FAQ  •  Register  •  Login

Data analysis development-wishlists and datasets?

Forum rules
Please be as geeky as possible. Reference, reference, reference.
Also, please note that this is a mixed bag of math-gurus and mathematically challenged, so choose your words wisely :-)
<<

jcvillasboas

Contributor

Posts: 41

Joined: Fri Apr 03, 2015 3:22 pm

Location: Rochester - MN

Post Wed Nov 15, 2017 4:46 am

Re: Data analysis development-wishlists and datasets?

Vinko and Mike,

I have explored cytofkit and started to like it a lot specially for offering a one-stop GUI-based solution. I was initially excited to see it supports flowSOM but I have not been able to generate daughter FCS files with the flowSOM cluster annotations (all I get are the phonograph clusters). I reached out to the developers and they acknowledged that deficiency in the code but - as far as I can tell - that has not been yet fixed. I was wondering if either of you (or anyone else) had found a way around it. Is there another package that supports this kind of flowSOM annotation so I can try a workflow similar to what Vinko presented on the link?

Best,

JC

ps: after the post was placed I went and checked the bioconductor website for cytofkit. Under the NEWS link (http://bioconductor.org/packages/release/bioc/news/cytofkit/NEWS in the documentation section I found indication that those bugs have now been potentially fixed. Will try it out.
<<

vtosevski

Contributor

Posts: 44

Joined: Wed Nov 20, 2013 12:50 pm

Location: Zurich, Switzerland

Post Wed Nov 15, 2017 12:06 pm

Re: Data analysis development-wishlists and datasets?

jcvillasboas wrote:I was wondering if either of you (or anyone else) had found a way around it. Is there another package that supports this kind of flowSOM annotation so I can try a workflow similar to what Vinko presented on the link?

I never bothered making a GUI to the workflow presented in the webinar but I can send you the code I used there. Recently I streamlined it and made it fully automated. In principle, you can just put the script in the folder with FCS files and execute it. The script will, then,
- read in data (no subsampling)
- transform appropriate parameters (default arcsinh with a factor of 5)
- take all reagent parameters and perform the tSNE
- take those same parameters and perform flowSOM clustering and metaclustering with k ranging from 5 to 15 (as it currently coded, you can change that).
- combine the 2 results and write them back into the initial FCS files (creates a copy in a subfolder and does not modify the input files)

You can then explore the results from a GUI, choose the best partitioning scheme for your data and report summary statistics.
This seem to work really well for our users so let me know if you want to give it a try.

Vinko
<<

jcvillasboas

Contributor

Posts: 41

Joined: Fri Apr 03, 2015 3:22 pm

Location: Rochester - MN

Post Wed Nov 15, 2017 2:43 pm

Re: Data analysis development-wishlists and datasets?

Thank you Vinko. I will reach out to you shortly via email to ask for the code. Much appreciated. JC
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Fri Nov 17, 2017 2:43 pm

Re: Data analysis development-wishlists and datasets?

mleipold wrote:So, yes, a versatile, user-friendly harmonizer would be a simple but highly useful addition. And, yes, we bench scientists like our GUIs, so that would be preferable ;)

Mike


Hi Mike,

We've added a simple FCS channel renaming tool to the cytutils package:

https://github.com/ismmshimc/cytutils#f ... l-renaming

A GUI is also in the works :)

Adeeb
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

vmotta

Participant

Posts: 13

Joined: Mon Jan 26, 2015 10:34 pm

Post Thu Dec 07, 2017 7:16 am

Re: Data analysis development-wishlists and datasets?

Hi,

I usually change parameter names using flowCore in R. I have re-written my code as a shiny application for changing names and parameters in FCS. It is a simple application that only change channel name and description (antibody name). It does not change any other metadata in the .fcs file. Thus, information associate with original file name, operator, instrument, date of acquisition is preserved. It creates new files and does not over-write your original file. It works for CyTOF and conventional flow files.

I have tested a few times and have not found any problems. However, let me know if you find any bug.

Please check the link if you are interested in: https://github.com/viniciusnmotta/ChangeFCSParameters

you can run from Rstudio, as follows:

1. install.packages(shiny)
2.library(shiny)
3.runGitHub("ChangeFCSParameters","viniciusnmotta")

Best Regards,
Vinicius
Mass Cytometry Core Facility, McGill University, Montreal, Canada
Previous

Return to CyTOF data analysis

Who is online

Users browsing this forum: AnnaBelkina and 10 guests