ParkerICI / premessa

R package for pre-processing of mass and flow cytometry data
GNU General Public License v3.0
42 stars 24 forks source link

Different Debarcode Event Yields using Zunder Lab, Catalyst & Premessa Debarcoders #42

Open JahangirSufi opened 2 years ago

JahangirSufi commented 2 years ago

Hi,

I hope you're well.

As per the subject, I've been testing out the different debarcoders:

Using the exact same normalised dataset, I've been getting different debarcoded event yields when comparing Catalyst and Premessa to the ZLD (which is my 'control' (As I've been using this for the last couple of years)), even though the % assigned is near enough the same between the three debarcoders.

As an example, for this one experiment I'm analysing (using the exact same debarcoding settings of 0.1 BCS and 10 MD), I'm getting a debarcoded event yield of around 49% for each debarcoder (which is normal as I'm barcoding organoids in situ using the TOBis 7-c-3 barcoding system - if you want to read more, see here). But the ungated event counts I'm getting differs greatly such that the Catalyst Debarcoder results in around 15% fewer events than the ZLD and the Premessa debarcoder results in a whopping 56% fewer events than the ZLD. Please see screenshots below.

Is there a reason for this? I read that you use an R-implementation of the Nolan Lab Debarcoder (which I think it's quite similar to the ZLD, but surely there shouldn't be a 56% discrepancy between your debarcoder and t the ZLD.

Any explanation/clarification would be extremely helpful.

Many thanks,

Jahangir Sufi

Screenshot 2022-05-09 at 16 05 36 Screenshot 2022-05-09 at 16 05 40