DITTO-UI

Easy to use web interface for biologists to look for any type of small genetic variant and understand their deleteriousness using DITTO scores.

!!! For research purposes only !!!

Description

A web app where one can lookup small variants and understand the deleteriousness using DITTO deleterious score and Clinvar reported significance. DITTO uses an explainable neural network model to predict the functional impact of variants and utilizes SHAP to explain the model's predictions. It is trained on variants from ClinVar and uses OpenCravat for annotations from various data sources. The higher the score, the more likely the variant is deleterious.

Data

DITTO-UI comprises DITTO scores for any variant using annotations from openCravat for making predictions. Annotations are extracted from openCravat API query on page render.

Usage

DITTO-UI is deployed on the Streamlit Cloud: DITTO-UI site. Here's an example on how it looks like

Installation

Installation simply requires fetching the source code. Following are required:

• Git

To fetch source code, change in to directory of your choice and run:

    git clone https://github.com/uab-cgds-worthey/DITTO-UI.git

Requirements

OS:

This app is developed and tested in Mac OS.

Tools:

• Python 3.11
• Pip 23.3

Environment:

• python virtual environment

# create environment. Needed only the first time. Please use the above link if you're not using Mac.
python -m venv ditto-env

source ditto-env/bin/activate

Install required packages

Change in to root directory and run the commands below:

# Install dependencies in the environment. Needed only the first time.
pip3 install -r requirements.txt

Run DITTO-UI locally

To run DITTO-UI locally make sure the environment has been succesfully made and then run the following commands

# run the DITTO-UI application using Streamlit
streamlit run src/Home.py

Once the application has started up it will open a new tab in your default browser to DITTO-UI

Contact Info

Tarun Mamidi | tmamidi@uab.edu