asrhou/natmi repository overview

⁠NATMI: Network Analysis Toolkit for the Multicellular Interactions

Recent development of high throughput single-cell sequencing technologies has made it cost-effective to profile thousands of cells from a complex sample. Examining ligand and receptor expression patterns in the cell types identified from these datasets allows prediction of cell-to-cell communication at the level of niches, tissues and organism-wide. Here, we developed NATMI (Network Analysis Toolkit for Multicellular Interactions), a Python-based toolkit for multi-cellular communication network construction and network analysis of multispecies single-cell and bulk gene expression and proteomic data.

How to cite:

1. Hou, R., Denisenko, E., Ong, H.T. et al. Predicting cell-to-cell communication networks using NATMI. Nat Commun 11, 5011 (2020). https://doi.org/10.1038/s41467-020-18873-z⁠
2. Ramilowski, J., Goldberg, T., Harshbarger, J. et al. A draft network of ligand–receptor-mediated multicellular signalling in human. Nat Commun 6, 7866 (2015). https://doi.org/10.1038/ncomms8866⁠

Contact: Rui Hou [[email protected]⁠]

More information: https://github.com/forrest-lab/NATMI⁠

⁠Pulling Docker Image

Once docker is installed, the next step is to pull the NATMI image from docker hub using the following command:

docker pull asrhou/natmi

⁠Running image

Predict ligand-receptor-mediated interactions in a mouse single-cell RNA-seq dataset using literature supported ligand-receptor pairs and four CPUs:

   docker run --rm -it --name natmi -v /home/path/workdir/:/opt/NATMI/workdir asrhou/natmi python ExtractEdges.py --species mouse --emFile toy.sc.em.txt --annFile toy.sc.ann.txt --interDB lrc2p --coreNum 4 --out /opt/NATMI/workdir/toy.sc

Predict ligand-receptor-mediated interactions in a human bulk RNA-seq dataset using putative and literature supported ligand-receptor pairs and one CPU:

   docker run --rm -it --name natmi -v /home/path/workdir/:/opt/NATMI/workdir asrhou/natmi python ExtractEdges.py --species human --emFile toy.bulk.em.xls --interDB lrc2a --out /opt/NATMI/workdir/result/folder

Detect changes in edge weight in two output folders (generated by ExtractEdges.py) using literature supported ligand-receptor pairs:

   docker run --rm -it --name natmi -v /home/path/workdir/:/opt/NATMI/workdir asrhou/natmi python DiffEdges.py --refFolder /opt/NATMI/workdir/reference/dataset --targetFolder /opt/NATMI/workdir/target/dataset --interDB lrc2p --out /opt/NATMI/workdir/delta.folder

Visualise cell-connectivity-summary networks from the results of ExtractEdges.py or DiffEdges.py:

   docker run --rm -it --name natmi -v /home/path/workdir/:/opt/NATMI/workdir asrhou/natmi python VisInteractions.py --sourceFolder /opt/NATMI/workdir/result/folder --interDB lrc2p --weightType mean --detectionThreshold 0.2 --plotFormat pdf --drawNetwork y --plotWidth 12 --plotHeight 10 --layout kk --fontSize 8 --edgeWidth 0 --maxClusterSize 0 --clusterDistance 1

Visualise cell-to-cell communication networks between all possible pairs of cell types using results of ExtractEdges.py or DiffEdges.py:

   docker run --rm -it --name natmi -v /home/path/workdir/:/opt/NATMI/workdir asrhou/natmi python VisInteractions.py --sourceFolder /opt/NATMI/workdir/result/folder --drawClusterPair y --keepTopEdge 10

Visualise cell-to-cell communication networks via a ligand-receptor pair from the results of ExtractEdges.py:

   docker run --rm -it --name natmi -v /home/path/workdir/:/opt/NATMI/workdir asrhou/natmi python VisInteractions.py --sourceFolder /opt/NATMI/workdir/result/folder --interDB lrc2p --drawLRNetwork LIGAND.SYMBOL RECEPTOR.SYMBOL