DuncanBrady / ST_Honours_2024

Record of work Completed as part of Robert Bradys 2024 Honors project at Vafaee Lab(UNSW). Project is focused on the downstream analysis of Spatial Transcriptomics Data.
0 stars 0 forks source link

ST_Honours_2024

Record of work Completed as part of Robert Bradys 2024 Honors project at Vafaee Lab(UNSW). Project is focused on the downstream analysis of Spatial Transcriptomics Data.

Supervisory Team

This project is being conducted as collaborative effort between Vafaee Lab and AliveX under the guidance and supervision of Associate Professor Fatemeh Vafaee, Dr Nona Farbehi and Dr Simon Junankar

Project Description

The field of Spatial transcriptomics has seen major development in recent years providing new opportunities for Transcriptomic analysis techniques. This project seeks to leverage the newly available spatial information from spatial transcriptomic technologies to explore the prediction and analysis of cell-cell communication using machine learning techniques.

Project Scope

This project explores the application of Graph Neural Networks to Spatial Transcriptomic data within the context of Cellular communication analysis. All data being examined is sourced from Spatial Transcriptomic technologies, specifically MERfish and Seqfish datasets.

Aims

The goals of the project are to explore the effectiveness of Graph Neural Networks at prediction of cell-cell interaction when treated as a link prediction task for Spatial transcriptomic data. In order to do so the following considerations will be addressed

Methods

Data Processing

Model Architecture

GCN Model

Attention Based Model

Edge Based Attetion Model

Training and Evaluation

Data Availability