Thank you Sir for posting.
• 0 views
•
link
written
by
Kevin Blighe
how to get ER, PR and HER2 data from TCGA BRCA
Hi I have dowloaded the BRCA data from TCGA using TCGABiolinks
I have done this:
BRCARnaseqSE <- GDCprepare(query.a, directory = "BRCA_all")
sample.info <- SummarizedExperiment::colData(BRCARnaseqSE)
Now I want to get data on ER, PR and HER2 - positive, negative samples, but I can not find any such columns. Here is the description of sample.info
names(sample.info)
[1] "sample" "patient" "barcode"
[4] "shortLetterCode" "definition" "days_to_recurrence"
[7] "ajcc_staging_system_edition" "days_to_last_follow_up" "classification_of_tumor"
[10] "age_at_diagnosis" "icd_10_code" "prior_malignancy"
[13] "morphology" "created_datetime.x" "last_known_disease_status"
[16] "tumor_stage" "updated_datetime.x" "days_to_last_known_disease_status"
[19] "ajcc_pathologic_t" "treatments" "year_of_diagnosis"
[22] "synchronous_malignancy" "state.x" "ajcc_pathologic_m"
[25] "progression_or_recurrence" "prior_treatment" "site_of_resection_or_biopsy"
[28] "ajcc_pathologic_n" "days_to_diagnosis" "tissue_or_organ_of_origin"
[31] "diagnosis_id" "tumor_grade" "primary_diagnosis"
[34] "ajcc_pathologic_stage" "created_datetime.y" "cigarettes_per_day"
[37] "state.y" "bmi" "weight"
[40] "exposure_id" "height" "alcohol_intensity"
[43] "alcohol_history" "updated_datetime.y" "years_smoked"
[46] "gender" "created_datetime" "days_to_birth"
[49] "state" "race" "ethnicity"
[52] "demographic_id" "year_of_birth" "vital_status"
[55] "age_at_index" "year_of_death" "updated_datetime"
[58] "days_to_death" "bcr_patient_barcode" "project_id"
[61] "disease_type" "dbgap_accession_number" "name"
[64] "released" "releasable" "primary_site"
[67] "is_ffpe" "subtype_patient" "subtype_Tumor.Type"
[70] "subtype_Included_in_previous_marker_papers" "subtype_vital_status" "subtype_days_to_birth"
[73] "subtype_days_to_death" "subtype_days_to_last_followup" "subtype_age_at_initial_pathologic_diagnosis"
[76] "subtype_pathologic_stage" "subtype_Tumor_Grade" "subtype_BRCA_Pathology"
[79] "subtype_BRCA_Subtype_PAM50" "subtype_MSI_status" "subtype_HPV_Status"
[82] "subtype_tobacco_smoking_history" "subtype_CNV.Clusters" "subtype_Mutation.Clusters"
[85] "subtype_DNA.Methylation.Clusters" "subtype_mRNA.Clusters" "subtype_miRNA.Clusters"
[88] "subtype_lncRNA.Clusters" "subtype_Protein.Clusters" "subtype_PARADIGM.Clusters"
[91] "subtype_Pan.Gyn.Clusters"
I cannot see any info related to ER status, or something like er_status_by_ihc, or pr_status_by_ihc or her2_status_by_ihc
Please help!
Thanks!
2 answers
Not sure about TCGAbiolinks but the information is definitely available at the GDC Data Portal: A: How to download triple negative breast cancer RNA-seq fpkm data from GDC.
You can feasibly use that information and link it up to your TCGAbiolinks output.
Kevin
• 0 views
•
link
written
by
Kevin Blighe
I came across this post because I had the same question. Here's the way I did it 3 years ago (saved in an old code file) and tested today (20-Nov-2023):
library(tidyverse)
library(TCGAbiolinks)
query <- GDCquery(project = "TCGA-BRCA",
data.category = "Clinical",
data.type = "Clinical Supplement",
data.format = "BCR Biotab")
GDCdownload(query)
clinical.all <- GDCprepare(query)
tcga_brca.clin <- clinical.all$clinical_patient_brca
tcga_brca.tnbc_samples <- tcga_brca.clin %>%
filter(er_status_by_ihc == 'Negative' &
pr_status_by_ihc == 'Negative' &
her2_status_by_ihc == 'Negative') %>%
pull(bcr_patient_barcode)
tcga_brca.er_samples <- tcga_brca.clin %>%
filter(er_status_by_ihc == 'Positive' &
her2_status_by_ihc != 'Positive') %>%
pull(bcr_patient_barcode)
tcga_brca.her2_samples <- tcga_brca.clin %>%
filter(her2_status_by_ihc == 'Positive') %>%
pull(bcr_patient_barcode)
The BCR Biotab gives info on very limited number of samples. The BCR XML data.format has info on a lot more samples but I cannot find a function that parses it. Even the GDCprepare_clinic function seems to work on a rather limited subset of XML fields. I'm writing my own hack, will update as soon as it's done.
I think I was wrong - both the BioTab and XML give us the same data, just in a different number of files. I ran a preliminary test: the 116 TNBC patient IDs (TCGA-XX-XXXX) overlap a 100% between the two formats.
Log in to answer this question.