Project 99: Development of AI Models for SERSome Metabolomics Analysis

Contact Information

Prof. Jian Ye/ Dr. Zhou Chen

Project Description and Objectives

Metabolomics hold significant promise for the early diagnosis of diseases, which can identify specific biomarkers associated with disease onset. For more rapid and sensitive metabolomics analysis of the biofluids, SERSome (Cell Reports Medicine, 2024) technique has been recently proposed with advantages of low-cost, ultra-high throughput and single-molecule sensitivity. However, the complexity and volume of SERSome data pose significant analytical challenges.

This project will focus on developing reliable AI models specialized for comprehensive SERSome metabolomics analysis. Deep learning architectures (e.g., convolutional neural networks, U-net) can extract latent feature information from massive data. By integrating cutting-edge AI methods with metabolomics, this project will pave the way for transformative advancements in early disease diagnosis and precision medicine.

Eligibility Requirements

Background in artificial intelligence and clinical diagnosis

Prerequisite of programming skills using Python and AI modelling using PyTorch

Enthusiasm to advancing the in vitro diagnostics (IVD) techniques

Main Tasks

Preprocessing Raman spectra: baseline reduction and denoising

Developing deep learning architectures based on the order-invariant SERSome for metabolomics analysis

Training and fine-tuning AI models

Writing final reports.

Website

Lab: http://www.yelab.sjtu.edu.cn/

School: https://bme.sjtu.edu.cn/