Page 1 of 1

2020-Trussart et al-preprint

PostPosted: Wed May 13, 2020 5:00 pm
by mleipold
"CytofRUV: Removing unwanted variation to integrate multiple CyTOF datasets"
Marie Trussart, Charis E Teh, Tania Tan, Lawrence Leong, Daniel HD Gray, Terence P. Speed
bioRxiv, posted May 09, 2020
https://doi.org/10.1101/2020.05.09.085621

- will update when/if this is peer-reviewed and published

- compares to CytoNorm and BatchAdjust

- "The fcs files from this study will be available at flow repository ID FR-FCM-Z2L2." "Sorry, you do not have access to the specified experiment."

- also reuses datasets
1. 2019-Schuyler et al-Front Immunol
2. FR‐FCM‐Z247 (2019-Van Gassen et al-Cytometry A)

Re: 2020-Trussart et al-preprint

PostPosted: Wed May 27, 2020 8:09 pm
by narulam
I tried this normalizing algorithm for raw fcs files. It works well for batch effect correction but markers lose resolution between negative and positive expression due to normalized protein distribution which makes it very difficult to manually gate subpopulations (for eg. Tregs and CD4+ Th subsets) and identify separated islands with tsne or downstream clustering differences. Any thoughts on that?

Thanks

Re: 2020-Trussart et al-preprint

PostPosted: Wed May 27, 2020 8:14 pm
by dtelad11
Any chance you could post before/after biaxial plots?

Re: 2020-Trussart et al-preprint

PostPosted: Sun May 31, 2020 11:46 pm
by narulam
Hi El-ad,

I've attached some slides to show the gating and resolution issue for some markers. Since the scaling from pre-CytofRUV files does not apply to the files post normalization, I tried to tweak the scaling on normalized fcs files and was able to improve the resolution for some of them (Th1,Th2,Th17) but not all (Treg,Tfh). Please let me know your thoughts.

Thanks,
Mansi

Re: 2020-Trussart et al-preprint

PostPosted: Mon Jun 01, 2020 12:05 am
by mleipold
Hi Mansi,

Your files didn't attach.

There could be a couple reasons for this:
1. You might not have hit "Add the file" after you chose it for upload.
2. The file size might have been too big to attach. I'm not sure of the max file size allowable, but this has happened before (eg viewtopic.php?f=4&t=1496&p=4240)


Mike

Re: 2020-Trussart et al-preprint

PostPosted: Tue Jun 02, 2020 8:30 pm
by narulam
Hi,

I split up the files due to size limit and attached now. Please let me know your thoughts.

Thanks,
Mansi