Survival comparison High vs low expression

389 views
Skip to first unread message

Rowyn Liebau

unread,
Jan 24, 2023, 2:42:30 PM1/24/23
to cbiop...@googlegroups.com
Hello, 
I want to use the cbioportal site to compare patient survival curves between tumors displaying high expression and low expression of a gene. I can analyze groups comparing altered and unaltered samples, but the altered group is a mix of high and low expressing patients, likely obscuring any effects. Is there a way to set up separate high and low expression groups based on mRNA expression and a z score cutoff, as I can do for the oncoprint tract?
Thank you for your help,

--
Rowyn Liebau
She/her 
PhD candidate Gautier Lab
Columbia University

Tali Mazor

unread,
Jan 24, 2023, 3:44:19 PM1/24/23
to Rowyn Liebau, cbiop...@googlegroups.com
Hi Rowyn,

There are two ways you can approach this. If you run a query, which sounds like the approach you've been taking, you can use OQL to generate separate tracks for high vs low expressing samples based on a z-score cutoff, eg
EGFR: EXP>1
EGFR: EXP<-1
You'll see this in OncoPrint as two distinct tracks. If you then click over to the Comparison tab, you have the option to deselect the 'Altered group' and 'Unaltered group' and instead select the groups corresponding to your query/OncoPrint tracks:
image.png


Another option is to use study view to define groups, which is demonstrated in more detail in this video: https://www.youtube.com/watch?v=HTiKUXk0j0s

-Tali




--
You received this message because you are subscribed to the Google Groups "cBioPortal for Cancer Genomics Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cbioportal+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cbioportal/CAPTU_s3DB8rzxtFwKgSfw5SFiKDr%2B2giFWnk6nhN47P6JTpEWQ%40mail.gmail.com.
Reply all
Reply to author
Forward
0 new messages