LEGEND: Identifying Co-expressed Genes in Multimodal Transcriptomic Sequencing Data

We present a novel method called muLtimodal co-Expressed GENes finDer (LEGEND) that performs integrated gene clustering on scRNA-seq and SRT data to identify co-expressed genes at both the cell type and tissue domain levels. LEGEND performs a hierarchical gene clustering with the aim of maximizing intra-cluster redundancy and inter-cluster complementarity.

Dependencies

• anndata>=0.8.0
• numpy>=1.21.6
• scanpy>=1.9.3
• loguru>=0.6.0
• scipy>=1.9.3
• hdbscan>=0.8.29
• scikit-learn>=1.2.0
• pandas>=1.5.2
• SpaGCN>=1.2.5
• igraph>=0.10.2
• leidenalg>=0.9.1
• squidpy>=1.2.2
• torch>=1.13.1
• opencv-python>=4.6.0

Installation

You can download the package from GitHub and install it locally:

git clone https://github.com/ToryDeng/LEGEND.git
cd LEGEND/
pip install dist/LEGEND-0.1.1-py3-none-any.whl

Getting started

If adata_rna is the sc/snRNA-seq dataset and adata_st is the ST dataset (the histology image is optional), you can run LEGEND as follows:

import LEGEND as lg

info_rna, _ = lg.GeneClust(adata_rna, return_info=True)
# if the histology image is available
info_st, _ = lg.GeneClust(adata_st, image=img, return_info=True)
# if the histology image is not available
info_st, _ = lg.GeneClust(adata_st, return_info=True)

integration_info, integrated_genes = lg.integrate(info_rna, info_st, return_info=True)

For more details about how to call the functions, please refer to the Tutorial.

Tested environment

Environment 1

• CPU: Intel(R) Xeon(R) Platinum 8255C CPU @ 2.50GHz
• Memory: 256 GB
• System: Ubuntu 20.04.5 LTS
• Python: 3.9.15

Environment 2

• CPU: Intel(R) Xeon(R) Gold 6240R CPU @ 2.40GHz
• Memory: 256 GB
• System: Ubuntu 22.04.3 LTS
• Python: 3.9.18

Citation

Coming soon.