question about parameters

[wbvguo]

Hi,

Thanks for maintaining this tool. Choosing a threshold is always subjective. May I ask if CellBender provides an automatical way to specify the parameters for --expected-cells and --total-droplets-included? if not, would this be in the future plan?

Thanks,

[sjfleming]

Hi Wenbin,

Yes, CellBender does provide an automatic way to set those parameters, although it is not that great as of version 0.2.1, I will admit. It does fine on "easy" datasets, but there are a lot of datasets out there that are not so easy.

Right now, you can simply run the command without including the arguments --expected-cells or --total-droplets-included at all. They are optional. If you do not provide them, CellBender will make its best guess.

I am actively working to make these guesses better for version 0.3.0, and I think I have made a lot of progress. If you want to try the very latest version and see if it works better, you can try to install from the branch called sf_dev_0.3.0_postreg. Or if you prefer, you can wait for the official v0.3.0 release. That will be out as soon as #189 gets merged.

For advice on how to set the values of --expected-cells or --total-droplets-included yourself, see the notes here:
https://cellbender.readthedocs.io/en/latest/usage/index.html#recommended-best-practices

Thanks,
Stephen

[sjfleming]

Just now released v0.3.0 which should do a much better job estimating those parameters for you. Now you really should not need to include the input arguments --expected-cells or --total-droplets-included at all.

Closed by #238