search

RGLab / flowStats

flowStats: algorithms for flow cytometry data analysis using BioConductor tools
15 stars 10 forks source link

Help with normalization using FlowStat #37

Open f0ng1203 opened 2 years ago

f0ng1203 commented 2 years ago

Hi there, my name is Yeh Fong and currently I am facing some problem using the normalization with gaussNorm.

I am using the vignette in FlowSOM package. Attached is the link: https://rdrr.io/github/saeyslab/FlowSOM_workshop/f/vignettes/FlowSOMWorkshop.Rmd

When I tried to execute #gaussNorm normalization, I have an error

"No curve with 2 landmarks was found for channel APC-Cy7-A . Decrease max.lms[ APC-Cy7-A ] and try again. Error in extract.base.landmarks(lms$filter, channel.names, max.lms) :"

I saw elsewhere it can be solved when we transformed it with the following code

trans <- estimateLogicle(fs_batch[[1]], channels = test.chnl) fs_batch.trans <- transform(fs_batch, trans) fs_scaled1 <- gaussNorm(fs_batch.trans, channel.names=test.chnl)

However, when I tried that code I still face a similar problem.

fs_gaussNorm <- gaussNorm(flowset = fs_batch.trans,

  • channel.names = channels_of_interest,
  • peak.density.thr=0.01) No curve with 2 landmarks was found for channel APC-Cy7-A . Decrease max.lms[ APC-Cy7-A ] and try again. Error in extract.base.landmarks(lms$filter, channel.names, max.lms) :

Could you kindly advise on how I should approach this problem?

Thank you in advance.

mikejiang commented 2 years ago

gaussNorm normalization is no longer part of our normal workflow for data analysis, therefore it's not actively maintained and supported. For automated cell discovery and clustering,FAUST proved to be more accurate and effective. https://www.cell.com/patterns/fulltext/S2666-3899(21)00234-8 Feel free to post to your questions there https://github.com/RGLab/FAUST if you decide to give it a try