search

ShobiStassen / PARC

MIT License
41 stars 11 forks source link

any plan to release an R version of PARC? #10

Open cirobotta opened 4 years ago

cirobotta commented 4 years ago

Very interesting work, I was thinking to include this method in my cytometry pipeline (it will be released soon). Are you planning a release fo R?

ShobiStassen commented 4 years ago

hi there, I didn't have any plans just yet but will try to spend some time next month to see if I can make a release for R. I'll get back to you once I have something running! Also, since you mentioned you are working on a cytometry pipeline, I imagine you might have some large datasets. I recently added a feature which uses the HNSW knn graph as a basis for a umap visualization, if you are working with datasets that are in the several 100,000s or millions of cell count, you could consider trying it (the Jupyter NB - which is still a wip) has some details of this. For smaller / mid size datasets the regular umap works as fast.

cirobotta commented 4 years ago

Thank you very much for your answer...At the moment I included PARC in my R pipeline by using the package reticulate.....I should say that it is extremely fast and works really well!! Thank you! Waiting for an R version to include it more easily... Thank you for suggestion regarding the dimensionality reduction approach, I'll give it a look...As soon as the pipeline will be officially released I'll inform you!

ShobiStassen commented 4 years ago

Hi, it has just come to my attention that the latest version of leidenalg is considerably slower than the previous version. For the time being, to ensure you are getting the best speedup possible, please can you double check you have version 0.7.0 of leiden alg. The updated 0.8.0 leidenalg came out this April 2020. pip install leidenalg ==0.7.0 should work

toobiwankenobi commented 4 years ago

Hi there, any news on an R version of PARC?