BioVix: An Integrated Large Language Model Framework for Data Visualization, Graph Interpretation, and Literature-Aware Scientific Validation

Table of Contents

1. overview
2. Features
3. Prerequisites
4. BioVix Installation
5. Running BioVix
6. Outputs
7. Deployment
8. Tips for Success
9. Reference
10. License and Issues
11. Authors and Contacts

Overview

BioVix is an AI-assisted visualization tool built on Streamlit that streamlines the workflow from data analysis to literature discovery. By integrating Plotly for visualization, DeepSeek V3.1 for query processing, and Semantic Scholar for bibliographic search, it offers a robust analytical environment. Furthermore, the system incorporates GPT-OSS-20B for structured dialogue and Qwen2.5-VL-32B-Instruct for visual graph reasoning, enabling users to gain a deeper understanding of data trends through natural language interaction.

Features

• Data Upload & Management: Supports CSV, TSV, and Excel (.xlsx) formats.
• AI-Powered Chart Generation: Create interactive Plotly visualizations from natural language queries.
• AI Insights: Automatically analyze and provide explanations for generated charts.
• Academic Research Integration: Discover relevant research papers via Semantic Scholar.
• Graph Interpreter: Analyze uploaded graph images using AI.
• Data Q&A: Ask questions about your dataset and receive AI-driven answers.
• Sample Datasets: Access pre-loaded datasets for quick testing (Apple Stock, Gene Expression, Hospital Data).

Prerequisites

• Python <3.9 and >=3.11
• API keys for:
  • DeepSeek API (via OpenRouter)
  • GPT API (via OpenRouter)
  • Qwen API (via OpenRouter)
  • Semantic Scholar API

BioVix Installation

1. Clone the repository

git clone https://github.com/MuhammadZain-Butt/BioVix.git
cd BioVix

2. Create a Virtual Environment (Recommended)

It is highly recommended to use a virtual environment to avoid dependency conflicts.

Windows:

python -m venv env
env\Scripts\activate

Note: Replace env_name with your preferred name for the virtual environment.

Linux / macOS:

python3 -m venv env_name
source env_name/bin/activate

Note: Replace env_name with your preferred name for the virtual environment.

3. Install Dependencies

pip install -r requirements.txt

4. Configure Environment Variables

Create a .env file in the project root directory (as in BioVix):

DEEPSEEK_API_KEY="your_deepseek_key_here"
GPT_API_KEY="your_gpt_key_here"
QWEN_API_KEY="your_qwen_key_here"
SEMANTIC_SCHOLAR_API_KEY="your_semantic_scholar_key_here"

Tip: If you do not have these API keys, you can create accounts here to generate them:

• OpenRouter
• Semantic Scholar

Running BioVix

After installing dependencies and setting up your environment, you can start BioVix using Streamlit.

streamlit run app.py

Once the command runs, the app will automatically open in your default browser at: http://localhost:8501

Outputs

The following panels illustrate the outputs of BioVix across varying datasets. (A) displays the raw input data, while (B,C,D) presents the corresponding interactive visualization rendered with Plotly. (E) provides the AI-generated interpretation of the graph, along with the derived search query. Finally, (F) lists the relevant research papers retrieved from Semantic Scholar using the formulated query:

• Figures:

  1. Gene-level Protein Expression Dataset

     

  2. Peak Annotation dataset

     

  3. Clinical Diabetic Dataset

     

Deployment

BioVix is deployed on Hugging Face and can be tested or used directly, click here

Tips for Success

• Ensure that input files are correctly formatted (e.g., CSV, XLSX, or TSV) and contain all information required for visualization.
• Write queries in a clear and detailed manner, and avoid using informal language.
• Use consistent naming conventions for columns and variables to improve clarity and interpretation.

References

In Process.

License and Issues

This BioVix is licensed under the MIT License - see the LICENSE file for details. Submit issues or contributions via GitHub Issues.

Authors and Contacts

Mr. Muhammad Zain Butt
Integrative Omics and Molecular Modeling Laboratory, Department of Bioinformatics and Biotechnology, Government College University Faisalabad (GCUF), Faisalabad, 38000, Pakistan
Email: zain.202302328@gcuf.edu.pk

Mr. Rana Sheraz Ahmad
Integrative Omics and Molecular Modeling Laboratory, Department of Bioinformatics and Biotechnology, Government College University Faisalabad (GCUF), Faisalabad, 38000, Pakistan
Email: ranasheraz.202101902@gcuf.edu.pk

Ms. Eman Fatima
Integrative Omics and Molecular Modeling Laboratory, Department of Bioinformatics and Biotechnology, Government College University Faisalabad (GCUF), Faisalabad, 38000, Pakistan
Email: eman.202204127@gcuf.edu.pk

Dr. Muhammad Tahir ul Qamar (Correspondence)
Integrative Omics and Molecular Modeling Laboratory, Department of Bioinformatics and Biotechnology, Government College University Faisalabad (GCUF), Faisalabad, 38000, Pakistan
Email: m.tahirulqamar@hotmail.com