There are many tools and resources to estimate Cell-Cell Communication from scRNA data - do you wonder which one you should use? We do too! To find out, we compared 6 methods and 15 resources doi.org/10.1101/2021.0….
For this, we built a framework that decouples Cell-Cell communication tools from their inbuilt resources. The framework is open access and available at github.com/saezlab/ligrec….
We analyzed the ‘lineage’ of the different resources that shows how they are all interconnected, but individual resources had varying proportions of the collective CCC prior knowledge.
We saw that both the resource and the method can have a considerable impact on the results. It is however difficult to know what works best given the lack of ground truths to benchmark (see Supp. Note 4). Do you have any benchmark suggestions? 🙂
We think these disagreements are a word of caution, as the choice of method can largely affect the interpretation of the data.
Did the antibody stop working or you cannot measure a node? We can predict it’s activity from other measured nodes. Predictions correlate strongly with test data and are accurate across time and conditions. Watch out for rare signaling patterns though.
We are happy to share our first contribution to the #cardiology field. @roramirezf94 & Jan Lanzer, et. al. present: “A Consensus Transcriptional Landscape of Human End-Stage Heart Failure” (medrxiv.org/content/10.110…) (1/7)
In our pre-print we compared and integrated 16 public transcriptomic studies of the last 15 years that profiled over 900 left-ventricle biopsies of #heartfailure (HF) patients and control individuals (2/7)
We demonstrated that single studies reported highly dissimilar gene expression markers (A), however, a common disease signature became apparent with the use of transfer learning approaches (C) (3/7)