cellink bridges donor genotypes and single-cell profiles for end-to-end single-cell genetics

cellink brings single-cell data together with the genomes of the donors the cells come from. It reads genotypes from VCF, PLINK, BGEN and PGEN files, keeps them paired with the cells of the same donors through every selection, and runs genetic analyses on both. Established tools such as PLINK, tensorQTL, SAIGE-QTL, LDSC and MAGMA run directly on your data, and cellink takes care of the file formats and conventions each of them expects.

Data

cellink works on DonorData objects, which link the cells of a cohort to their donors. In cellink the donor level dd.D holds the genotypes, with one row per donor and one column per variant, and dd.C holds the cells as an AnnData or MuData. Selecting donors, variants or cells keeps both levels aligned, so genotypes and expression always describe the same people.

import cellink as cl

genotypes = cl.io.read_plink("cohort")
dd = cl.DonorData(D=genotypes, C=adata, donor_id="individual")
dd = dd[..., dd.C.obs["cell_type"] == "CD4 T", :]
dd.write_dd("cohort.dd.zarr")

DonorData is imported from the DonorData package and is installed with cellink. The DonorData documentation explains how these objects are built and stored, and how to add a sample level or the electronic health records of the donors.

Analyses

Association testing

• eQTL mapping on pseudobulk with jaxQTL or tensorQTL
• eQTL mapping on single cells with SAIGE-QTL
• A simple pseudobulk eQTL mapping
• Rare variant association tests
• Clumping and pruning
• Colocalization
• Writing association tests as formula strings

Heritability, gene programs and GWAS integration

• LD score regression for cell types
• Gene program scoring with sc-linker
• Gene set enrichment with MAGMA
• Disease relevance of single cells from GWAS with scDRS and seismic
• Cell neighborhoods associated with disease with scEPS
• GWAS on spatial data with gsMap

Deep learning

• Learning donor representations with LIVI
• Predicting single cell coverage from sequence and scoring variants with scooby
• Dataloaders for deep learning, including multiple instance learning over the cells of each donor

Variants

• Quality control, annotation and selection of genetic variants

Getting started

cellink is not on PyPI yet and will be published there as cellink-tools. Install the development version from GitHub, which also installs DonorData:

pip install git+https://github.com/theislab/cellink.git@main

The installation guide lists the optional extras for the external tools. If you are new to cellink, start with DonorData basics, which builds an object from genotype and expression data, subsets it and saves it without running any analysis.

Documentation

Installation guide	Requirements and the optional extras (pip install cellink[extra-name])
Tutorials	Step by step guides for the analyses above
API reference	Every public function and class, organized by module
DonorData documentation	The DonorData object, its sample level and its file format
Contributing guide	Development setup, tests and the pull request workflow
Changelog	What changed in each release

Contact

If you found a bug, please use the issue tracker.

Release notes

t.b.a

Citation

t.b.a