Yeah that's how it ended up for me: large datasets get churned through for speed in Python, but I then usually switch over to R with the summary data because there's just way more biology-specific ecosystem in R than in Python.

R/Bioconductor has packages for human genome-specific analyses so it's easy to download gene positions etc., there are packages for read simulation, amplicon sequence variant detection, gene distance simulations, any kind of RNAseq analysis you can think of... none of these packages exist in Python. If you'd rerun it in Python you'd save 10 minutes or hours of running time but you'd lose days or months re-implementing analyses that are in R packages (plus those R packages often call on C++ code, anyway)

plus ggplot2 is miles ahead of any plotting in Python (to me :) ).

Yes. But the big problem is that all the things that make Python Python is also the things that make it slow. People have tried again and again to make a fast Python, and failed. And from my first impressions of Mojo, it's not very much like Python at all.

That seems to be exactly what Mojo is/wants to be. At least that's how I understand their landing page: https://www.modular.com/max/mojo