HyunSBong / CVAE-Transformer-BRCA

Research on subtype classification, data augmentation, and drug prediction of breast cancer using human genome and transcriptional information

CVAE-Transformer-BRCA

Overview

RNA-seq generation model | Paper | Code

• 

---

Modified it to Pytorch (gene level self-attention) | Code

• 
• Original moBRCA-net | Paper | Code
• 

---

Attention-based architecture for diagnosis and prognosis from omics data | Code

• 

---

BRCA_weighted_DEGs.R - Top 1,000 genes with high weights were selected from the miRNA-TF-mRNA gene regulatory networks.

• Reference Statistical Analysis | Paper

• Subtypes	Basal-like	Her2	LumA	LumB	Normal-like
  Weighted DEGs	376	157	249	206	249

---

Dataset

• GTEx(Genotype-Tissue Expression) Dataset
• TCGA(Cancer Genome Atlas) Dataset
• L1000 landmark gene set
• RNA-seq(human transcriptomics) Dataset (9147 samples and 18154 genes)
• TCGA-BRCA (PAM50)

Install dependencies

• torch >= 1.12.1
• umap-learn >= 0.5.3
• scikit-learn >= 1.1.1