Has anyone looked at CINS (https://doi.org/10.1371/journal.pcbi.1010468)?
Learns a Bayesian network of cell type dependencies from changes in cell type proportions (#scRNAseq) in case-control studies. Causal ligands from cell-cell edges are predicted by LASSO regression of #ligand and predicted response genes (using NicheNet's ligand-reponse network).
Cool idea, weird lack of validation in their paper... But I guess that's the challenge in #CellCellInteraction prediction - validation is _hard_
CINS: Cell Interaction Network inference from Single cell expression data
Author summary Single cell transcriptomics has emerged as a leading technology for studying the composition of organs and tissues in the human body, development and several other biological processes. More recent studies, including studies of various diseases (such as cancer), treatment—response studies and aging studies aim at comparing samples at the single cell level. To date, such analysis mainly focused on the differences in expression of genes in the different cell types. However, in addition to differences in expression such studies also provide information on the differences of cell type proportions between the conditions. To use such information for inferring cell interactions we developed a new computational framework termed CINS. CINS combines Bayesian network learning (which is used to infer cell type–cell type interactions) with constrained regression analysis (used to infer the specific proteins involved in such interactions). We applied CINS to a number of different case-control scRNA-Seq datasets including a lung disease and an aging study. By analyzing public data and profiling new scRNA-Seq data we show that CINS is able to correctly identify several proteins as playing a significant role in interactions controlling disease progression and aging, improving on prior methods suggested for this task.
Finally read the REMI paper by Alice Yu (aliceomics@twitter) et al. Calculating the partial correlation structure of #ligand #receptor interactions across #scRNAseq samples (cancer, in their case) to _more_specifically_ identify context-dependent interactions. #CellCellInteraction prediction has a specificity problem, and this method outperforms NicheNet, which uses predicted transcriptional response to improve specificity of predictions.
https://twitter.com/PlevritisLab/status/1509196799520620544?s=20&t=6jB2saz7rcO2u-4Gh54C4Q
Stanford Plevritis Lab on Twitter
“Cellular crosstalk is fundamental to tissue biology. We are very excited to share our novel computational method - REMI - for reconstructing cell-cell communication from omics data! Read the article here: https://t.co/vyIlbOSRer”
https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1010715
I reviewed this paper, and was pretty excited about it. I don't do any #spatial #scRNAseq, but the tools they've developed for #CellCellInteraction inference in C. elegans (stereotyped cellular loactions + published scRNAseq atlas = genius test-bed for spatial inference methods) are powerful and neato.
Author's tweet: https://twitter.com/eagut/status/1593334779134386176?s=20&t=abTEVMJ2ngD0IXF5RC2Bpg
Inferring a spatial code of cell-cell interactions across a whole animal body
Author summary Neighboring cells coordinate gene expression through cell-cell interactions, enabling proper functioning in multicellular organisms. Hence, intercellular interactions can be inferred from gene expression. We use this strategy to define a molecular code bearing spatial information of cell-cell interactions across a whole animal body. We develop a computational framework to infer the first cell-cell interaction network in Caenorhabditis elegans from its single-cell transcriptome, and show a negative correlation between interactions and intercellular distances, which is driven by a combination of ligand-receptor pairs following spatial patterns across the C. elegans’ body, i.e., the spatial code. Thus, our framework uncovers molecular features crucial to defining spatial cell-cell interactions across a whole body; a strategy that can be readily applied in higher organisms.
Brendan Innes
Malte Petersen