Moving away from Seurat/Signac
I have been using Seurat / Signac for 10x data analysis for a long time but the ecosystem of related libraries seems to be getting fewer updates, running into build errors and is missing a several modern features.
I am looking to move away from that pipeline but I'm not really able to spend a while benchmarking various packages in R / Python.
Can people recommend more up to date software libraries / pipelines they typically rely on or point me to a good up-to-date guide? Many of the single-cell tutorials I have found are 2+ years old.
• 259 views
•
link
written
by
yura.grabovska •
0 answers
No answers yet.
Log in to answer this question.
More posts like this
-
After doublet detection in 10x scRNA-Seq data, there is still evidence of doublets
written by yura.grabovska •
-
Seurat v5 and how to correctly integrate across multiple experiments
written by yura.grabovska •
-
Converting scMultiome data to loom using SEURAT
written by halimaakhter014 •
-
How do I resolve an integration error in Seurat?
written by pcitrano3 •
-
GeneActivity without Fragments file in Seurat for Integrating scRNA-seq and scATAC-seq
written by el24 •
-
Job: Postdoctoral research fellow in bioinformatics of eDNA
written by salanova.elliott •
-
Herald: The Biostar Herald for Tuesday, July 20, 2021
written by Biostar
-
Job: Where is a good place to find experienced, ongoing, bioinformatics help?
written by twizzlo •
-
Not having root access sucks; installing software without root privileges
written by Drio
-
What Kind Of Bioinformatics Tutorials Would You Like To See Online?
written by User 59
Tutorials are "old" because the single-cell hype dampens, and it is a standard thing to do these days. Basics have been written down in tutorials, so there is no need for monthly updates. It comes down to either use the SCverse/ScanPy ecosystem. Or R/Bioconductor. I use the latter, never needed anything else, and if these is some nice SCverse stuff I use reticulate.
At least in my field, the majority of big papers have moved away from graphical clustering and into approaches like NMF-based methods tailored towards sparse data which isn't really something that is readily implemented in R to the same extent. Also, at least in my experience R/Bioconductor libraries tend to struggle with larger datasets, and especially with the Seurat/Signac-compatible methods the memory overhead is a big problem.