CelLink: integrating single-cell multi-omics data with weak feature linkage and imbalanced cell populations
CelLink is a Python package designed for single-cell multi-omics integration. It excels uniquely in integrating datasets with weak feature linkage and imbalanced cell populations. CelLink normalizes and smooths feature profiles to align scales across datasets and integrates them through a multi-phase pipeline that iteratively employs the optimal transport algorithm. It dynamically refines cell-cell correspondences, identifying and excluding cells that cannot be reliably matched, thus avoiding performance degradation caused by erroneous imputations. A classic example of weak linkage is seen in the integration of scRNA-seq and CODEX (spatial proteomic data) from the Human Pancreas Analysis Program (HPAP).
CelLink uniquely enables cell subtype annotation, correction of mislabelled cells, and spatial transcriptomic analyses by imputing transcriptomic profiles for spatial proteomics data. Its great ability to impute large-scale paired single-cell multi-omics profiles positions it as a pivotal tool for building single-cell multi-modal foundation models.
Cellink can be installed from PyPI using pip. For best practices, create a new virtual environment before installation. Below, we demonstrate how to set up this environment using conda.
conda create -n CelLink python=3.9
conda activate CelLink
pip install cellink-luoA tutorial on integrating scRNA-seq and CODEX datasets from donor HPAP023 is provided. The feature linkage information between coding genes and proteins is stored in protein_gene_relationship.csv. Please check our tutorial website.
If you use Cellink in your research, please kindly cite our paper using the following reference: Xin Luo et al. CelLink: Integrate single-cell multi-omics data with few linked features and imbalanced cell populations, Biorxiv 2024.
This project is licensed under the GNU General Public License v3.0. For more information, see the LICENSE file in this repository.