search

Teichlab / tracer

TraCeR - reconstruction of T cell receptor sequences from single-cell RNAseq data
Other
122 stars 48 forks source link

Tracer clonotype network output #91

Closed Meng0625 closed 4 years ago

Meng0625 commented 5 years ago

Hi there,

When I run the summarise process in TraCer, it's stuck in the process of dot drawing for very long time(more than 1 day) and I have to kill it in the end. I have check the detail of the drawing, and it seems like the dot cannot handle a large picture(about 900 cells). Could anyone give some tips about this issue?

And the dot files are here. dot.zip

Many thanks

mstubb commented 5 years ago

Hi,

Yes, Dot is not good at laying out large networks (depending upon how many of your cells are connected).

I'd suggest running summarise with the --no_networks option to stop it trying to draw the networks. It will then complete without needing to be killed and you can find alternative ways to interpret your data.

It's worth noting that large network graphs with huge amounts of connectivity are very hard to interpret and so of relatively little value.

Best,

Mike

VicenteFR commented 4 years ago

Hi,

Regarding this issue... Just wondering if you can give a rough approximation of the maximum amount of cells dot can deal with in order to create the graph.

Many thanks

mstubb commented 4 years ago

We haven't systematically tested this. My sense is that it should be fine with hundreds of cells but will start to struggle beyond a few thousand. It'll depend a bit upon the amount of connectivity in your graph, the compute resources you have and your level of patience.