jkmorrison
commented
3 years ago Hi Jiarui,
My apologies for not getting back to you sooner! CellPhoneDB does allow you to use gene name or HGNC symbol annotation as input data for the metadata file, so you shouldn't need to convert to Ensembl ID before running the analysis.
In regards to your other questions, we haven't tried other packages for ligand-receptor analysis yet, but it is certainly on our to do list!
Best, Josh
Hi, Joshua,
Noticed your really nice work and follow your code in recent days. I tried to search for your email address but didn't manage to get that. So I have to drop my question here.
I had problem with cellphoneDB since the first row of the counts data is the gene symbol, however, cellphoneDB only allows the ensembl gene id.
I tried to use biomart package (getBM function) and clusterProfiler package (bitr function) to do id conversion with the aim to get the ensembl gene id from gene symbol (the zebrafish ortholog converted human gene symbol). However, there are multiple ensembl gene ids mapping to the same gene(gene symbol).
I am wondering have you ever encountered such a problem and do you have any recommendations to solve that.
BTW, we are studying zebrafish pancreas so basically I did the similar zebrafish-to-human ortholog conversion step before running ligand-receptor interaction.
Final two questions, for ligand-receptor analysis, have you ever tried some other packages, such as NicheNet? and what about motif analysis (it is a shame that zebrafish is not included in the SCENIC but do you know if there are any other tools available?)
best Jiarui