AI-MARRVEL

AI-MARRVEL (AIM) is an AI system for rare genetic disease diagnosis.

It takes as input patient VCF and phenotype (formatted with HPO) to predict the causal variant(s).
In making prediction, it takes variant annotation from MARRVEL database and more, and generates prediction score + confidence score as output.

You can use AI-MARRVEL from our website or follow the documentation to run locally.

:new: Our paper is now published in NEJM AI!

Quick Start

Install Required Data Dependencies

AIM utilizes various databases for variant annotation, all of which have been compiled and are available for download. We use AWS S3 for data access, and the data can be downloaded by following these steps:

Install the AWS CLI: Follow the instructions provided in the AWS CLI Installation Guide.
Navigate to Your Desired Directory: Change to the directory where you want your data dependencies downloaded. For example, in Ubuntu, use:
```
cd <desired/folder/path>
```

Use the following command to sync the S3 bucket to your local directory:

aws s3 sync s3://aim-data-dependencies-public . --no-sign-request
### Get the software
AIM is released as a Docker image for easy distribution. To get it:

docker pull chaozhongliu/aim-lite:latest

Run with your sample

docker run -v <Path to VCF File>:/input/vcf.gz \
           -v <Path to HPO file>:/input/hpo.txt \
           -v <Path to downloaded database>:/run/data_dependencies \
           -v <Path to output folder>:/out \
       chaozhongliu/aim-lite /run/proc.sh [Sample ID] [Reference genome: hg19/hg38] [Memory Limit (G)]

License

AI-MARRVEL is licensed under GPL-3.0. You are welcomed to use it for research purpose.
For business purpose, please contact us for licensing.

Disclaimer

Some of the data and software included in the distribution may be subject to third-party constraints. Users of the data and software are solely responsible for establishing the nature of and complying with any such restrictions.
AI-MARRVEL provides this data and software in good faith, but make no warranty, express or implied, nor assume any legal liability or responsibility for any purpose for which they are used.

Citing AI-MARRVEL

@article{doi:10.1056/AIoa2300009,
author = {Dongxue Mao  and Chaozhong Liu  and Linhua Wang  and Rami AI-Ouran  and Cole Deisseroth  and Sasidhar Pasupuleti  and Seon Young Kim  and Lucian Li  and Jill A. Rosenfeld  and Linyan Meng  and Lindsay C. Burrage  and Michael F. Wangler  and Shinya Yamamoto  and Michael Santana  and Victor Perez  and Priyank Shukla  and Christine M. Eng  and Brendan Lee  and Bo Yuan  and Fan Xia  and Hugo J. Bellen  and Pengfei Liu  and Zhandong Liu },
title = {AI-MARRVEL — A Knowledge-Driven AI System for Diagnosing Mendelian Disorders},
journal = {NEJM AI},
volume = {1},
number = {5},
pages = {AIoa2300009},
year = {2024},
doi = {10.1056/AIoa2300009},

URL = {https://ai.nejm.org/doi/abs/10.1056/AIoa2300009},
eprint = {https://ai.nejm.org/doi/pdf/10.1056/AIoa2300009}
,
    abstract = { AI-MARRVEL is an AI system for genetic diagnosis that improves diagnostic accuracy, surpassing state-of-the-art benchmarked methods. }
}

LiuzLab / AI_MARRVEL

readme