Is there a limitation to the number of cells swappedDrops can handle?
2
0
Entering edit mode
@josephjoonsang-20642
Last seen 5.6 years ago

Hi,

I have used swappedDrops function to remove potential index hopping UMIs from 10x genomics single-cell data. Recently, we sequenced +100K single cells. When I was processing the data, I found that swappedDrops crashed. I am wondering if there is a limitation to the number of cells swappedDrops can handle. Your inputs/insights will be much appreciated. Thank you.

Best, Joon

swappedDrops DropletUtils • 1.2k views
ADD COMMENT
1
Entering edit mode

Hi Joon,

Can you provide some more information about how you ran swappedDrops (e.g., locally or on a compute cluster), and whether there was an error message that was reported?

Loading in very many unfiltered matrices from CellRanger should be quite tough on your RAM. Are you running in an environment with sufficient memory, even before swappedDrops is called?

Jonny

ADD REPLY
0
Entering edit mode
Aaron Lun ★ 28k
@alun
Last seen 16 minutes ago
The city by the bay

As Jonny said, more information is necessary to provide meaningful advice. In the absence of such information, I'll just throw out a couple of scenarios for you to ponder.

  • You ran out of memory and your cluster killed the job. Solution: allocate more memory. Expect swappedDrops to use up about 5 GB of RAM per sample (I don't remember the exact numbers).
  • You have more than .Machine$integer.max molecules in one of your files. Solution: switch to DropletUtils 1.3.* or higher, which avoids this limitation (up until 2^64 molecules, at least).
  • You get a segfault due to a stack overflow. Solution: increase the size of your stack, as discussed here.
  • Something else happened, in which case you should show the commands you used, the error message, sessionInfo(), etc. A reproducible example would be great but I doubt you will be able to provide it.
ADD COMMENT
0
Entering edit mode
@josephjoonsang-20642
Last seen 5.6 years ago

Jonny and Aaron,

Thank you. I will rerun swappedDrops following your suggestions. If I fail, I will get back to you with error messages.

In this analysis, we sequenced ~40 10x GEM samples (~100K cells in total) three times by using NovaSeq. We included all samples in each run.

Joon

ADD COMMENT
0
Entering edit mode

For future reference, use "add comment" to an existing answer to post updates like this. You should only use "add answer" if you're answering your own question.

ADD REPLY

Login before adding your answer.

Traffic: 696 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6