Genetic mapping of developmental trajectories for complex traits and diseases

This repository contains scripts, explanations, and examples for our pipeline for genetic mapping of developmental trajectories for complex traits and diseases. The pipeline is based on integrative analysis of Genome-Wide Association Studies (GWAS) and single-cell RNA-seq (scRNA-seq). The analysis performs the following three main tasks:

1. Identification of connections between developmental trajectories and traits.
2. Elucidate molecular pathways that underlie the link between the trajectory and trait
3. Prioritize genes that carry the link between the pathway, trait and trajectory

Prerequisites

• MAGMA
• Download our R functions here.
• Make sure the following packages are Installed: Monocole (v2), plyr, ggplot2, data.table, parallel, speedglm, Seurat, clusterProfiler (a Bioconductor package).
• Some input files are also required. For details and sample files, see individual vignettes below.

Examples for usage

We provide vignettes for the analysis using scRNA-seq dataset of pancreatic islet development (Byrnes et al.), and a GWAS dataset of type 2 diabetes (Mahajan et al.).

See the following link for a vignette that covers all the steps.

Alternatively, the following details the analysis steps and provides links to the vignettes:

• Step 1: Identification of connections between developmental trajectories and traits.
  • 1a. Converting GWAS variant scores into gene-trait association scores. This is performed using MAGMA gene analysis. We provide example output files for the type 2 diabetes dataset. Refer to MAGMA’s website and manual for a detailed explanation.
  • 1b. Calculating cell-trait association scores. See our vignette.
  • 1c. Trajectory inference. This step is performed using tools such as Monocle 2, Monocle 3, destiny. In principle, any tool that gives quantitative maturation scores for cells, e.g., pseudotime, is suitable. See Monocle’s version 2, and version 3 websites for thorough explanations. For the pancreatic development dataset used here, the analysis codes were published by the authors and are available for download here. Also, the output, Monocle object, is available here.
  • 1d. Examining the association between trait and trajectory. See our vignette.
• Step 2: Elucidate molecular pathways that underlie the link between the trajectory and trait.
  • 2a. Finding pathways enriched in the trajectory. This is covered here.
  • 2b. Examining the trajectory-enriched pathways from 2a for trait association. This is covered here.
• Step 3: Prioritize genes that carry the link between the pathway, trait and trajectory. See our vignette here.

The following flowchart summarizes the analysis steps:

Authors

• Eldad David Shulman

• Prof. Ran Elkon

License

This project is licensed under the BSD 3 License - see the LICENSE.md file for details.