TCGA data: low / high grade ovarian cancer

Question

[erica.fary]

Dear All,

I have downloaded ovarian cancer annotation data from the TCGA using the TCGAbiolinks package (recount2 data) as follows:

ovary_rec2_tcga <- TCGAquery_recount2(project="tcga", tissue = "ovary")
tcga_sampleAnnot <- data.frame(ovary_rec2_tcga_scaled@colData)

I can check the cancer type with command like this one

unique(tcga_sampleAnnotgdc_cases.project.name)
# [1] "Ovarian Serous Cystadenocarcinoma" 

as I am only interested in the low-grade tumors, how can I know if it is a high-grade or low-grade "Ovarian Serous Cystadenocarcinoma" ?

are there only high-grade ones in TCGA data ?

(I was not able to find this information in the recount2 annotation table...)

thanks in advance !

erica

Answer 1

[Josip]

Hi Erica,

They're mainly high-grade, but are also poorly annotated what is a relic of the time when serous ovarian cancers were considered as a continuum rather than 2 distinct pathologies. TCGA-OV study is very 'clean' and represents HGSOC only.

Firehouse legacy and Pan-cancer atlas are poorly annotated and since only 80ish % of samples are p53 mutant proteins, I assume there are a lot of low-grade serous cancers there, but I just don't see a way of pulling them out with confidence. The presence of a lot of MMR mutations also indicates there are endometrioid cancers there too so you can't rely on p53 status either.

I've played around in cBioportal and I managed to pull out just 11 cases with reported G1 disease, of which only 3 were sequenced.

Good luck

[erica.fary]

thanks a lot for your input !