Gene Imputation

This tutorial performs continuous gene expression reconstruction via Graph Autoencoder (GAE) and Implicit Neural Network (INR) in both 2D and 3D spatial transcriptomics examples. We further describe how accuracy is quantitatively evaluated, between predicted and ground-truth gene expression at Quantitative Assessment.

Contents:

• Gene imputation on 2D spatial transcriptomics data
  • Read the Data
  • Step1: Train GAE
  • Step2: Train INR + fine-tune GAE
  • Step3: Prediction/Imputation
• Gene imputation on 3D spatial transcriptomics data
  • Read the Data
  • Step1: Train GAE
  • Step2: Train INR + fine-tune GAE
  • Step3: Prediction/Imputation