CNV data for Urothelial Bladder carcinoma

Question

[arpa530]

Dear All,

I am going to analysis of CNV data of Bladder cancer. My specific aim is to identify the most commonly altered region in the urothelial bladder carcinoma. In this regard I have planned to retrieved the data from GEO and cBioportal. I want to consider more CNV data but unfortunately failed to collect the more CNV data of Bladder cancer. I don't know whether TCGA will provide us to access the Bladder Cancer CNV data. In addition to this I have downloaded three files of Bladder cancer CNV data from GEO. But in GEO they have provided the raw data. So I cant understand how to process it further. Hence, my request if you kindly assist me from where I can get the more Bladder cancer CNV data and how shall I process if then I shall be highly obliged.

[Irsan]

You need a local bio-informatician (preferably with experience in cancer genomics!) that will help you do the job. For example, samples for which only exome sequencing data is available, you should use the off-target reads to make copy number profiles. They will be of very good quality, almost similarly to whole genome sequencing, depending on how high your resolution needs to be. Once you have collected data of various sources (SNP-array, CGH-arrays, ExomeSeq, (shallow) whole genome seq) you should process the raw/semi-raw data to segmented data files with circular binary segmentation (since these are cancer samples, don't use HMMs to estimate some discrete level op copy number) implemented by e.g. DNAcopy (bioconductor). Once you have segmentation files, you can easily browse the results in IGV and/or use GISTIC to find the most recurrent/high-level/focal regions, i.e. the ones that you want to focus on.

Answer 1

[TriS]

You can try the TCGA firehose website. The link should send you to the Bladder Urothelial Carcinoma page.

There you can select e.g. the MAF file (Mutation Annotation File) that contains info on SNV, insertions and deletions, or download the SNP array data.

[arpa530]

Thanks for your response. I have checked the site and what they have showed I have pasted below now I didn't get which I have to download for the work.

   Clinical Analyses  
   CopyNumber Analyses  
Aggregate AnalysisFeatures
CopyNumber Clustering CNMF
CopyNumber Clustering CNMF thresholded
CopyNumber Gistic2
CopyNumberLowPass Gistic2
Correlate Clinical vs CopyNumber Arm
Correlate Clinical vs CopyNumber Focal
Correlate CopyNumber vs mRNAseq
Correlate molecularSubtype vs CopyNumber Arm
Correlate molecularSubtype vs CopyNumber Focal
Pathway Paradigm RNASeq And Copy Number

[TriS]

First...I think you need to have a focused question and then pick the data based on that...I kinda of feel that "looking at the most common CNVs in bladder cancer is not much focus, but, with the benefit of doubt, it can be a good starting point

Second...I see you are new-ish so remember to use the "comment" section to add a reply, unless you suggest a solution to the initial question

Third...not sure we are looking at the same page because I'm looking at SNP6 CopyNum and it has the following:

• genome_wide_snp_6-FFPE-segmented_scna_hg18 (MD5)
• genome_wide_snp_6-FFPE-segmented_scna_minus_germline_cnv_hg18 (MD5)
• genome_wide_snp_6-segmented_scna_minus_germline_cnv_hg19) (MD5)
• genome_wide_snp_6-segmented_scna_hg19 (MD5)
• genome_wide_snp_6-FFPE-segmented_scna_hg19) (MD5)
• genome_wide_snp_6-segmented_scna_hg18 (MD5)
• genome_wide_snp_6-FFPE-segmented_scna_minus_germline_cnv_hg19 (MD5))
• genome_wide_snp_6-segmented_scna_minus_germline_cnv_hg18 (MD5)

Those are CNV data from SNP6 arrays....which one...well, it kinda of depends from what you want.

You might want those mapped to hg19 minus germline so that it shows the somatic mutations specific for the tumor.

Also, FFPE is for Formalin Fixed and Paraffin Embedded tumors so it's up to you if you want to use those data

The firehose help page has a brief description of the data you wrote, Gistic is an algorithm for calling copy number alterations (gistic2 is the newer one published in genome biology), the CNMF is a clustering algorithm but I admit I don't know much about it. I wouldn't use the low pass.

The other data are pre-analyzed results correlating copy number with other parameters...

Overall, you can also get a good overview of the bladder data by using this tool from firebrowse

if after reading the help pages, looking at the file fomats and getting maybe a few more ideas, you are still unsure about how to use those data then probably the best thing is to ask to your local bioinformatician or biostatistician so that they will be able to show you step by step what/how to do it