Skip to content

Code and analyses supporting the GeNA single-cell csaQTL manuscript.

License

Notifications You must be signed in to change notification settings

immunogenomics/GeNA-applied

Repository files navigation

In this repository

We provide code and analyses supporting the GeNA manuscript. We provide scripts to:

  • Apply GeNA to real single-cell profiling and simulated genotypes to evaluate GeNA's calibration (null folder) and statistical power (nonnull_sims folder)
  • Apply GeNA to identify cell state abundance QTLs (csaQTLs) in the OneK1K dataset (run_gwas, leadsnps_perm, suggestive_loci, retest_subcohorts, molecularQTLs folders)
  • Test associations to each lead SNP in single-cell objects with cis-genes removed (mask_cis_genes folder) or suggestive trans-eGenes removed (mask_trans_eGenes folder)
  • Perform GWAS of cluster-based cell type proportion traits for comparison (cluster_gwas folder)
  • Evaluate the replication of csaQTLs from the OneK1K discovery cohort in five replication cohorts (replication folder)
  • Evaluate the replication of csaQTLs previously identified using flow cytometry in our neighborhood-based framework for single-cell data (published_csaQTLs folder)
  • Examine cell state abundance associations to polygenic risk scores (prs)
  • Evaluate the sensitivity of our results to various aspects of the primary analysis (ccg_retained, k_sensitivity, conditional_testing folders)
  • Apply GeNA to a dataset of cells in early neural differentiation (neural_dset folder)

We also provide the notebooks used to generate figures and key reported values.

Citation

The GeNA manuscript can be found and cited at [link to preprint]

To use GeNA

Please refer to the GeNA repository at immnogenomics/GeNA

Data availability

All datasets used in these analyses are previously published:

  1. Yazar, S. et al. Single-cell eQTL mapping identifies cell type–specific genetic control of autoimmune disease. Science 376, eabf3041 (2022).
  2. Perez, R. K. et al. Single-cell RNA-seq reveals cell type-specific molecular and genetic associations to lupus. Science (American Association for the Advancement of Science) 376, eabf1970–eabf1970 (2022).
  3. Oelen, R. et al. Single-cell RNA-sequencing of peripheral blood mononuclear cells reveals widespread, context-specific gene expression regulation upon pathogenic exposure. Nat Commun 13, 3267 (2022).
  4. Randolph, H. E. et al. Genetic ancestry effects on the response to viral infection are pervasive but cell type specific. Science 374, 1127–1133 (2021).
  5. Jerber, J. et al. Population-scale single-cell RNA-seq profiling across dopaminergic neuron differentiation. Nat. Genet. 53, 304–312 (2021).

Contact

Please contact Laurie Rumker (Laurie_Rumker AT hms.harvard.edu) with any questions about these analyses.

About

Code and analyses supporting the GeNA single-cell csaQTL manuscript.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages