Home

Welcome to the Post Mortem Interval (PMI) Prediction App. This application helps you predict the PMI (Hour) based on tissue type, circRNA (or other RNAs), expression levels, and temperature.




About the Lab

Key Laboratory of Epigenetics and Oncology, the Research Center for Preclinical Medicine

Director: Prof. Junjiang Fu

Address: Southwest Medical University, Luzhou 646000, Sichuan, China



Please kindly cite following paper to support further development:

Options

Results

Options

Results



                        

Transfer Learning Details

Options

Website Usage Statistics

Detailed Statistics

Help

Application Overview

This app predicts Post Mortem Interval (PMI) using circRNA expression data with Support Vector Machines (SVM) and Transfer Learning technology.

Section Descriptions:

1. Home

Provides an introduction to the application and information about the lab. Includes citation information for the research paper this tool is based on.

2. Mouse Data

Predict PMI using mouse data. Features include:

  • Select tissue type and circRNA from dropdown menus
  • Adjust temperature and expression values
  • View 3D visualization of predictions
  • See statistical plots comparing actual vs predicted values
  • Access detailed model information
3. Human Data

Predict PMI using human data with optional transfer learning from mouse models. Features include:

  • Same input options as Mouse Data section
  • Toggle transfer learning option (enabled by default)
  • View transfer learning performance plots
  • Access QQ plots for model diagnostics
4. User Data

Upload and analyze your own data. Features include:

  • Upload Excel files with your experimental data
  • Select relevant columns from your data
  • Generate predictions using your custom dataset
  • View 3D visualizations specific to your data
5. Website Statistics

Track application usage metrics including:

  • Total views and daily views
  • Number of unique users
  • Detailed access logs
  • Interactive charts of usage patterns
6. Publications
For reading and helping

Data Requirements:

For user uploads, your Excel file must contain these columns:

  • Tissue - Tissue type
  • circRNA - circRNA identifier
  • Expression - Expression level
  • Temperature - Temperature in Celsius
  • PMI_Hour - Actual PMI in hours (for model training)

Technical Notes:

The application uses:

  • SVM with RBF kernel for modeling
  • Transfer learning for mouse-to-human predictions
  • 3D visualization with convex hulls for data boundaries
  • Interactive plots with Plotly

Note: All predictions should be interpreted by qualified professionals in context with other forensic evidence.



Support Information

For support, please contact us at:

Name: Mazaher Maghsoudloo

Email: mazaher@swmu.edu.cn

                babak1146@gmail.com