigitalHealth-informational)](https://www.thelancet.com/journals/landig/article/PIIS2589-7500(21)00249-1/fulltext)
8		7
9	</div>	8	</div>
10		9
11	## Description	10	## Description
12	Code related to the paper "Neural network-based integration of polygenic and clinical information: Development and validation of a prediction model for 10 year risk of major adverse cardiac events in the UK Biobank cohort".	11	Code related to the paper "Neural network-based integration of polygenic and clinical information: Development and validation of a prediction model for 10 year risk of major adverse cardiac events in the UK Biobank cohort".
13	This repo is a python package for preprocessing UK Biobank data and preprocessing, training and evaluating the NeuralCVD score.	12	This repo is a python package for preprocessing UK Biobank data and preprocessing, training and evaluating the NeuralCVD score.
14		13
15	![NeuralCVD](./src/neuralcvd_fig1.png?raw=true "NeuralCVD")	14
16
17	## Methods	15	## Methods
18	NeuralCVD is based on the fantastic [Deep Survival Machines](https://arxiv.org/abs/2003.01176) Paper, the original implementation can be found [here](https://github.com/autonlab/DeepSurvivalMachines).	16	NeuralCVD is based on the fantastic [Deep Survival Machines](https://arxiv.org/abs/2003.01176) Paper, the original implementation can be found [here](https://github.com/autonlab/DeepSurvivalMachines).
19		17
20	## Assets	18	## Assets
21	This repo contains code to preprocess [UK Biobank](https://www.ukbiobank.ac.uk/) data, train the NeuralCVD score and analyze/evaluate its performance.	19	This repo contains code to preprocess [UK Biobank](https://www.ukbiobank.ac.uk/) data, train the NeuralCVD score and analyze/evaluate its performance.
22		20
23	- Preprocessing involves: parsing primary care records for desired diagnosis, aggregating the cardiovascular risk factors analyzed in the study and calculating predefined polygenic risk scores.	21	- Preprocessing involves: parsing primary care records for desired diagnosis, aggregating the cardiovascular risk factors analyzed in the study and calculating predefined polygenic risk scores.
24	- Training involves Model specification via pytorch-lightning and hydra.	22	- Training involves Model specification via pytorch-lightning and hydra.
25	- Postprocessing involve extensive benchmarks with linear Models, and calculation of bootstrapped metrics.	23	- Postprocessing involve extensive benchmarks with linear Models, and calculation of bootstrapped metrics.
26		24
27		25
28	## How to train the NeuralCVD Model	26	## How to train the NeuralCVD Model
29	1. First, install dependencies	27	1. First, install dependencies
30	```bash	28	```bash
31	# clone project	29	# clone project
32	git clone https://github.com/thbuerg/NeuralCVD	30	git clone https://github.com/thbuerg/NeuralCVD
33		31
34	# install project	32	# install project
35	cd NeuralCVD	33	cd NeuralCVD
36	pip install -e .	34	pip install -e .
37	pip install -r requirements.txt	35	pip install -r requirements.txt
38	```	36	```
39		37
40	2. Download UK Biobank data. Execute preprocessing notebooks on the downloaded data.	38	2. Download UK Biobank data. Execute preprocessing notebooks on the downloaded data.
41		39
42	3. Edit the `.yaml` config files in `neuralcvd/experiments/config/`:	40	3. Edit the `.yaml` config files in `neuralcvd/experiments/config/`:
43	```yaml	41	```yaml
44	setup:	42	setup:
45	project_name: <YourNeptuneSpace>/<YourProject>	43	project_name: <YourNeptuneSpace>/<YourProject>
46	root_dir: absolute/path/to/this/repo/	44	root_dir: absolute/path/to/this/repo/
47	experiment:	45	experiment:
48	tabular_filepath: path/to/processed/data	46	tabular_filepath: path/to/processed/data
49	```	47	```
50		48
51	4. Set up [Neptune.ai](https://www.neptune.ai)	49	4. Set up [Neptune.ai](https://www.neptune.ai)
52		50
53	5. Train the NeuralCVD Model (make sure you are on a machine w/ GPU)	51	5. Train the NeuralCVD Model (make sure you are on a machine w/ GPU)
54	```bash	52	```bash
55	# module folder	53	# module folder
56	cd neuralcvd	54	cd neuralcvd
57		55
58	# run training	56	# run training
59	bash experiments/run_NeuralCVD_S.sh	57	bash experiments/run_NeuralCVD_S.sh
60	```	58	```
61		59
62	## Citation	60	## Citation
63	```	61	```
64	@article{steinfeldt2022neural,	62	@article{steinfeldt2022neural,
65	title={Neural network-based integration of polygenic and clinical information: development and validation of a prediction model for 10-year risk of major adverse cardiac events in the UK Biobank cohort},	63	title={Neural network-based integration of polygenic and clinical information: development and validation of a prediction model for 10-year risk of major adverse cardiac events in the UK Biobank cohort},
66	author={Steinfeldt, Jakob and Buergel, Thore and Loock, Lukas and Kittner, Paul and Ruyoga, Greg and zu Belzen, Julius Upmeier and Sasse, Simon and Strangalies, Henrik and Christmann, Lara and Hollmann, Noah and others},	64	author={Steinfeldt, Jakob and Buergel, Thore and Loock, Lukas and Kittner, Paul and Ruyoga, Greg and zu Belzen, Julius Upmeier and Sasse, Simon and Strangalies, Henrik and Christmann, Lara and Hollmann, Noah and others},
67	journal={The Lancet Digital Health},	65	journal={The Lancet Digital Health},
68	volume={4},	66	volume={4},
69	number={2},	67	number={2},
70	pages={e84--e94},	68	pages={e84--e94},
71	year={2022},	69	year={2022},
72	publisher={Elsevier}	70	publisher={Elsevier}
73	}	71	}
74	```	72	```