James S. Nagai¹, Nils B. Leimkühler², Michael T. Schaub ³, Rebekka K. Schneider^4,5,6, Ivan G. Costa^1*

¹Institute for Computational Genomics, Faculty of Medicine, RWTH Aachen University, Aachen, 52074 Germany

²Department of Hematology and Stem Cell Transplantation, University Hospital Essen, Germany

³Department of Computer Science, RWTH Aachen University, Germany

⁴Department of Cell Biology, Institute for Biomedical Engineering, Faculty of Medicine,RWTH Aachen University, Pauwelsstrasse 30, 52074 Aachen, NRW, Germany

⁵Oncode Institute, Erasmus Medical Center, Rotterdam, 3015GD, the Netherlands

⁶Department of Hematology, Erasmus Medical Center, Rotterdam, 3015GD, the Netherlands

Motivation: Ligand-receptor (LR) analysis allows the characterization of cellular crosstalk from single cell RNA-seq data. However, current LR methods provide limited approaches for prioritization of cell types, ligands or receptors or characterizing changes in crosstalk between two biological conditions.

Results: CrossTalkeR is a framework for network analysis and visualisation of LR networks. CrossTalkeR identifies relevant ligands, receptors and cell types contributing to changes in cell communication when contrasting two biological states: disease vs. homeostasis. A case study on scRNA-seq of human myeloproliferative neoplasms reinforces the strengths of CrossTalkeR for characterisation of changes in cellular crosstalk in disease state.

Install

You can install CrossTalkeR with the simple commands below:

install.packages("devtools")
devtools::install_github("https://github.com/CostaLab/CrossTalkeR", build_vignettes = TRUE)
require(CrossTalkeR)

Note: Please avoid to use the following characters in celltype name: ‘$’

Possible system dependencies

libudunits2-dev
libgdal-dev
gdal-bin
libproj-dev
proj-data
proj-bin
libgeos-dev

CrossTalkeR Plots examples and vignette

We provide in our vignette examples on how to analyse cell interactions from a human myelofibrosis single cell RNA-seq.

vignette('CrossTalkeR-HumanMyfib')

Integration of liana with CrossTalkeR

We provide here a tutorial on how to integrate predictions from liana into Crosstalker.

Run liana

New Features v1.4.0

Splitted generate_report function in two parts:
- analise_LR() to only run the analysis without generating the CrossTalkeR report
- make_report() to only generate a new CrossTalkeR report for existing CrossTalkeR results
Added node types to the network:
- we now consider the annotation of a gene as ligand (L) or receptor (R) to consider the biological function
Less constrains on the cell cluster name annotation (only ‘$’ must be avoided in the cluster naming)
Integration with liana-py for ligand-receptor interaction predictions

Old Features v1.3.0

Single and Comparative Reports
- Cell Cell Interaction visualization
- Sending and Receiving Cells Ranking
- Gene Target based Sankey Plots
- CCI and GCI PCA ranking
  - All measures and PC table
  - PC1 and PC2 based barplot
- Leimkühler et. al. [2] data were added to the package
- Fisher Test were implemented to highlight the CCI edges significance (new) 🔥NEW🔥
- Change input format: Please see the Documentation 🔥NEW🔥
  - A python3 notebook are available to cast the old input to the new input.
- Liana (Dimitrov et. al. [3]) Output can be used as CrossTalkeR input. 🔥NEW🔥
- LR pair visualization plot can be done using a Seurat Object 🔥NEW🔥

References

[1] CrossTalkeR: Analysis and Visualisation of Ligand Receptor Networks link

[2] Heterogeneous bone-marrow stromal progenitors drive myelofibrosis via a druggable alarmin axis. link

[3] Comparison of Resources and Methods to infer Cell-Cell Communication from Single-cell RNA Data link