Closed MingBit closed 4 years ago
Hi, thanks for trying STREAM. Sorry about the lack of documentation for the attributes.
Re your question, branch_id_alias
and S0_pseudotime
should suffice in terms of selecting the cells needed.
branch_lam
and branch_id
are both the 'raw' attributes during the structure learning, which are the lamda
, i.e. the arc-length from the beginning of the curve, and the id
internally used to label branches, respectively. They are preserved just for the purpose of debugging. You most likely won't need them.
Anyhow, thanks for the feedback. We will keep improving it. Hope the above is helpful to you.
Yeah! it's really helpful. Thanks a lot! Does it mean that the pseudotime at terminal point is sufficient for ordering? For instance, S2_pseudotime can be used for ordering cells from pathway (S3->S1->S2)?
S2_pseudotime
means the ordering of all cells when choosing 'S2' as the starting node.
That being said, to order cells along the trajectory (S3->S1->S2), you should consider S3_pseudotime
. Also you might want to only keep the cells assigned to that path by choosing the cells from branch ('S3','S1') and ('S1','S2')
Score! Thank you so much! :+1:
Hallo,
Thanks for the great tool! I'm still new to STREAM and doing the first analysis according to this tutorial. Everything works properly so far. Just a little bit confused about the results stored in adata object.
I'm wondering where I could find the documentation for the new attributes (e.g. branch_id, branch_lam, S0_pseudotime) stored in adata object. For instance, I'd like to select cells from one branch (S3 -> S0) and order cell by pseudotime. However, I'm not sure which one should I use?
looks forward to your reply. Thank you.