GlioReg

GlioReg is a research model for learning gene representations that support regulatory activity inference in glioblastoma. This public repository contains the minimal runtime model code: the gene encoder, cell encoder, decoder heads, and the top-level GlioReg module.

The development repository, benchmark code, internal notes, generated results, large data assets, trained checkpoints, and manuscript materials are maintained separately.

Install

pip install -e .

Runtime Import

from glioreg import GlioReg

model = GlioReg(
    gene_input_dim=128,
    n_genes=128,
    gene_hidden_dim=64,
    gene_latent_dim=32,
    cell_embedding_dim=32,
)

Assets

Large training data, benchmark assets, and model artifacts are stored outside GitHub. Public asset download instructions will be added after the release artifact set is finalized.